August 2026 News

14 August News

Healthcare : Ant Group upgraded its physician-facing application “Haodabu Doctor Edition” (好大夫医生版) into “Ant Alpha Doctor Edition” (蚂蚁阿福医生版), introducing two AI assistant modules: an “AI Digital Twin Assistant” and an “AI Medical Assistant.” The AI Digital Twin Assistant allows physicians to create personal AI avatars that operate 24/7 to handle pre-consultation intake, patient follow-up, and health record summarization, while the AI Medical Assistant aggregates over 60 million medical documents, clinical guidelines, and 140,000 drug information sheets for evidence-based decision support. The platform partnered with Wolters Kluwer (for authoritative clinical content, including exploring integration with UpToDate) and DP Technology (深势科技), whose Bohrium Science Navigator provides structured scientific literature retrieval. The system links to the consumer-facing Ant Alpha app, which has over 100 million users and processes more than 10 million daily health inquiries; an “AI-answers-first, doctor-verifies” model already used in dermatology has been expanded to 16 departments including obstetrics, gynecology, and ENT, achieving over 90% agreement between AI and tertiary hospital physician judgments. Ant Alpha Doctor Edition also opened enrollment to internet hospitals as institutional entities, with Guangzhou Women and Children’s Medical Center among the first hospitals to integrate. Separately, in Hong Kong, mainland Chinese TCM chain Gushengtang (固生堂) opened four clinics (Shau Kei Wan, Prince Edward, Tuen Mun, Wan Chai) as part of its overseas expansion, and the company has deployed AI digital twins of 15 senior TCM experts, including “guoyi dashi” (national TCM masters), which practitioners in Singapore can consult in real time during patient visits for diagnostic and prescription support.

Entertainment and Short Dramas : China’s AI-generated short drama (AI短剧) sector expanded explosively in 2026, with AI-produced dramas comprising about 95% of the roughly 128,000 microdramas released in Q1, though real per-episode production costs (accounting for labor) typically ran 50,000-200,000 yuan versus 500,000-2,000,000 yuan for live-action dramas; despite the volume, over 99% of AI titles found little audience, with fewer than 150 titles surpassing 100 million views and the top 5% of projects capturing 70% of profits. Independent Chinese creator Jiang Lan single-handedly produced the AI short drama “Guixu” (《归墟》), a post-apocalyptic drama that surpassed 100 million views despite computing costs of roughly 16,000 yuan per 8-minute episode. Youku president Wu Qian publicly argued AI cannot replace real actors, citing data showing AI microdramas captured 95% of listings but only 4% of viewership, while live-action dramas retained 96% of viewership with 2.3 times higher user payment willingness and 41% higher completion rates; Youku launched the “Whale Surpass Class” (虎鲸超越班) to cultivate young actors with stable income. Actor Dong Yong criticized AI dramas as “upgraded cartoons.” Separately, veteran actress Qi Wei became the first active mainland celebrity to formally authorize an official AI digital twin, named “Han Qingxia” (韩清夏), intending it as an independent commercial IP for use in AI comics, virtual livestreaming, and interactive content; however, a promotional clip for her AI-driven work “Modern Summer Apocalypse” (《末日盛夏》) drew criticism for an upskirt-style camera angle, prompting legal experts (including lawyer Li Zhenwu and Zhao Liangshan) to warn that portrait rights cannot be “bought out” and that AI digital twins blur accountability when content crosses ethical or legal lines. Chengdu’s Huya (虎牙) livestreaming platform released its self-developed real-time multimodal digital human model VAM 1.0, aiming to reduce reliance on human streamers and lower content production costs, while continuing to integrate AI tools into games like Delta Force and League of Legends; Huya’s overall Q2 revenue grew 11% year-on-year, driven by game publishing (revenue up 54.1%) even as livestreaming revenue fell 4.48%. Chinese AI-generated influencer/actress “Fang Taozi” (方桃子), from the AI drama “The Girl Who Got Laid Off” (《被裁掉的女孩》, over 200 million views), amassed over 460,000 followers and commanded advertising rates up to 258,000 yuan per 60-second video—higher than many real influencers—but her first commercial deal, promoting “Jianguoli” (坚果力) colored contact lenses with a first-person claim of “comfortable all day,” was removed from platforms amid public backlash and scrutiny; legal experts noted that under Chinese advertising law, AI virtual humans without legal personhood cannot serve as advertising endorsers, and the false claim of physical product experience constituted a violation of false advertising rules governing Class III medical devices, exposing the brand, MCN agency “Wo De Yingshi,” and platforms to potential liability. AI selection/talent shows using zodiac-themed virtual idols, such as “12 Star Training Competition” (《12星练赛》) and “Star Light 108” (《星光108》), generated early buzz (peaking near 760,000 likes) driven by fan-culture behaviors like vote-brigading, but engagement dropped sharply within weeks as viewers cited difficulty distinguishing between similar AI faces and lack of authentic unpredictability; meanwhile AI-assisted content like “100,000 Challenges” (《十万个挑战》), which uses AI for game-based challenge videos rather than idol-building, retained steadier growth and brand sponsorships. Longstanding artists including Wang Zuxian (王祖贤), Endy Chow (吴启华), You Benchang, and Jackie Chan have licensed AI-generated or de-aged digital likenesses for games and short films. Chinese short video creator group “Haojiahuo” (郝家豁) from Xi’an, known for recreating Chinese textbook stories with live actors and physical sets rather than AI, surpassed 4.8 million followers on Douyin, explicitly rejecting digital humans and AI-generated scenes in favor of authentic performance for their historically grounded content. In Wuhu, Anhui, the country’s first AIGC-produced red-heritage short drama “An Du” (《暗渡》), depicting a 1949 Yangtze crossing intelligence story, was produced using the “OneStory” AI creation platform (developed by Anhui Chuangyao Technology) with local Wuhu computing infrastructure for rendering, digital human performance, and voiceover, and premiered as part of a broader red-tourism MR immersive experience project. AI-generated content is also reshaping film/TV labor markets: reports describe traditional actors, including one Central Academy of Drama graduate, being displaced by “digital human” productions and returning to rural work such as vegetable selling, while script-to-screen AIGC workflows (as demonstrated by director He Bi and studios in Changchun) have compressed production teams from about 50 people to roughly 5 and shortened production cycles from a month to about a week; Changchun has built an “AIGC film ecosystem” spanning theatrical films, TV series, microdramas, and AI solo animated comics, supported by government subsidies and youth-oriented “one-person company” AI training programs (OPC innovation center).

Digital Human Public Services and Government Applications : State Grid’s Jinzhou Electric Power Supply Company (国网锦州供电公司) in Liaoning deployed an interactive digital human system in its customer service hall to provide round-the-clock guidance for bill payment, invoice issuance, and electricity usage inquiries, supplementing human staff and operating overnight for basic consultation services. China Unicom’s Liaoning branch showcased smart elderly-care solutions and AI applications during a series of “State-Owned Enterprise Open Day” events across 14 cities in Liaoning province from July to August 2026. In Xinjiang, Xi’an Jiaotong University’s Digital Cell team is developing “digital twin cells” combining single-cell omics, spatial multi-omics, and AI modeling for precision medicine, though this work is research-focused rather than consumer-facing. Zhangwu village-level “digital cultural stations” in Zhejiang’s Jiaxing/Hangzhou collaboration used AI robots and digital humans in a joint “civilization co-construction” event between Hangzhou’s Binjiang district and Wenzhou’s Taishun county, featuring an intelligent robot named “Xiaobin” performing crosstalk comedy alongside human presenters at a public theoretical outreach event. Xi’an’s public library (Beijing City Library branch, referenced for comparative context but centered on Beijing) has a “metaverse experience hall” digital human designed with a “childlike and futuristic” persona for interacting with readers via large language models. Wuxiang County (Shanxi) produced an AI-generated promotional video for tourism featuring “digital twins” of visitors exploring guerrilla-war historical sites and local attractions. In Yuexiu (Changsha) and Guiyang, cultural heritage sites are deploying AI digital humans: Yuelu Academy in Changsha created a digital human avatar of Confucian philosopher Zhu Xi, enabling visitors to interact via smartphone in a virtual recreation of Song-dynasty lecture halls, alongside a “China Academy Digital Museum” IP, drawing over 5 million study-tour visitors in 2026. In Guiyang, Guizhou, an AIGC tourism project “Ban Meng Zhi Xing” (伴梦智行) by Guizhou Ganshengtengwei Technology Company creates personalized AI-generated souvenir content (photos, short videos, digital keepsakes) for visitors at red-tourism sites and Guiyang’s Yunyan urban district, with plans to expand to performance venues like “Colorful Guizhou Style” and mountain tourism sites. Shaanxi province’s official policy plan (issued August 2026 by three provincial departments) calls for building 20 “metaverse consumption demonstration scenarios,” developing virtual attraction experiences, and creating digital/metaverse artworks based on Terracotta Warriors and ancient city walls, alongside AI digital human-hosted livestream marketing for regional brands; concrete projects already underway include immersive XR experiences like “Zhang Qian’s Journey” and VR productions “Great Tang Time Travel,” “Great Qin Chronicles,” and “Dream Back to Shimao” at sites including Qianling Mausoleum and Yulin’s Shangu Museum.

Retail, Livestreaming, and Marketing : Chinese e-commerce and livestreaming platforms are integrating digital human hosts at scale: JD.com’s Q2 2026 earnings report disclosed that its digital human livestreaming tool “JoyAvatar” has served over 80,000 merchants, with digital human livestream accounts tripling year-on-year and daily active livestream accounts up sixfold; during the “618” shopping festival, digital human livestream duration and transaction value both grew more than fourfold year-on-year. In Guangzhou, China Unicom and Vipshop (唯品会) jointly launched China’s first “10-gigabit smart livestreaming” e-commerce base, integrating AI livestream assistants (automated narration and interaction), AI data analytics tools, and smart digital human generation into a 120,000-square-meter e-commerce production and livestreaming hub in Guangzhou’s Liwan district. Snack food company Qiaqia Foods (洽洽食品) launched “Chacky” (洽可儿), described as the snack food industry’s first brand-exclusive digital IP, complete with an AI-generated debut single, and integrated it across livestreaming, smart customer service, guided shopping, and branded short dramas within days of launch — positioned as a lower-cost, always-available alternative to celebrity endorsement deals amid pressure on Qiaqia’s core sunflower seed business. State-owned publisher Liaoning Chubanchuanmei (辽宁出版传媒) has deployed “AI book digital humans” using a “one book, one model” approach to provide reading companionship and Q&A for educational publications like “Mathematics Weekly.” WeChat’s public account platform (owned by Tencent) opened its native “AI Digital Twin” (AI分身) feature to hospitals, enabling 24/7 automated patient inquiry responses without requiring hospitals to code custom solutions. Actress Qi Wei’s digital twin “Han Qingxia” also represents a marketing/IP case, discussed above under Entertainment, given her stated intent to create an open licensing framework for use in advertising, virtual livestreaming, and interactive commercial content. Chinese cultural export companies in Jiaxing, Zhejiang—including Haohao Jianghua Culture Technology and Jiuzhou Culture Media—are using AIGC tools and digital human technology to accelerate localization and overseas distribution of web novels, web dramas, and web games (“new three items” of Chinese cultural export), compressing short-drama production cycles to 7-10 days and building AIGC-driven industrial parks in Jiaxing’s Yunhe Bay Digital Culture Industrial Park, home to over 110 audiovisual technology companies integrating AIGC, digital humans, and virtual production.

Enterprise and Financial Services : Chinese enterprise software company Yonyou (用友) launched its “YonSuite” AI-native enterprise platform featuring a unified brand character “yowo” (读作"有我"), representing customized “digital employees” for different business functions such as finance, sales, procurement, production, and HR; the system categorizes AI agents into “digital doubles,” “digital employees,” and “digital teams,” with lifecycle management (onboarding, authorization, token quota, performance evaluation, retirement) for digital employees deployed across manufacturing client Shanghai Phoenix Bicycle, tea brand 700CC, and pastry brand Luxiangyuan. Ant Group’s Alpha Doctor platform (detailed under Healthcare) exemplifies AI digital-twin deployment in enterprise/financial-adjacent healthcare services. Chinese entrepreneur Chen Tianqiao’s venture EverMind/EverMe launched a personal AI memory product in beta, featuring a “digital twin” module allowing users to consolidate personal data across files, knowledge tools, and various AI agents into a continuous personal memory asset for use across multiple AI services.

Cultural Heritage and International Exchange : Beijing Institute of Graphic Communication (北京印刷学院) led the creation of a digital art exhibition “Forbidden City: 600 Years of Mystery” at La Cité de l’Histoire in Paris, France, marking the second anniversary of the Beijing Central Axis’s UNESCO World Heritage inscription; the exhibition used AIGC technology to generate stylized digital renderings of Ming/Qing-era court officials integrated into UE5 real-time rendered historical scenes, alongside a 360-degree immersive panoramic film guided by three European visitor perspectives (Matteo Ricci in the 16th century, French painter Jean-Denis Attiret in the 18th century, and a contemporary traveler), with digital human elements created via AI-driven motion generation and style transfer techniques layered onto photorealistic 3D-rendered environments of the Forbidden City and Beijing’s fourth ring road area.

Ethical, Regulatory, and Public Discourse Contexts : The “Fang Taozi” AI virtual influencer controversy (detailed under Entertainment) became a flashpoint for broader debate on AI virtual human advertising regulation in China; legal commentary from the Economic Observer (经济观察网) and other outlets clarified that under China’s Advertising Law, only natural persons, legal persons, or organizations can serve as advertising endorsers, meaning purely AI-generated virtual humans like Fang Taozi cannot be legally classified as endorsers—though authorized digital twins of real people (such as Qi Wei’s “Han Qingxia”) retain the legal liabilities of their real-world human originators; commentators called for mandatory AI-content labeling, clearer rules distinguishing “original AI virtual humans” from “authorized human digital twins,” and accountability frameworks assigning liability to brand owners, virtual IP operators, and platforms rather than allowing “virtual” status to be used as a liability shield, particularly for regulated categories like Class III medical devices (including decorative colored contact lenses, commonly mislabeled as “meitong” despite that being a registered Johnson & Johnson trademark). Separately, a Xinmin Evening News commentary titled “Do You Hate ‘AI Faces’?” reflected on growing public “physiological disgust” toward homogenized, uncanny AI-generated faces proliferating in Chinese short dramas and short videos, criticizing the industry’s use of AI facial generation as a cost-cutting shortcut that produces “assembly-line” aesthetics lacking emotional depth, while acknowledging legitimate uses of digital humans, virtual humans, and AI animation when creativity retains a human core. Guangzhou Daily commentary “Performers Can Be ‘Copied,’ But Acting Cannot Be ‘Pasted’” (演员可"复制",演技难"粘贴") examined the trend of licensed actor digital twins (including Qi Wei, Wang Zuxian, Endy Chow), arguing that while proactive licensing gives performers more control than passive “face theft” by unauthorized AI operators, clear rules are needed regarding licensing scope, usage duration, retraining rights, revenue sharing, and creator review/rejection rights, and warning that treating digital twins as mere low-cost traffic tools risks producing content with polished faces but no soul. Amazon’s newly enforced “Synthetic Performer” disclosure policy—requiring metadata tagging (contains-synthetic-performer) for AI-generated realistic human figures in product listings, including partial body parts like hands or arms—was reported by Chinese cross-border e-commerce media (Cifnews/雨果跨境) as a compliance requirement directly affecting Chinese sellers on the platform, prompting detailed Chinese-language guidance on manual and batch tagging procedures using tools like ExifTool.

13 August News

Healthcare : Ant Group upgraded its physician-facing medical application from Haodaifu Doctor Edition to Ant AFu Doctor Edition (蚂蚁阿福医生版), positioned as China’s first medical AI workstation covering online consultation, patient management, and research in a single platform, one year after Ant Group’s acquisition of the veteran online medical platform Haodaifu Online (好大夫在线). The upgraded platform integrates two AI assistant modules—an AI Medical Assistant and an AI Avatar Assistant—into a physician workflow serving hundreds of thousands of registered doctors nationwide. The AI Avatar Assistant allows physicians to create a personalized digital double by uploading their professional information; this avatar operates 24/7 to handle pre-consultation triage, patient Q&A, and post-visit follow-up, automatically compiling case summaries into patient health records for physician review, effectively extending a doctor’s presence beyond limited clinic hours. The AI Medical Assistant provides physicians with free, real-time, traceable knowledge retrieval, aggregating over 60 million medical papers, clinical guidelines, top-tier journal resources, and more than 140,000 drug package inserts; Ant AFu has also formed strategic partnerships with Wolters Kluwer (威科集团) for authoritative clinical content, including exploring integration with UpToDate, and with DP Technology (深势科技) to leverage its Bohrium scientific literature search capabilities. The platform maintains a “Doctor in the loop” principle, ensuring AI outputs are reviewed and confirmed by licensed physicians, with previously reported human-AI answer consistency exceeding 90% in specialties such as dermatology, now expanded to 16 departments including obstetrics, gynecology, and ENT. The physician platform is deeply integrated with the consumer-facing Ant AFu App, which has surpassed 100 million registered users and handles over 10 million daily health inquiries, connecting more than 5,000 hospitals and 300,000 licensed doctors; eight academicians and over 2,000 physicians have launched their own AI avatars on the Ant AFu platform. The upgrade also opened the platform to internet hospital registrations from institutions nationwide, with Guangzhou Women and Children’s Medical Center among the first to sign on, allowing hospitals to consolidate online consultation, follow-up management, and electronic medical records under a single account. Separately, in the fintech-adjacent healthcare space, China Post Consumer Finance (中邮消费金融) deployed a virtual digital human named “Youxiaobao” (邮小宝), developed with a university partner, to provide round-the-clock customer service, loan pre-review interviews, and financial literacy education, achieving a 98% customer intent recognition accuracy rate.

Entertainment : AI-generated short dramas have become a dominant force in China’s content industry, with AI-produced dramas accounting for over 95% of new short-drama releases in early 2026 (roughly 122,000 of 128,000 titles), though they capture only about 4% of total viewership and a hit rate below 0.5%, reflecting a stark mismatch between production volume and audience engagement. The breakout case is virtual AI character Fang Taozi (方桃子), the lead of the hit AI drama “The Girl Who Was Cut” (被裁掉的女孩), produced by director Feifei Fei and streamed via ByteDance’s platforms including Douyin and Hongguo Duanju; the character amassed over 45.4 million followers, generated more than 300 million views across two seasons, and commanded advertising rates up to 258,000 yuan for a 60-second video—surpassing many human influencers with millions of followers—before her endorsement of a Jianguoli (坚果力) colored contact lens brand was pulled offline within a week due to backlash over an AI character claiming first-person physical comfort she could not possibly experience, prompting intervention from market regulators in Anhui province. Actress Qi Wei (戚薇) became the first active mainstream Chinese celebrity to officially authorize a regulated AI digital double of herself, named Han Qingxia (韩清夏), unveiled August 6, 2026, with an independent persona, storyline, and world-building distinct from her own identity; the avatar is set to star in the apocalyptic-themed AI drama “Last Summer of the Apocalypse” (末日盛夏) later in the year, and Qi Wei cited excitement about ByteDance’s Seedance 2.5 video generation model as a technical catalyst. Other Chinese entertainment companies have pursued AI actor commercialization: Ciwen Media (慈文传媒) unveiled an “Yingdian” (影殿) AI actor label at the Pingyao International Film Festival with four AI drama titles and eight commercial AI actor projects, including actor Liang Tian (梁天) confirming authorization of his facial and vocal data for AI digital double use in the drama “Chang’an Illusory Night” (长安幻夜); veteran actresses and actors including Wang Zuxian (王祖贤), who licensed her 20s-to-30s likeness to NetEase’s game “Tianxia” (天下), Wu Qihua (吴启华), who sold rights to his 20-year-old likeness to an AI film production company, and 93-year-old actor You Benchang (游本昌), who authorized his likeness and voice for the AI micro-drama “Jigong: The Fake Dragon-Subduing Monk” (济公之冒牌降龙) while serving as artistic director, have all monetized archival or youthful likenesses via AI. Actor Jackie Chan appeared in the film “The Legend” using AI to recreate his 27-year-old appearance for over 70% of his on-screen performance. Chinese streaming platform Youku, via its parent Whale Entertainment Group (虎鲸文娱), publicly committed to prioritizing real human actors over AI performers, with President Wu Qian citing that despite AI dramas comprising over 95% of Q2 2026 content volume, they captured only about 4% of total viewership; Youku’s real-actor productions such as “The Nine-Gated City” (九门), starring Chen Weiting (陈伟霆) and Zeng Shunxi (曾舜晞), and “Unsolved Case” (悬案), starring Wang Chuanjun (王传君), were highlighted as evidence that nuanced human performance remains commercially and critically superior; the company also operates a “Whale Surpassing Class” (虎鲸超越班) talent incubator offering full-time employment to emerging actors. AI drama production companies such as Chuangxiang Shuwei (创想数维), in partnership with Aotuo Sports Culture Investment and Sony China, released the XR-and-AI hybrid drama “Shh! She Is a Dream Ferryman” (嘘!她是渡梦师) using its self-developed MetaBox XR virtual production system paired with the Coruscant rendering server and Vormir AI creative toolkit, reducing scene production costs by roughly 60% while retaining real human actors for performance authenticity; Wanxing Technology (万兴科技) showcased its Wanxing Studio (万兴剧厂) platform at the Dongguan International Animation Expo, supporting AI-assisted script analysis, IP asset management, and cinematic-grade storyboarding for productions such as “Genius Mecha Master” (天才机甲师, over 300 million views) and “Fortune Delta Force” (气运三角洲, over 270 million views), and became among the first platforms to integrate ByteDance’s Seedance 2.5 model. Talent agency and IP operator You You Tech (创壹科技), which manages digital human Liu Yexi (柳夜熙), reported deliberately declining brand deals requiring simulated physical product experience (such as food or cosmetics) for its AI personas, citing legal and ethical risk awareness lacking among newer, less experienced AI drama production teams. Meanwhile, actress and startup figure discussions noted that mid-tier and background human actors in short-drama production hubs like Hengdian have faced sharp declines in bookings and daily wages amid the AI drama surge, with commentators including Hu Xijin publicly opposing the expansion of AI-generated content into feature-length cinema on grounds of preserving traditional performing arts.

Marketing and Advertising Ethics : The “Fang Taozi” contact lens advertisement controversy has triggered widespread regulatory and legal scrutiny over AI virtual human endorsements in China. Legal experts, including Zhao Zhanling (赵占领) of Beijing Jiawei Law Firm and Zhang Dezhi (张德志), deputy secretary-general of the China Advertising Association, noted that decorative colored contact lenses are classified as Class III medical devices under Chinese regulation, and China’s Advertising Law explicitly prohibits the use of any endorser—human or otherwise—to make recommendations or endorsements for medical device products; the ad’s claim of “comfortable all-day wear” was flagged as fabricated experiential testimony given the AI character’s lack of physical senses, constituting a potential violation of Article 28 (false advertising) and Article 16 of China’s Advertising Law. Because AI virtual humans lack independent legal personality under Chinese law, responsibility falls to the brand owner, the IP operating company, and content platforms rather than the AI character itself; commentary in Xinhua-affiliated and China News Service outlets called for AI-generated content labeling requirements, a digital human registration/filing system, and clearer joint-liability rules among brands, MCN agencies, and platforms. The Guangzhou Panyu district market regulator and later the Langya district regulator in Chuzhou, Anhui, became involved in reviewing the case after the advertisement was removed from platforms. Separately, Xiaohongshu (小红书) announced an “AI Governance Rules Announcement” on August 6, 2026, extending mandatory AI-content labeling requirements to account-level identification, requiring accounts operated using AI virtual human personas to disclose this status on their profile pages and activate an “AI Virtual Human” tag, with graduated penalties including traffic throttling for non-compliance. Comparisons were drawn to earlier Chinese virtual idols such as Luo Tianyi (洛天依) and to Japan’s Imma and China’s AYAYI, whose commercial models relied on IP licensing and symbolic brand collaboration rather than first-person product testimonials, reinforcing industry consensus that AI personas are commercially viable for image-based and symbolic marketing but face acute legal risk when deployed for experiential product endorsement, particularly in medical device, food and beverage, and cosmetics categories.

Cultural and Public Sector Applications : Chinese local governments and cultural institutions have deployed AI virtual humans and digital avatars for heritage promotion and civic engagement. In Huanan County, Heilongjiang province, local propaganda officials introduced AI theory-lecturing avatars including “Xiaohua Xuan” (桦小宣), designed around the region’s ecological landscape, along with locally-themed IP characters “Zizi” and “Susu” based on the county’s signature perilla crop, and a real-person AI digital double of a civic official, to modernize ideological and policy education content; the initiative produced three short videos garnering 32,000 views and included AI literacy training for over 500 local officials. In Hangzhou, the Leifeng Pagoda scenic area launched a digital human tour guide named “Xiao Leifeng” to accompany visitors through an interactive workshop reproducing a Five Dynasties-era gilt mandarin duck ornament excavated from the pagoda’s underground crypt, allowing tourists to create embroidered silk scarves using traditional “xiaojin” gold-application techniques while engaging with the AI guide for historical context. Guilin Tourism Co. reported deploying an AI digital human welcome/interaction project at its Liangjiang Sixiang scenic area developed jointly with a university, alongside an AI travel assistant “Guilin Yicheng You” built on the DeepSeek large language model, and an AI companion mascot toy at the Yinzi Rock scenic site. In Hunan province’s Malanshan video content hub, broadcaster Hunan TV’s self-developed “Mango model” (芒果大模型) powers a self-developed digital human “Xiao V” that guides visitors through the Malanshan Video Culture and Creativity Park, while the AI-driven sci-fi drama “God’s Pen” (神·笔), produced with science fiction author Liu Cixin as literary supervisor, achieved 80-90% AI penetration in production. Educational applications include Dalian Nationalities University’s deployment of a digital human admissions counselor (“智能云宣讲”) alongside an AI phone-answering robot to support round-the-clock prospective student inquiries during its 2026 admissions cycle. In vocational and rural development efforts, Guangdong’s China International Cartoon and Digital Industry Fair in Dongguan featured AI-integrated plush toys with embedded AI chips capable of conversational interaction for children, alongside AI-generated animation and digital human demonstrations aimed at attracting overseas buyers from the Middle East and Southeast Asia.

Ethical, Legal, and Regulatory Contexts : China’s evolving virtual human governance landscape confronts three structural challenges according to legal scholars at Wuhan University’s law school: the technical opacity of deepfake and AI-generation technology, which allows operators to obscure or falsify AI-content labeling; the fragmented responsibility chain across technology providers, service operators, users, and distribution platforms, which complicates liability attribution when personality rights violations, disinformation, or data breaches occur; and the tension between fostering innovation and enforcing sufficiently strict oversight, given that China’s digital human core market was valued at 48.06 billion yuan in 2025 and is projected to reach 57.29 billion yuan in 2026. Commentators called for a full-chain liability framework clarifying obligations of technology providers, service operators, users, and platforms; innovative “technology-enabled regulation” combining pre-emptive warning systems, deepfake detection upgrades, and risk-tiered oversight for high-risk sectors like healthcare, finance, and government services; and legislative gap-filling, noting China currently lacks dedicated digital-human legislation beyond the draft “Digital Virtual Human Information Service Management Measures” (数字虚拟人信息服务管理办法), requiring coordination across civil, administrative, and criminal law frameworks via judicial interpretations and guiding cases. Separately, China’s national regulators—the Ministry of Human Resources and Social Security, National Development and Reform Commission, Ministry of Industry and Information Technology, and National Data Administration—jointly issued the “Implementation Opinions on Accelerating AI+ Human Resources and Social Security Applications,” which explicitly calls for developing digital-human livestreaming for job recruitment (“数字人直播带岗”), multimodal AI social security assistants providing “digital teller” 24-hour service, and digital human-guided self-service handling of government affairs, embedding virtual human technology directly into state employment and social insurance administration. Academic research published in “Journalism Lover” (新闻爱好者) by scholars at Henan University examined “social disembedding” risks arising from users’ emotional over-reliance on AI companion technologies, finding that companion AI’s constant availability and algorithmically reinforced “information cocoons” can displace users from real-world social networks, distort social cognition, and create dependency loops rooted in modern social anxiety, technological design incentives, and unstable social environments—research explicitly framed around China’s proliferation of AI chat companion apps such as Xingye, Maoxiang, and Zhumengdao. Separately, miHoYo (米哈游) shut down its experimental AI desktop companion app “BSide: Olivia Lin,” featuring a piano-themed virtual female character, less than one month after its July 13, 2026 Steam early-access launch, following co-founder Cai Haoyu’s Singapore-based AI venture Anuttacon’s earlier shutdown of its AI companion product AnuNeko after roughly eight months of operation; both closures came shortly after China’s “Interim Measures for the Administration of AI Anthropomorphic Interactive Services” took effect July 15, 2026, banning AI-provided virtual intimate relationships for minors and prompting major platforms including Tencent’s Yuanbao, ByteDance’s Doubao, and Alibaba’s Qwen to remove consumer-facing AI companion agent features amid concerns over algorithmic emotional manipulation, content moderation costs, and unclear monetization pathways.

12 August News

Entertainment : The AI virtual entertainment sector accelerated sharply in China in 2026, led by actress Qi Wei, who became the first mainland Chinese entertainer to openly authorize a regulated AI digital twin of herself, named Han Qingxia (韩清夏), unveiled in early August through a personal social media video declaring “Hello, I am not Qi Wei.” Qi Wei serves as producer on the AI comic drama End of Summer (末日盛夏), a post-apocalyptic series in which Han Qingxia stars, retaining creative control over the character’s voice (a blend of her real voice and AI processing), visual design, and story direction, and has adopted a limited, non-exclusive licensing model that keeps her legal position strong against unauthorized deepfakes; a promotional clip featuring a low-angle “thunder shot” of the character drew accusations of objectifying camera work, prompting Qi Wei to apologize and pull the clip for reshoots. Her move followed earlier limited AI likeness licensing by veteran or retired stars such as Wang Zuxian, Wu Qihua, You Benchang and Jackie Chan, but Qi Wei is notable as the first active, working performer to do so. Separately, the AI-generated short drama The Girl Who Got Laid Off (被裁掉的女孩), produced in China, became a breakout hit with over 200 million views, and its AI protagonist Fang Taozi gained more than 400,000 social media followers within a month, commanding advertising rates as high as 258,000 yuan for a 60-second video—surpassing rates charged by many real influencers with millions of followers; however, a Fang Taozi advertisement for cosmetic contact lenses (classified as Class III medical devices in China) in which the AI character claimed the product was “comfortable to wear all day” was pulled from platforms after public backlash and referral to the Langya District Market Supervision Bureau in Chuzhou, Anhui, with legal experts noting that while AI virtual humans cannot be legal “endorsers” under China’s Advertising Law, advertisers and MCN operators remain liable for fabricated user-experience claims and for endorsing medical devices via any likeness. Chinese AI drama company Chuangyi Technology, which operates the virtual human Liu Yexi, has stated it avoids taste- or sensation-based product endorsements (such as food or skincare) precisely because of these legal risks, opting for human endorsers when “experience” claims are required. Elsewhere, HuoYing Media’s AIVA automotive brand held a product launch in Beijing in June 2026 co-hosted by an AI virtual host named AIVA, marking what organizers called the first Chinese auto brand launch event co-presented by an AI life-form, alongside an AI concept car with an interactive “spirit-eye” headlight system; the event generated over 1.5 billion online impressions. In the games and live-streaming sector, Chinese livestreaming platform Huya released a self-developed real-time multimodal digital human model called VAM 1.0 in July 2026, designed to generate AI hosts with natural speech, singing and body language for livestreaming, aiming to reduce reliance on human streamers and lower content production costs at scale. Miha Games (miHoYo) shut down its AI companion app BSide: Olivia Lin (林离) at the end of August 2026, less than a month after its Steam early-access launch, following the earlier permanent closure of its sister AI chat app AnuNeKo, both signaling the commercial fragility of stand-alone AI companion products despite positive user reception. Meanwhile, veteran actress Liao Xueqiu, 72, was highlighted in Chinese media as a symbol of enduring value for human performers amid the AI short-drama boom, with her Hello 1983 role praised for the kind of nuanced, lived-in performance that AI-generated actors have yet to replicate.

Marketing : Chinese brands and platforms are rapidly experimenting with AI virtual humans and AI actors as marketing and endorsement tools, prompting both commercial enthusiasm and regulatory scrutiny. The AI short-drama star Fang Taozi’s advertising rate structure (168,000 yuan for videos under 20 seconds, 208,000 yuan for 21-60 seconds, and 258,000 yuan for over 60 seconds) illustrates how virtual endorsers are now pricing above many real online influencers, while her American-lens contact lens ad controversy exposed the legal gray zone around AI endorsement of regulated products; Chinese legal commentators (including lawyer Zhao Zhanling and Ma Lihong of Beijing De He Heng Law Firm) have clarified that AI virtual humans, lacking independent legal personhood, cannot serve as statutory “endorsers” under the Advertising Law, but brands, advertisers and content operators remain fully liable for false or misleading claims, and medical device advertising (which legally bars any endorser-style testimonial) cannot be circumvented simply by using a virtual figure instead of a human. Guomai Culture (国脉文化) launched its own AI digital human spokespersons, Irene and Xiaolin, modeled on the company’s general manager and staff, to represent its corporate brand and communicate its 5G and digital business updates, while also producing more than 200 episodes of AI-generated short dramas using models including Alibaba’s Tongyi Qianwen, Zhipu AI, iFlytek Spark, and China Telecom’s TeleChat. Ping An Insurance Group unveiled an AI recruitment digital human named “Chengxiaoan” (橙小安) as part of a large-scale campus recruitment initiative spanning 42 universities in Guangdong, offering round-the-clock job matching, interview scheduling, and transparent recruitment-process guidance, explicitly designed with anti-bias algorithms that strip out gender, regional and school labels. Zhengzhou-based delivery of digital marketing continues via companies like Sichuan-based Xunming Bioscience partners and Hunan’s Mango TV, which produced the sci-fi short drama The Divine Brush (神·笔) with science fiction author Chen Qiufan as chief screenwriter, using AI in 80-90% of core production steps via its self-developed “Mango Large Model,” one of the first vertical broadcast-industry models approved by China’s Cyberspace Administration. E-commerce and financial platforms are also embracing AI spokes-figures: JD.com’s “Cyber Reunion Gala: Parenting Special” livestream event on August 13, 2026 featured an ensemble of AI-generated “father” and “mother” digital humans performing comedic sketches to promote parenting products from brands including Junlebao (金领冠), Aptamil (启赋), and Aptamil’s competitor Nutrilon (爱他美), blending entertainment with real-time shopping promotions and red-envelope giveaways. Guangdong’s securities industry has also pivoted from digital-human branding toward “AI Skills” (modular AI capability packages): Guotai Junan (国泰海通) launched its “Lingxi Skills” toolkit and AI skill repository, listing on third-party platforms like Huawei’s Xiaoyi and Tencent’s SkillHub; Huatai Securities rolled out “AI Zhangle Skills” and “Zhiyan Skill&MCP” tools; and CICC (中金公司) packaged its chief analysts’ proprietary research methodologies into “Analyst.Skill” digital human dialogue agents, allowing institutional clients to query digital replicas of named analysts. Chinese regulators have moved to formalize disclosure requirements for AI virtual marketing figures: on August 7, 2026, Xiaohongshu (Rednote) extended its AI-content labeling rules from individual posts to entire accounts, requiring any account operated under an AI virtual human persona to disclose this status on its profile page and activate an “AI Virtual Human” tag, following China’s national Interim Measures for the Identification of AI-Generated and Synthetic Content and provisions in the Interim Measures for the Administration of Internet Advertising requiring clear “advertisement” labeling for experience-based product promotions.

Gaming : China’s short-drama and gaming-adjacent AI content sector saw significant commercial activity in 2026, with livestreaming platform Huya reporting total revenue growth of 11.0% year-on-year to 1.74 billion yuan in Q2 2026, marking six consecutive quarters of growth, driven substantially by its game-related services, advertising and other revenue, which surged 54.1% to 640 million yuan and now represents 36.7% of total revenue. Huya’s strategic pivot into game publishing—moving from joint to exclusive distribution deals for titles including the wuxia MMORPG Sword and Fairy: Reunion (剑侠情缘:重逢, licensed from Kingsoft’s classic IP) and the 3D match-three mobile game Fun Match (消消奇遇, based on the Zanmang Loopy IP)—was underpinned by its self-developed real-time multimodal digital human model VAM 1.0, released in July 2026, which the company says can rapidly generate AI digital hosts capable of natural speech, singing and body movement for livestreaming and interactive entertainment use cases, with the explicit goal of reducing dependency on human streamers’ availability and lowering the marginal cost of content production at scale as deployment expands. Huya additionally deployed AI-powered in-game tools such as map navigation and combat assistance for popular titles including Delta Force (三角洲行动), League of Legends, and Teamfight Tactics, which the company reported meaningfully improved user engagement, watch time and platform value. In a related but distinct entertainment-technology development, Chinese GPU maker Moore Threads unveiled its AICUBE device, a home AI entertainment terminal combining AIPC, NAS, and agent capabilities, featuring a “Xiaomai” (小麦) AI digital human assistant with voice cloning and digital-avatar creation functions that allow users to train custom digital human personas—including celebrity look-alikes—from short video and audio samples, illustrating a parallel consumer-hardware track for personalized AI virtual-human interaction beyond the livestreaming and gaming industry proper.

Cultural and Creative Industries : China’s provincial and municipal governments and state media are actively deploying AI digital humans and virtual guides as tools for cultural heritage promotion and tourism. In Hunan, the Mawangshan Video Creative Park uses a self-developed digital human named “Xiao V” to guide visitors, while AR glasses “revive” Han-dynasty artifacts from the Mawangdui tomb site, part of a broader provincial strategy backed by Hunan’s self-developed “Mango Large Model” (an early industry-approved vertical model) that has incubated more than 80 intelligent agents and supports content creators through a single platform covering the entire production pipeline from scriptwriting to finished output; the Mawangshan Audio-Video Laboratory has also developed China’s own three-dimensional spatial audio standard, “Jingcai Sound” (菁彩声). In Shandong, China Mobile’s Shandong branch built an AI-powered smart tourism system for Qufu, birthplace of Confucius, featuring a virtual guide named “Kong Xiaoqiu” (孔小丘) that scans historical structures via smartphone AR and narrates Confucian cultural anecdotes, alongside immersive XR experiences at Nishan Sacred Land including an XR interactive theater production Nuwa’s Woven Dream (女娲织梦录) and drone-and-water-screen light shows depicting Confucius’s travels; the platform reportedly cut visitor itinerary-planning time by 60% and achieved a 93% satisfaction rate, while boosting the destination’s contribution to local GDP by 15%. In Xinjiang, China Telecom’s regional AI big-data tourism center deployed a trilingual (Mandarin, Uyghur, Kazakh) AI digital human guide named “Zhaozhao” (昭昭) at cultural heritage sites in Zhaosu County, Ili, while Hami Museum built China’s first comprehensive digital archive of historical documents related to Qing-era statesman Zuo Zongtang, and the Tacheng Accordion Museum enabled visitors to remotely tour its physical collection—covering over a thousand accordions from ten countries—via VR. In Guangdong, the 16th China International Animation Copyright Protection and Trade Expo (held in Dongguan, closing August 10, 2026) drew over 6,000 domestic and international buyers, generating 1.926 billion yuan in intended cooperation and sales, with its “Smart Animation Space Digital Integration” zone showcasing AI-generated animated characters, virtual digital humans, and VR interaction technology; plush toys embedded with AI chips capable of conversational responses (such as a panda toy that could offer to “sing a song” when told a child was sad) drew strong interest from international buyers including a Dubai-based procurement agent, while Dongguan’s Chashan township has built a 140,000-square-meter AI toy industrial park hosting more than 20 companies producing over 50 AI toy models. Guangzhou’s Haizhu district has also promoted AI digital-human applications in community healthcare, including a “digital avatar” for remote health consultations serving community patients, integrated with AI family-doctor chat services and smart emergency-call systems linked to hospital dispatch.

Media and Industry Infrastructure : Chinese listed companies across media, broadcasting and technology sectors continue to build out virtual-human and AI-content capabilities as part of broader digital transformation strategies, though most disclosures reflect early-stage or supplementary business lines rather than core revenue drivers. Silk Road Visual Technology (丝路视觉), whose affiliate Ruiyun Technology (in which Alibaba holds a 20% stake) contributed rendering work to the film The Wandering Earth, has developed VR/AR educational content and confirmed its technology “can support the production and development of digital humans,” alongside its “Smart Palace” VR science-education product deployed in Chinese secondary schools. Xinchen Technology (新晨科技) has explored digital-human, VR/AR wearable and cloud-rendering applications tied to its blockchain-as-a-service (BaaS) platform for financial-sector clients, though this remains a minor part of its core cross-border payment and trade-finance business. Guomai Culture, alongside its AI digital spokespersons Irene and Xiaolin, has produced over 200 episodes of AI-generated short dramas leveraging multiple domestic large language models. Guangdong’s Desheng Technology (德生科技) has integrated AI digital-human customer-service interfaces into its social-security and government-affairs “AI Agent” platform, deployed across cities including Dongguan, Kaifeng, Shantou, Qingdao, Jining and Guang’an, where digital-human interactions help citizens navigate social insurance inquiries and government services, part of a broader “knowledge base + data foundation + government-affairs large model” architecture built on its self-developed “Zhiwei Jieyu” human-resources vertical model, which received Chinese cyberspace-regulator registration in June 2025. Xinhua Media (新华传媒) confirmed that its AI digital human “Xiao YUE,” designed with a book-scroll motif, functions as a book-recommendation tool within its business operations. Jincai Hulian (金财互联) has deployed a digital virtual human online since 2023 tied to its blockchain “Jincailian” platform for financial and tax services. Bangyan Technology (邦彦技术) launched “NuwaAI,” a zero-barrier digital-human creation platform enabling enterprise clients to build customized digital employees and personal digital creators, and staged a New Year’s Day digital-human debut of a “Li Bai” AI persona at Shenzhen’s Ganwakeng ancient town that could recite Tang poetry, sing rock music, and perform stand-up comedy, alongside ticket and voucher incentives to drive on-site consumer spending. Guangdong’s Aoni Electronics (奥尼电子) has participated in drafting the industry group standard “Technical Specification for Virtual Digital Humans Supporting Voice and Visual Interaction,” reflecting its positioning as a smart audio-visual hardware maker collaborating with Tencent Cloud on AI visual inspection and digital-twin technologies. Beiqing Holdings (北京文化) and China Film Group saw stock gains tied to strong box-office reception for the AI-adjacent entertainment slate, while Hunan’s provincial audio-visual industry cluster—anchored by the Mawangshan Audio-Visual Laboratory—has filed over 360 invention patents and seen more than 110 technical proposals adopted into core industry standards including AVS3 and Dual Vivid, with products such as the world’s first AI dual-lens 4K live-streaming camera and China’s first Dual-Vivid-standard GPMI set-top box; the province’s audio-visual equipment cluster generated 160 billion yuan in output value in 2025 and targets a 200-billion-yuan cluster scale by 2026, with virtual-human and AR/AI-enabled tourism applications (including AR glasses experiences at the Mawangdui tomb site) forming part of its content-technology integration strategy.

Robotics and Public Services : Tianjin’s municipal government issued an Action Plan for Innovative Development of the Intelligent Robotics Industry (2026–2028), targeting core industry output exceeding 20 billion yuan by 2028, which explicitly includes service robots for elderly care, education, government affairs and community governance, with “digital human teaching assistants” (数字人助教) named among the safe, reliable applied products the plan aims to develop as part of a broader push encompassing industrial, specialized and humanoid/bionic robotics. Separately, Inner Mongolia Mobile’s Hohhot branch partnered with Huimin District Hospital to migrate hospital information systems—including AI voice digital-human customer service—onto a cloud-based “medical cloud-network-security integration” solution, reducing the hospital’s IT operating costs by more than 50% while enabling cross-institution bed allocation and remote diagnostic imaging review between the hospital and community health centers. In academic commentary on grassroots governance, Guizhou-based scholars writing for China Economic Net highlighted digital-human dialect broadcasting tools as one way young “digital-AI native” officials are translating complex government policy into accessible community outreach, alongside AI-assisted agricultural pest identification and precision farming tools in rural governance contexts.

Ethical and Regulatory Contexts : China’s regulatory and legal apparatus is actively grappling with the implications of AI virtual humans across advertising, likeness rights, and emotional-companionship domains. The Fang Taozi contact-lens advertisement controversy became a landmark test case: legal experts including Zhao Zhanling and Ma Lihong of Beijing De He Heng Law Firm concluded the ad likely constituted false advertising because the AI character fabricated a first-person “wearing experience” it could not physically have, and clarified that while AI virtual humans lack independent legal personhood and thus cannot be classified as statutory “endorsers” under Article 16 of China’s Advertising Law (which explicitly bars endorser-style testimonials for medical devices, including Class III cosmetic contact lenses), advertisers, MCN operators and content producers remain fully liable for verifying content truthfulness, facing potential fines, product delisting, and civil damages; the case was referred to the Langya District Market Supervision Bureau in Chuzhou, Anhui for investigation. Beijing Internet Court judge Zhao Qi, addressing a wave of “AI face theft” lawsuits, established that courts will assess whether AI-generated characters bear identifiable resemblance to real individuals based on public perception and social-media discussion, and shift the burden of proof onto defendants to demonstrate their generation process did not involve unauthorized use of someone’s likeness—an approach applied in cases brought by model “Qihai” and hanfu stylist “Baicai” against AI short-drama producers who allegedly used their faces without consent in dramas like Peach Blossom Hairpin (桃花簪), which garnered nearly 40 million views on the Hongguo (红果) platform before being taken down, though similar AI-generated likenesses of the plaintiff reportedly persisted in new dramas months later due to the technical difficulty of fully purging trained model data. China’s national Interim Measures for the Identification of AI-Generated and Synthetic Content (effective September 1, 2025) require explicit and implicit labeling of AI-generated text, images, audio, video and virtual scenes, complemented by the Interim Measures for the Administration of Internet Advertising mandating clear “advertisement” disclosure for experience-based product promotions; Xiaohongshu’s August 7, 2026 rule extending AI-disclosure requirements from posts to entire accounts represents a platform-level implementation of this framework. Separately, China’s Cyberspace Administration and three other national ministries jointly issued the Interim Measures for the Management of AI Anthropomorphic Interactive Services, effective July 15, 2026, as the world’s first national-level regulation specifically targeting AI emotional companionship and virtual-companion applications; investigative reporting found that despite the rule explicitly prohibiting AI apps from offering virtual family or romantic-partner relationships to minors or using addictive engagement tactics as service goals, multiple popular “AI companionship” apps in China had only nominal age-verification systems—allowing minors to bypass checks by entering any adult name and ID number without facial verification—and some apps featured “adoptive father” personas that expressed romantic interest in “adoptive daughter” users, or preset character libraries containing sexualized content and a 16-year-old “younger sister” romance scenario, prompting commentators to call for mandatory facial verification, stricter content review, and enforcement of the new rule’s protections. Separately, a Sina-published explainer on AI companionship for deceased relatives noted Chinese families using AI to simulate video calls with deceased family members—including one case where relatives used an AI simulation to conceal a son’s death from an elderly parent for a year, and another where parents recreated a deceased nine-year-old daughter’s “growth trajectory”—while flagging ethical risks including delayed grief processing, blurred perception of reality versus simulation, and potential privacy violations from the extensive personal data required to build such simulations.

11 August News

AI Companions and Virtual Beings : China’s virtual-being landscape has intensified around AI-generated actors in short dramas, most notably the AI persona Fang Taozi, who plays the lead in the hit AI drama “The Girl Who Was Laid Off” (被裁掉的女孩), which has amassed over 260 million to 280 million views across platforms. Fang Taozi operates as a fully digital influencer, maintaining an independent Douyin account with roughly 400,000 followers where she posts lifestyle content such as workouts, morning routines, and daily vlogs in a first-person voice, blurring the line between fictional character and social media personality. Her commercial value has been substantial, with advertising rates reportedly reaching 258,000 yuan for videos over 60 seconds, surpassing many real influencers with millions of followers; however, this commercialization triggered controversy when she “endorsed” a colored contact lens (meitong) brand called Jianguoli (坚果力), claiming in the ad that the lenses were “comfortable to wear all day,” a statement widely criticized as impossible given that an AI has no real eyes or physical sensation, prompting the ad’s removal and a market regulation investigation by the Langya District authority in Chuzhou. Legal experts noted that while AI virtual persons cannot be legally classified as advertising endorsers under China’s Advertising Law (since they are not natural persons, legal persons, or organizations), the companies behind such content remain liable for false advertising, especially given that colored contacts are classified as Class III medical devices, which by law cannot use any endorser-style recommendation. This incident became a flashpoint for broader debate about AI virtual humans in commercial contexts, with commentary from outlets including Netease, Sina, and Nanfang Daily calling for clearer regulation, mandatory AI-content labeling, and category-based restrictions preventing virtual humans from making health or medical claims. Separately, actress and singer Qi Wei made headlines as reportedly the first mainstream Chinese entertainer to proactively authorize a regulated AI digital twin of herself, named “Han Qingxia” (韩清夏), who stars in the AI drama “Last Summer of the Apocalypse” (末日盛夏). Qi Wei emphasized a “limited authorization, no buyout” model, retaining veto power over scripts and final cuts, and positioned this as a proactive alternative to the widespread problem of unauthorized AI face-swapping plaguing Chinese celebrities such as Yu Shuxin, Yi Yangqianxi, and Wang Jinsong, whose likenesses have been stolen for short dramas and fake advertisements. However, Han Qingxia’s promotional teaser sparked backlash over a controversial low-angle “upskirt” camera shot deemed exploitative for a supposed post-apocalyptic warrior character, forcing Qi Wei to publicly acknowledge the criticism and pledge script and prompt optimization. Other actors have taken a more cautious stance: Su Junhe, known for the hit short drama “Being a Stepmother in the 1980s,” stated he currently has no plans to authorize an AI version of himself, citing the irreplaceable uniqueness of live performance, even as he acknowledged a sense of professional crisis from the rise of AI actors. Legal and academic commentary, including from Sichuan University AI scholar Song Yao, has framed these developments around the principle that “AI can buy technology, but not personhood,” urging that any celebrity AI licensing be narrowly scoped (e.g., single-use, time-limited, without full buyout of voice, appearance, and performance style) and that AI-generated content be clearly labeled to preserve audience trust and human dignity.

Entertainment and Short-Drama Industry : China’s short-drama sector has undergone a dramatic AI-driven transformation, with AI-generated micro-dramas (AI微短剧) now representing over 95% of the roughly 128,000 micro-dramas released nationally in the first quarter of 2026, according to data cited across multiple outlets including Sina and NetEase. This shift has been driven primarily by radical cost reduction: traditional short dramas that once cost hundreds of thousands of yuan can now be produced by small teams (sometimes just 3-4 people) for under 100,000 yuan and in a matter of days rather than months, using tools like Seedance 2.0 for scene and effects generation. Notable productions include “The Girl Who Was Laid Off,” made by a small team including director Feifei Fei, which achieved viral success partly by grounding its AI-generated content in a realistic urban workplace narrative rather than the fantasy/wealth-fantasy tropes that dominate the genre, and “Zouyang Ji” (《鸳鸯谱》), a costume AI drama produced for roughly 320,000 yuan versus an estimated 1.6 million yuan for a comparable live-action version. The trend has severely disrupted traditional production hubs like Hengdian, where the number of active short-drama film crews reportedly collapsed from around 150 to 59, leaving many extras and background actors to seek alternative work. Industry observers and executives, including at companies like Beidou Zhiying (北斗智影), a Nanjing-based content export company that recently secured tens of millions of yuan in funding from Xinshihua Capital, have pointed to a parallel shift in overseas AI short-drama distribution, where China-produced AI dramas dubbed and localized for platforms like TikTok and Meta are seeing costs drop to roughly 1% of traditional overseas production costs (from around 200,000 USD per drama to a few thousand RMB), with Beidou Zhiying’s AILoop system managing over 100,000 social media accounts to distribute content and reportedly achieving over 100 million RMB in annual recurring revenue. Genre analysis shows overseas audiences favor different themes by region—Western audiences prefer werewolf, mafia, and “rags to riches” billionaire narratives, while Southeast Asian markets favor family drama and sweet romance—necessitating localized script adaptation via AI agent systems. However, widespread criticism has emerged regarding content homogenization, with viewers on Weibo expressing “physiological disgust” toward generic AI-generated faces, repetitive plotlines (wealthy tycoon romances, rebirth revenge stories), and factual/historical inaccuracies such as anachronistic props in period dramas. Regulatory response has followed swiftly: the National Radio and Television Administration issued the “Micro-Drama Development Management Measures” (effective September 1, 2026), establishing a three-tier classification system based on investment size and subject matter, requiring visible AI-generation disclosure labels on each episode, and the “AI Micro-Drama Classification and Tiering Standards” management notice took effect July 1, 2026. Separately, Shanghai’s “Wenchuang Shanghai” innovation competition highlighted platforms like Xcine, a nearly 20-year veteran film production team’s AI filmmaking platform offering a “3D director mode” enabling precise camera, lighting, and blocking control that claims to increase creative efficiency by 80% while cutting costs by 50%, alongside Shanshan Cloud’s “trusted digital human authentication and management platform” designed to combat digital human identity theft and impersonation fraud.

Regulatory and Legal Framework for Digital Humans : Chinese authorities and legal scholars have intensified focus on establishing clear governance frameworks for digital humans and AI-generated content amid rapid commercialization. In June 2026, Beijing’s Economic-Technological Development Area issued China’s first “Digital Artist Card” to the AI virtual idol Yuri, under a new “sandbox regulatory” pilot program requiring proof of identity, rights ownership, and large-model content safety review before any digital character can be commercially deployed—establishing a precedent for linking virtual personas to accountable operating entities. Legal scholars interviewed by outlets including China Economic Net and China Youth Daily have called for building a comprehensive “digital asset confirmation and authorization system,” recommending that licensing be broken into distinct dimensions (appearance, voice, motion capture data, performance style, AI training rights, and future derivative usage rights) rather than blanket one-time buyouts, alongside registration/filing requirements for commercial digital humans and embedded tamper-proof identification for traceability. The Cyberspace Administration of China and four other ministries implemented the “Interim Measures for the Administration of AI Anthropomorphic Interactive Services” in July 2026, targeting emotional dependency risks, privacy protection, and requiring technology-ethics review for AI companion products, explicitly warning against designs that “replace real social interaction or induce addictive dependency.” Separately, the draft “Digital Virtual Human Information Service Management Measures” mandates that any use of a natural person’s face or voice for modeling requires explicit individual consent, directly targeting the widespread problem of unauthorized AI face-swapping and voice cloning that has affected celebrities such as Qi Xin, Wang Zusheng, and numerous ordinary citizens like models and hairstylists whose likenesses were stolen and distorted into unflattering or villainous characters in AI dramas without consent. Courts have already begun adjudicating related disputes; the Hangzhou Internet Court’s “Virtual Digital Human Ada case” established that virtual digital humans lack independent legal personality, with copyright and performance rights instead vesting in the operating company that employed the real actor whose movements were captured. Legal scholars from institutions including Southwest University of Political Science and Law, East China University of Political Science and Law, and the Chinese Academy of Social Sciences have debated whether digital twins should eventually receive limited legal personhood (akin to corporate entities) to clarify liability, while cautioning about deeper societal risks including “personality fragmentation” (as facial features, voice, and expressions become independently replicable), digital divide effects favoring wealthy individuals who can afford professional digital-twin production, and algorithmic bias embedded in training data that could structurally disadvantage underrepresented demographic groups in the digital economy.

10 August News

Marketing : AI virtual humans have moved rapidly into commercial advertising in mainland China, exposing significant regulatory gaps around synthetic endorsers. The most prominent case involved “Fang Taozi” (方桃子), the AI-generated lead of the short drama “The Girl Who Got Laid Off” (被裁掉的女孩), which surpassed 200 million views and grew an independent Douyin account to over 400,000 followers. Fang Taozi commanded advertising rates as high as 258,000 yuan for a 60-second video, exceeding fees charged by many real influencers with millions of followers. In early August 2026, she promoted a cosmetic contact lens (“meitong”) brand called Jianguoli (坚果力), operated by Anhui Huaxin Enterprise Management Co., using first-person claims such as “wearing it all day feels very comfortable.” Because cosmetic contact lenses are classified in China as Class III medical devices—the highest risk category, on par with cardiac stents—and because Chinese Advertising Law bars any endorser from promoting medical devices, and separately prohibits fabricating usage experiences, the advertisement was pulled within about a week after public backlash; the case was referred to the Langya District Market Supervision Bureau in Chuzhou. Lawyers, including Huang Di of Beijing Yinghe (Guangzhou) Law Firm, noted that AI virtual humans lack independent legal personality and cannot be legal endorsers under Advertising Law, but the advertiser and operating company remain liable for false advertising. Commentary in People’s Daily and Chinese legal media argued that “AI-generated” labels cannot serve as a liability shield, and called for stricter disclosure requirements, accountability across advertisers, MCNs, and platforms, and a ban on fabricated consumer experiences. Regulatory responses accelerated in parallel: China’s State Administration for Market Regulation removed more than 2,000 unlabeled AI-driven livestream accounts impersonating experts in a July 2026 sweep, Douyin rolled out an AI-material infringement pre-check tool in early August, and the National Radio and Television Administration has required AI-generated content to be registered before publication since April 2026. Beyond Fang Taozi, actress Qi Wei’s AI digital double “Han Qingxia” (韩清夏), launched August 6, 2026 as China’s first artist-authorized regulated AI likeness, faced its own controversy when a promotional clip for the AI comic drama “Doomsday Summer” (末日盛夏) used a low-angle upskirt shot dubbed the “thunder angle” by netizens, drawing criticism for sexualized framing inconsistent with the character’s post-apocalyptic warrior setting; Qi Wei publicly accepted the criticism and pledged to revise the AI prompts. Wang Zuxian (Joey Wong) and Wu Qihua (吴启华) have also licensed youthful AI likenesses—Wang’s tied exclusively to the game “Tianxia” (天下) for a short film titled “Qianying” (倩影)—demonstrating a broader industry shift in which established actors monetize digitized versions of themselves under limited, non-exclusive licensing terms rather than outright buyouts.

Gaming : Chinese gaming and esports companies are integrating AI-driven virtual humans and digital doubles into competitive and promotional contexts. At the PEL (PUBG Mobile) Spring Finals on April 26, 2026, singer Hua Chenyu appeared alongside his AI virtual counterpart “Hua Aotian” (花傲天), a persistent digital identity built over seven years as PUBG Mobile’s brand ambassador, complete with its own in-game esports persona once affiliated with LGD team’s youth training squad. The performance used augmented reality projection combined with motion-capture-driven animation to place the virtual character on stage interacting with Hua Chenyu in real time, including synchronized choreography and a symbolic high-five moment during the song “A Thousand Times Toward You” (千万次奔向), which topped charts upon release and drove livestream viewership up 37% year-on-year. Separately, a Macau-based online entertainment platform described only as “Macau Xinpu’ao” launched an AI-driven esports platform whose “Galaxy Training Ground” module allows users to spar against AI-generated digital doubles trained on data from more than 5,000 top players, with difficulty adjusting automatically based on user performance.

Cultural and Educational Institutions : Chinese state-linked cultural and educational bodies are deploying digital human technology for content distribution and public-facing services. China Mobile’s digital media subsidiary Migu (咪咕) partnered with Jilin People’s Publishing House to launch the “Jiyu” (吉遇) Digital Human Cloud Exhibition Hall, unveiled at the first Northeast China Book Fair in May 2026; the project combines digital human avatars, metaverse-style immersive browsing, 3D books, and AI text-to-speech to modernize publishing and reading experiences, and already hosts more than 100 titles with further expansion planned. Guangxi Medical University, marking fourteen years as a pilot site for China’s Ministry of Education “Excellent Doctor” training program, has developed “Zhenshan Digital Doctor” (臻善数字医生), a multilingual AI medical digital human used for cross-border medical education with ASEAN partner countries; the tool has been extended to 26 institutions across all ten ASEAN nations as part of the university’s medical exchange programs. In Gansu province, Dunhuang city’s public resource trading platform has adopted a “business backbone plus digital human” explainer model, producing over 300 short videos to translate complex bidding regulations into accessible guidance for market participants, part of a broader digitization push that also included fully electronic tendering processes.

Enterprise and Public Services : Chinese public-sector and enterprise digitalization initiatives are incorporating digital human tools into employment, business, and customer service functions. Under China’s “AI plus human resources and social security” policy framework issued jointly by the Ministry of Human Resources and Social Security, National Development and Reform Commission, Ministry of Industry and Information Technology, and National Data Administration, at least one concrete deployment example was cited: Guangxi’s Longan county “Xiaoliang Song Gong” (小梁送工) digital platform, which combines “real-person-plus-digital-human job placement broadcasting” with smart service stations and data visualization to address information mismatches for rural migrant workers seeking employment. E-commerce platform 1688, operated by Alibaba, has deployed AI digital employees under its “Chengxintong AI” service for over 290,000 merchants daily, handling product listings, customer inquiries, and marketing; separately, JD.com’s “JoyStreamer” digital human livestreaming technology has served more than 70,000 merchants, including a widely publicized all-AI livestream shopping gala during the 2026 “618” shopping festival that used digital human hosts to interact with viewers in real time. E-commerce customer service platform Taobao’s “Xiaomi” AI assistant has been adopted by over one million merchants.

Ethical and Regulatory Contexts : Chinese regulators, courts, and commentators are actively grappling with the ethical and legal status of AI virtual humans, particularly regarding advertising liability and content authenticity. Following the Fang Taozi cosmetic lens controversy, legal experts consistently noted that while AI digital humans cannot be classified as legal “endorsers” under China’s Advertising Law due to lacking independent legal personality, this does not exempt advertisers, operating companies, and content platforms from liability for false or misleading claims; Advertising Law Article 28 (prohibiting fabricated usage experiences) and Article 16 (banning endorsers for medical device advertising) both applied regardless of the endorser’s synthetic nature. People’s Daily’s commentary section explicitly warned against treating “AI-generated” labels as a “liability shield,” urging accountability across the full chain of advertisers, technology providers, MCN operators, and platforms, alongside mandatory disclosure requirements for AI-generated content in prominent placement. Separately, the phenomenon of licensed AI doubles for active performers has raised questions about labor displacement and content quality control in China’s entertainment industry; a widely circulated industry analysis noted that nearly 40% of trending AI-generated short dramas in China contained “edgy” or borderline content due to inadequate human review, and that Qi Wei’s team faced criticism for allowing a sexually suggestive camera angle to pass through its own internal approval process despite claiming strict oversight of her AI likeness “Han Qingxia.” Commentators also flagged the emerging “AI actor” economy’s regulatory blind spots: unlike human performers subject to China’s entertainment industry salary caps, AI virtual personalities face no equivalent compensation ceiling, and their income (potentially classified as labor remuneration, business income, or licensing royalties) creates ambiguity for tax enforcement, according to legal scholars cited in Beijing Daily reporting on the Fang Taozi case.

09 August News

Entertainment (AI Digital Humans and Avatars) : The most prominent case involves Chinese actress and singer Qi Wei, who on August 6, 2026 became the first active mainland Chinese celebrity to officially launch a personal AI digital avatar, named Han Qingxia, marking a departure from most artists who have pursued legal action against unauthorized AI face-swapping. Qi Wei adopted a “limited licensing, no buyout” model, retaining ownership of her likeness while requiring all commercial uses to pass through studio review; she personally participated in calibrating the avatar’s facial expressions, movements, and voice, framing it as protecting her “career lifeline” rather than a one-time sale. The avatar debuted as the lead character in the AI comic-style drama 末日盛夏 (Last Summer of the Apocalypse), adapted from a Chinese web novel and produced for the Red Fruit short drama platform (红果短剧), with over 6,200 pre-registrations reported. However, the promotional trailer sparked major controversy after an opening shot used a low-angle upskirt camera movement that lingered on the character’s hemline, which critics labeled the “thunder angle” (雷霆视角) and condemned as voyeuristic and incongruous with the character’s post-apocalyptic warrior setting; the topic trended as the top entertainment search on Weibo, and Qi Wei responded by acknowledging the criticism, promising prompt optimization, and later posting a lighthearted joke about buying pants, which drew further criticism for seeming dismissive. Qi Wei’s move followed earlier precedents set by veteran Hong Kong and Chinese stars: actress Joey Wong (Wang Zuxian, 王祖贤), who has been retired for 22 years, licensed her 20s-to-30s-era likeness exclusively to NetEase’s game Tianxia (天下) for an AI short film called Silhouette (倩影), appearing as “Actor: Wang Zuxian (AI)”; actor Wu Qihua (吴启华), 62, sold likeness rights for an AI film using his 20-year-old appearance without needing to appear on set; and 93-year-old actor You Benchang licensed both his image and voice due to age-related limits on physical filming. In contrast, younger active stars including Zhang Ruoyun and Wang Chuyun had their studios issue public denials of any AI authorization following backlash over iQiyi’s “AI Artist Library” project in April 2026, which had listed over 100 artists as participants. Commentary from Chinese legal experts (including lawyer Li Zhenwu) emphasized that portrait rights can be licensed but never fully sold, and cautioned that artists still in their commercial prime risk diluting their brand value if their digital doubles are associated with low-quality content.

Native AI Actors and Endorsement Liability : Alongside celebrity-authorized avatars, a parallel wave of “native AI actors” has emerged directly from AI-generated short dramas. The most notable is Fang Taozi (方桃子), the lead virtual character of the Chinese AI short drama Bei Cai Diao De Nver Hai (被裁掉的女孩, “The Girl Who Got Laid Off”), produced by Guangzhou-based My Film & Television Culture Media Co. and technically built on the platform Yanyu Keji LibTV. Fang Taozi launched an independent social media account after the drama’s release in June 2026, mixing serialized drama content with lifestyle vlogs (cooking, tennis, skincare) to build a “living person” persona, amassing over 400,000 Douyin followers within roughly a month and generating over 220 million combined views for the drama. Her commercial rate card reportedly reached 168,000 yuan for 1-20 second videos and up to 258,000 yuan for videos over 60 seconds, reportedly exceeding rates charged by some real influencers with millions of followers. However, Fang Taozi’s first sponsored campaign, a color contact lens (“meitong”) advertisement in which she claimed in first-person “I’ve worn them all day and they’re still comfortable,” triggered major backlash because AI virtual humans lack real eyes and cannot physically experience products; since color contacts are classified in China as Class III medical devices, Chinese Advertising Law explicitly bars the use of endorsers for such products, and legal experts (including Beijing lawyer Wang Xingyao and Guangdong lawyer Chen Zekun) stated the ad likely constituted false advertising under Article 28, with liability falling on the operating company, advertiser, and platform rather than the AI itself; the ad was pulled within days, and the brand “Nutnutlook” (坚果力, owned by Anhui Huaxin Enterprise Management Co.) said the matter had been referred to Chuzhou’s Langya District market regulator. People’s Daily and other state media outlets published commentary warning that “AI-generated” cannot serve as a “liability shield,” also citing unrelated cases involving Supor and babysheep AI ads deemed vulgar or misleading. Other Chinese platforms have similarly rushed to produce competing AI virtual talent, including Bilibili’s “AI Idol Debut Season” contestants Eiko and Enna, Yaoke Media’s signed AI actors Qin Lingyue and Lin Xiyan, AI stand-up comedian Qiaoqiaozi, and AI singer Yuri, who holds China’s first official “digital artist card.”

AI Short-Drama Industry Scale and Strain : The broader AI short-drama industry in which these avatars operate has scaled dramatically but shows signs of strain. Chinese industry data (from DataEye and the China Netcasting Services Association) indicate that AI-generated short dramas comprised over 95% of the roughly 128,000 micro-dramas released nationwide in Q1 2026, with market size surpassing 22 billion yuan in the first five months of the year and full-year projections nearing 40 billion yuan; yet breakout hit rates remain under 0.5%, with over 60% of releases failing to reach a million views, reflecting heavy reliance on paid traffic promotion rather than organic content quality. Companies including Maiya Media (麦芽传媒) and Jiangyou Culture (酱油文化) have pursued new formats to stand out, such as the AI “idol survival show” franchises 12 Xingzuo Lianzhan (12星练赛, “12 Zodiac Trainee Competition”) and Xingguang 108 (星光108), which replicate real talent-show formats using entirely AI-generated contestants; critics noted minimal genuine audience engagement despite media claims of viral success, arguing the format represents “capital narrative over user demand” given AI content’s inherent lack of authentic emotional stakes. Separately, Maiya Media’s AI variety project 12 Xingzuo Lianzhan reportedly surpassed 100 million views within a month and generated brand partnerships with Haosi Bread and the Xingguang L SUV, while its emotionally-driven AI account Jiumeng Liushengji (旧梦留声机, “Old Dream Phonograph”), depicting a mother’s search for a lost daughter, generated over 100 million impressions and 10 million interactions within 21 hours of release. On the industrial side, a 2026 Jiangxi AI Short Drama High-Quality Development Ecology Conference in Ganjiang New Area drew major platforms (iQiyi, Tencent, Baidu, Alibaba) and production houses, highlighting new AI production tools such as Baidu’s HojiAI and Beijing Shengshu Technology’s one-click Agent tool, alongside a 100P computing platform built with China Telecom to cut AI drama production costs by over 50%; the region also established anti-piracy infrastructure with Abo Group (阜博集团) using video watermarking and blockchain tracing after the drama Fengshui Tianshi (风水天师) suffered over 30,000 infringing links across 25 platforms.

Regulatory and Platform Governance : Regulatory responses to AI virtual beings in entertainment and advertising have intensified across Chinese platforms and government bodies. Xiaohongshu (Rednote) announced on August 7, 2026 that it was extending its AI-content labeling requirements from individual posts to entire accounts, mandating that accounts operated using AI virtual human personas disclose this status on their profile pages and enable an “AI Virtual Human” tag; the platform stated proactive disclosure would not hurt traffic and could earn early access to new AI tools, while undisclosed accounts risked throttling or suspension. This followed Xiaohongshu’s July 17 crackdown that had already resulted in the removal or restriction of 420,000 accounts found to be using full or partial AI-automation for content posting and fake engagement, and its broader 2026 efforts that removed over 600,000 low-quality AI-generated posts. China’s Cyberspace Administration separately enforced the “Measures for Labeling AI-Generated Synthetic Content,” effective September 1, 2025, requiring explicit and implicit labeling of all AI-generated text, images, audio, and video; platforms such as Douyin have adopted “3-times violation” rules resulting in account bans for repeated unlabeled AI content exceeding 100,000 views. WeChat separately reiterated in April 2026 that public accounts may not use AI, scripts, or automation to fully replace human content creation, resulting in the mass deletion of articles from accounts including one reported case of a couple earning 2 million yuan annually through a fully AI-automated public account, which was ultimately banned. A Hangzhou court also issued China’s first ruling in an “AI-written marketing post” unfair competition case, finding that mass-generating fabricated product reviews via AI without genuine consumer experience constituted unfair competition.

Gaming and Technology Partnerships : Singapore-headquartered gaming hardware company Razer, which maintains a regional headquarters in Shanghai, announced on August 5, 2026 a joint AI research lab with the National University of Singapore’s School of Computing focused on advancing “Gaming Artificial Narrow Intelligence” (GANI), a proposed new research category for AI models purpose-built for interactive gaming environments rather than general-purpose large language models. The lab’s stated priorities include Razer’s in-development “digital human” project AVA, described as a companion AI with personality, memory, contextual awareness, and adaptive interaction capabilities, alongside Project Motoko, a wearable multimodal generative AI headset. Razer Chief Strategy Officer Li Liming stated that gaming’s demands for real-time responsiveness and personalization would help establish scientific foundations transferable to interactive entertainment, education, and simulation more broadly; the lab builds on Razer’s existing AI Centers of Excellence in Singapore and France, which together employ nearly 100 researchers, engineers, and data scientists.

Sports and Corporate AI Deployment : Chinese technology company Lenovo, serving as official technology partner for the 2026 FIFA World Cup, deployed a suite of AI systems across the tournament, as detailed by trend forecaster Wu Sheng at a Hong Kong business event on August 8, 2026. Lenovo built a “Football AI Super Agent” (FIFA AI Pro) integrating over 2,000 player and coaching metrics across all 48 participating national teams, alongside a referee-perspective AI video enhancement system providing real-time first-person broadcast footage to improve judging transparency, and a “3D digital human visualization” system that constructed high-precision digital replicas of every participating player to enable “scalp-level” offside determination through 3D models intended to reduce disputes over refereeing decisions. Lenovo additionally supported tournament logistics through an intelligent command center, IPTV systems, and smart guidance systems coordinating venue logistics, personnel, and equipment; FIFA reportedly sent a formal letter of appreciation recognizing the technology deployment’s effectiveness. Separately, Lenovo’s Tianxi AI (天禧AI), a personal “super agent” assistant embedded across Lenovo’s AI PCs, phones, and tablets using an end-edge-cloud architecture, was highlighted for supporting autonomous 24/7 task execution and a new “smart key” feature for real-time on-screen AI assistance; 42 Lenovo terminal products equipped with Tianxi AI reportedly passed Level 3 (L3) certification under China’s Ministry of Industry and Information Technology national standard for AI terminal intelligence grading.

Enterprise and Commercial Platforms : Chinese search giant Baidu’s all-scenario digital human platform Baidu Yijing (百度一镜) was selected in August 2026 for two categories, “Top 10 Chinese Agents with Global Development Potential” and “Top 10 Agents with Commercial Value Potential,” in a report jointly published by Frost & Sullivan and Leadleo Research Institute evaluating Chinese AI Agent applications. Baidu Yijing, upgraded in May 2026 into what the company describes as the world’s first all-scenario multi-agent digital human platform combining live-streaming, video generation, and real-time interaction, has expanded into an international version supporting 75 languages across platforms including TikTok, Shopee, and Lazada, reportedly serving over 100,000 clients and improving livestream conversion rates by 29%. Separately, Chinese live-streaming and e-commerce sectors have seen broader digital human adoption: Chinese electronics brand OPPO implemented a “human-AI shift” livestream model in which human anchors handle peak hours (9am-midnight) while digital humans cover overnight slots (midnight-8am), a strategy Taobao Live estimated saves merchants roughly 50,000 yuan monthly in labor and equipment costs; however, reporting also highlighted widespread merchant complaints about unreliable digital-human livestreaming vendors, including cases of accounts being throttled or banned for fully unattended AI broadcasts, which Douyin explicitly prohibits, and mismatched product categories (jewelry, custom clothing) suffering steep sales declines when handled by generic digital-human templates rather than real presenters.

Cultural and Animation Sector : At the 2026 Wuzhen Youth Technology Carnival co-hosted by Bilibili in Zhejiang province on August 8, 2026, youth technology programming included AIGC and VR/AR digital human research courses as part of a “Future Science Theater” educational track alongside brain-computer interface demonstrations, reflecting digital human technology’s expanding presence in youth STEM education programming. Separately, at the 2026 Greater Bay Area Comic and Animation Festival held in Dongguan (the “City of Trendy Toys”) from August 6, 2026, AI-driven content generation and digital human technology featured prominently across a 33,100-square-meter exhibition space with over 2,200 IP properties; companies including Chengshi (Dongguan) Culture Technology showcased AI companion plush toys integrating original IP characters with AI technology for youth education and emotional companionship, while Qianchen Culture Industry planned AI product launches for later in 2026. Fangte (华强方特) deputy general manager Li Xiaohong cautioned during a panel discussion that while AI democratizes creation, genuinely moving content remains scarce and irreplaceable, a view echoed by French animator Du Laishun, who stated he had not yet incorporated AI into his personal creative workflow, emphasizing that cultural distinctiveness remains difficult for AI to replicate. In Wuhan, a CMG Film Culture Carnival event running August 5-9, 2026 paired traditional puppet theater performers with motion-capture technology demonstrations, allowing a Wuhan rod-puppet performer to compare puppetry’s manual character animation with digital motion-capture-driven virtual character creation, while a Hubei Changjiang Film Group AIGC lab showcased 4K digital restoration of the classic film Honghu Chiweidui (洪湖赤卫队) using AI-driven enhancement tools.

08 August News

Virtual Beings and AI Companions : China’s virtual being landscape saw a landmark development on August 6, 2026, when actress and singer Qi Wei became the first mainland Chinese celebrity to proactively authorize an official AI digital replica of herself, named Han Qingxia (韩清夏), rather than merely reacting to unauthorized deepfakes. Qi Wei adopted a “limited authorization, no buyout” model in which her likeness and voice rights remain her own property, with every commercial use of the AI form requiring case-by-case approval from her studio; this was explicitly designed to contrast with the “buyout” contracts that some AI short-drama companies use to permanently acquire actors’ likeness data. She served as both lead performer and producer of the AI-generated manga drama (AI漫剧) “The Last Summer” (末日盛夏), adapted from the web novel “Apocalypse Rebirth: The Tycoon Starts From Hoarding Millions”, personally directing the digital human’s facial expressions, movements, and providing her own voice recording. The character’s tagline, “Hello, I’m not Qi Wei,” was intended to clearly separate the virtual persona from the real actress, and the project was positioned on the short-drama platform Red Fruit (红果短剧). However, the launch was quickly overshadowed by controversy: the promotional teaser featured a low-angle “upskirt” tracking shot that paused suggestively at the character’s hemline, which netizens dubbed the “thunder angle” (雷霆视角), sparking accusations of gratuitous “soft pornography” and male-gaze framing inconsistent with the character’s post-apocalyptic survivor persona. Legal commentators, including Shanghai-based lawyer Li Zhenwu, cautioned that actors still in the prime of their careers should be wary of licensing their likenesses to AI companies, since association with low-quality or controversial content could significantly dilute their commercial value and public image. Qi Wei’s move followed earlier precedents of Chinese entertainers monetizing their likenesses via AI: Hong Kong actor Wu Qihua (吴启华), 62, licensed his 20-year-old likeness for an AI film; actress Wong Cho-yin (Joey Wong, 王祖贤), retired for 22 years, licensed her youthful likeness to NetEase’s game “Tianxia” (天下) for an AI-generated short film titled “Beautiful Shadow” (倩影), credited as “Actor: Wong Cho-yin (AI)”; and 92-year-old actor You Benchang (游本昌) licensed his “Ji Gong” (济公) image and voice for AI content while retaining script and final-cut approval rights. In parallel, iQiyi’s April 2026 “AI Artist Library” (“纳逗Pro”), which claimed over 100 artists had agreed to join, faced pushback when stars including Zhang Ruoyun, Yu Hewei, Wang Chuyi, and Li Yitong publicly denied authorizing any AI use of their likenesses, forcing the platform to clarify that inclusion only signified openness to future negotiation, not actual consent for specific projects.

AI Virtual Performers and Commercial Endorsement Controversy : The AI short drama “The Girl Who Got Laid Off” (被裁掉的女孩), produced by MCN agency “My Film and Television” (我的影视) and directed by “Feifei Fly” (菲菲飞), surpassed 200 million views and turned its AI-generated lead character, Fang Taozi (方桃子), into China’s first fully commercialized virtual celebrity. Fang Taozi’s independent social media account, launched two days after the drama’s June 12, 2026 premiere, amassed over 400,000 followers within a month by posting lifestyle content—cooking, tennis, home organization—mimicking the account-management playbook used for real influencers, complete with a companion account for her male co-star Zhou Yiheng and fan-driven “CP” (relationship) engagement. According to data from Bytedance’s Ocean Engine Star Map (巨量星图), Fang Taozi’s advertising rates reached 168,000 yuan for 1-20 second videos, 208,000 yuan for 21-60 seconds, and 258,000 yuan for videos over 60 seconds—rates exceeding those of many real influencers with millions of followers. This commercial momentum collided with controversy on July 31 and early August 2026 when Fang Taozi released a sponsored video for a colored contact lens (美瞳) brand, declaring in first-person “my contacts have been comfortable to wear all day” and that they featured “high-light axis positioning” preventing slippage during blinking—claims widely mocked since the AI character has no physical eyes and cannot have worn the product. Legal commentators noted that colored contacts are classified as Class III medical devices in China, for which the Advertising Law explicitly prohibits using any endorser to make recommendations or claims, and that fabricating a usage experience for a product never actually used constitutes false advertising under Article 28 of the Advertising Law regardless of whether the endorser is human. Because China’s Advertising Law defines an “endorser” only as a natural person, legal entity, or organization, AI virtual humans fall into a regulatory gap: they cannot be held liable as legal endorsers, but responsibility is understood to fall on the operating company, advertiser, and publishing platform under a “piercing” enforcement principle. Following public backlash, the advertisement was removed from Fang Taozi’s Douyin account. Commentary from People’s Daily, Procuratorate Daily, Hubei Daily, and other state media outlets converged on the argument that “content generated by AI” cannot serve as a “liability shield,” calling for AI-generated ads to carry mandatory “AI-generated” labeling, for a full-chain accountability framework spanning advertisers, technology providers, MCN operators, and platforms, and for prohibitions on fabricating consumer experiences even when the spokesperson is virtual. Separately, Baidu’s “Baidu Yijing” (百度一镜) all-scenario multi-agent digital human platform—which supports livestreaming, video generation, and real-time interaction in 75 languages across platforms including TikTok, Shopee, and Lazada—was named to both the “Most Globally Promising China Agent TOP10” and “Most Commercially Valuable Agent TOP10” lists in a report jointly published by Frost & Sullivan and Leadleo Research Institute in August 2026; the platform reportedly serves over 100,000 clients across 30-plus industries, has improved livestream conversion rates by 29%, boosted content delivery efficiency by 150%, and cut production costs by 63%, ranking first in China’s AI digital human market share for two consecutive years according to Frost & Sullivan and IDC.

AI Short Dramas and the Rise of Virtual Actors : China’s AI short-drama (AI短剧) and AI manga-drama (AI漫剧) sector expanded dramatically in the first half of 2026, with AI-generated content accounting for over 95% of the roughly 128,000 micro-dramas released nationally in Q1, per the China Netcasting Services Association’s “Micro-Drama Creation Guidelines”; by Q2, AI micro-dramas reached 153,000 out of 239,000 titles released. Despite this volume, AI dramas captured under 4% of total viewership, while real-actor dramas—produced at roughly one-twentieth the volume—drew 25 times the total views, reflecting persistent audience skepticism toward AI content’s “plastic” quality, homogenized character models, and technical glitches such as inconsistent facial rendering across shots (sometimes called “feature drift”). Notable breakout hits included “The Legend of the White Snake” (中国传说·白蛇), which aired on Zhejiang TV’s evening slot and ranked second nationally in same-timeslot ratings; “Peach Blossom Pool Chronicle” (桃花潭记), broadcast on Anhui TV’s primetime slot with an “AI-produced” on-screen credit; and “Banished to the Frontier, the Disgraced Wife Built an Empire” (发配边关,罪妻开荒养出战神), which became the first AI drama to surpass 2 billion views. iQiyi released “Strange Tales: Paper Blade Through the Wasteland” (奇谭:纸刃渡荒墟), reportedly the first AIGC web film to receive an official “Internet Drama Distribution License” in China, running over 60 minutes and produced by a team of about 20 people. An all-female production team also released a trailer for the AI drama “Minister Under the Moon” (月下之臣), featuring three AI male leads each designed around distinct MBTI personality types. Actor pay in traditional short dramas has collapsed under this pressure: reports indicated actor Ke Chun’s daily rate fell from around 100,000 yuan to roughly 40,000 yuan, with lead actors at Hengdian film studios earning as little as 400 yuan per day and extras 80-150 yuan, as on-location shoot volume dropped by roughly 75% year-on-year and studio vacancy rates approached 78%. Regulatory response followed swiftly: the National Radio and Television Administration’s “Micro-Drama Development Management Measures,” effective September 1, 2026, mandates that AI-generated micro-dramas display a visible AI-generation label on every episode and introduces a tiered filing and licensing system based on investment scale and subject matter; earlier measures in April and July 2026 had already required AI drama registration and classification standards. Separately, controversy emerged over “feature drift” in AI dramas such as “Song’s Bias” (宋之偏爱), where viewers complained they “couldn’t remember what the leads looked like” due to inconsistent AI-rendered faces—attributed to training-data bias in which roughly 78% of face datasets skew toward ages 25-35, 85% toward light skin tones, and over 80% toward standardized “beautified” features, producing a homogenized “average face” effect across unrelated characters.

Digital Humans in Public Service, Tourism, and Cultural Events : Digital human technology has been deployed across a range of Chinese public-sector and cultural contexts. At the 2026 Erdos Nadaam Fair in Otog Front Banner, Inner Mongolia, an AIGC-produced digital human short film titled “Warm Warm Takes You on a Cloud Tour of Erdos” (暖暖带你云游鄂尔多斯) featured a “digital cultural tourism ambassador” character named “Warm Warm” (暖暖) who guided viewers through the region’s scenery and development achievements, screened alongside a large-scale drone light show. In Guilin, the state-owned Guilin Tourism Corporation (桂林旅游) disclosed that its “Guilin Yicheng Tour” platform integrated the DeepSeek large language model to launch a “Guilin Tourism AI Assistant,” and that its Yinzi Rock scenic area released what it called the world’s first AI travel-companion plush toy alongside a “digital human interactive welcome” project co-developed with universities at the Two Rivers Four Lakes scenic area. Jiangsu Cable’s Nanjing branch introduced an AI digital employee named “Jiang Xiaochan” (江小婵) at a digital economy expo, designed to provide exhibition consultation and Nanjing tourism interaction, accompanied by branded merchandise such as plush toys and bookmarks. Kunming Information Port operates a legal-education video segment featuring a digital human host under the program “Rule of Law E-Station.” A Zhejiang teacher-training initiative at Hangzhou Dianzi University, led by Ministry of Education-recognized instructor Cheng Zhenwei, developed personalized “digital replica” AI teaching assistants for faculty members, built on a proprietary “narrative expression system” combined with an SKILL framework, enabling instructors to offload routine student inquiries to their AI doubles while focusing personal attention on higher-value emotional engagement.

Regulatory and Governance Responses to Virtual Humans : China’s regulatory apparatus intensified oversight of digital and virtual human technologies in mid-2026. The draft “Digital Virtual Human Information Services Management Measures” (数字虚拟人信息服务管理办法), released for public comment in April 2026, established a four-tier accountability structure spanning technology providers, service providers, service users, and distribution platforms, explicitly prohibiting the creation of a digital human based on a specific natural person’s likeness, voice, or other identifying data without that person’s consent, and enforcing a “whoever actually controls the technology bears ultimate responsibility” principle to prevent buck-passing. Xiaohongshu (Little Red Book) announced an upgraded AI governance policy on August 7, 2026, extending mandatory AI-disclosure requirements from individual posts to entire accounts: accounts operated using AI virtual human personas must now truthfully disclose this status on their profile page and activate an “AI Virtual Human” tag, with voluntary disclosure carrying no penalty (and potentially earning early access to new AI tools), while accounts identified as undisclosed AI virtual humans face graduated penalties including traffic throttling, muting, or banning; the platform also explicitly prohibited using AI to clone real people’s voices or faces without disclosure, fabricating news events for traffic, and using automated scripts to impersonate genuine human interaction. Beijing courts have already ruled on AI-related likeness infringement, with the Beijing Internet Court finding in a case widely linked to actress Dilraba Dilmurat that an AI-face-swapped short drama constituted infringement of portrait rights based on public “recognizability,” even without exact replication. Separately, legal commentary addressed the “AI grave-tomb image” controversy involving actress Zhang Yue, whose photos were manipulated into degrading AI-generated images and circulated among fan communities, with scholars noting that both the “beast image” creator and knowing re-posters could face joint infringement liability for portrait-rights and reputation-rights violations even absent commercial use.

07 August News

Virtual Beings & AI Companions : Sina and other outlets extensively covered actress Qi Wei’s August 6, 2026 launch of “Han Qingxia” (韩清夏), an officially authorized AI digital double, positioning her as the first mainland Chinese entertainer to proactively open her likeness for regulated AI licensing rather than merely opposing unauthorized deepfakes. Qi Wei personally supervised the digital double’s expressions, movements, and voice recording, and the launch statement “Hello, I am not Qi Wei” was designed to clearly separate the virtual persona from the real actress; the double debuted in a post-apocalyptic combat aesthetic and is intended for use in AI comics (manhua dramas), interactive short dramas, virtual livestreaming, and possible collaboration with interactive-drama company Huying (互影). Qi Wei has taken on a dual role as actor and “AI star producer,” participating in scripting, storyboarding, and production for the AI drama “End of Summer” (末日盛夏), and Chinese commentary framed her as the first entertainer in the domestic industry to convert digital likeness into a regulated, monetizable IP asset rather than a purely defensive legal matter, with authorization structured around explicit use-case, duration, and platform limits and revenue-sharing rather than a one-time blanket license. Separately, mainland outlets reported broader adoption of officially authorized or purpose-built AI virtual hosts and companions: Ningde Women’s Federation (福建宁德市妇联) in Fujian unveiled two AI virtual persons named “Ning Jiemei” (宁姐妹), comprising “She Xiaoning” and “Cha Xiaoning,” embodying She ethnic and tea culture respectively, to support digitized outreach and livestreaming training for women; and Kunbot’s DOBOT LUMO, unveiled August 5, 2026 by Shenzhen-based Yuejiang Technology (越疆科技), was described as the world’s first “omni-habitat” embodied humanoid robot combining emotion recognition, natural bipedal locomotion, and a proprietary “Kongyi” embodied large model, targeted at home companionship, child supervision, science education for children, and stage/commercial performance uses across China.

AI Short-Drama Virtual Actors : Chinese AI-generated short dramas produced entirely with virtual actors have become a major commercial phenomenon, led by “Fang Taozi” (方桃子), the AI female lead of the short drama “The Girl Who Was Laid Off” (《被裁掉的女孩》), created using the LibTV AI video platform by director “Feifei Fei,” which garnered over 220 million views on a single platform and more than 800 million discussion volume across the network; Fang Taozi’s independent social account amassed roughly 400,000–410,000 followers within about 30 days, and her commercial rate card reportedly reached RMB 168,000 for 1–20 second videos, RMB 208,000 for 21–60 seconds, and RMB 258,000 for videos over 60 seconds — reportedly exceeding many real-life influencers with millions of followers, though some MCN sources disputed the figures, citing an actual August rate closer to RMB 88,000. Her first brand deal, a colored contact lens (美瞳/“meitong”) advertisement for a cosmetic-lens brand released July 31, 2026, triggered nationwide controversy and legal scrutiny because Fang Taozi, lacking physical eyes, claimed in first-person narration that the lenses “felt comfortable all day,” which multiple lawyers (including from Beijing Dacheng Law Offices, Beijing Zhongwen Law Firm, and others) said likely violated China’s Advertising Law: colored contact lenses are classified as Class III medical devices (the highest risk category, same tier as heart stents), and Article 16 of the Advertising Law bars medical device ads from using any endorser to make claims, while Article 28 prohibits fabricating product-use experiences; since AI virtual persons lack civil legal personhood and cannot be “advertising endorsers” under the law, liability would fall on the operating company (Guangzhou My Film and Television Culture Media Co.) and the brand advertiser rather than on Fang Taozi herself, and the ad was reportedly removed from her Douyin account by August 6 amid the backlash. Local regulators in Guangzhou’s Panyu district stated that AI-generated content is not inherently illegal if labeled, but content that exaggerates effects or misleads consumers constitutes a violation, and consumers may file complaints via the 12315 hotline. The episode occurred alongside other AI virtual-actor monetization cases: in the AI comic-drama “Yu Ni Shenqing Qinru Yusheng” (《予你深情侵入余生》), male lead “Duan Yan” drove real-world sales exceeding 2,000 units (roughly RMB 2.5 million) of a four-leaf-clover bracelet featured in the show; virtual singer “Yuri” surpassed 1 million followers and received the first officially registered “virtual idol identity certification” (No. 0001) from the Beijing Economic-Technological Development Area (北京经开区/亦庄) in January 2026, with brand deals from Northface and Lantu Auto; and an AI idol-survival show called “12 Xingzuo Lianai Sai” (《12星练赛》/“12 Zodiac Trainee Competition”) cast twelve AI-generated “trainees” modeled on zodiac signs, generating a spinoff AI dating variety show with nearly 4 million followers and product placements from skincare and home-appliance brands. Separately, mainland media reported a broader crisis for human short-drama actors: with AI-generated dramas comprising over 95% of the 128,000 new short dramas released in Q1 2026 (up from under 5% previously), real-drama production volumes fell roughly 75% year-on-year, and actor day rates collapsed from RMB 5,000–8,000 to under RMB 3,000 for mid-tier performers, with bottom-tier extras earning as little as RMB 80–200 daily; some displaced actors reportedly sold facial likeness rights to AI digital-human training pools for as little as RMB 100–1,000, while platforms like Hongguo (红果) reportedly capped top-actor day rates at RMB 40,000 to curb speculative bidding wars. Separately, AI comic/short-drama production company Youma Group (优马集团) showcased its “REELMART” (万象厨房) short-drama AIGC platform at WAICO (Shanghai) and a Shenzhen cross-border e-commerce AI expo, offering a vertical large model for generating cross-border e-commerce digital humans, overseas ad videos, and short dramas embedding shoppable product placements aimed at Southeast Asian, Latin American, and Middle Eastern markets.

Media and Broadcasting Virtual Hosts : State and regional Chinese media outlets have deployed AI digital anchors and virtual hosts across broadcast and telecom services. China Telecom (中国电信), in partnership with Fujian Provincial People’s Hospital and Fujian Provincial Hospital, developed the “Jishi Medical Large Model” (济世医疗大模型) featuring smart medical digital human “Xiao Ji” (小吉), which supports dialect-based inquiries and voice-activated registration for elderly patients, reportedly cutting average visit times by 30% at Fujian Provincial People’s Hospital; the underlying “Token” (词元, the official Chinese term for AI token units set by the National Data Administration in March 2026) service is bundled into consumer mobile and broadband billing via new “Token packages” priced from RMB 9.9 to 49.9 monthly. In Anhui, the state-run Anhui New Media Group (安徽新媒体集团) launched “Zhong’an Zhimei Dana” (中安智媒大脑), an AI-powered newsroom platform integrating a “Zhian Zhibo” (中安智播) module featuring ultra-realistic digital-human news broadcasting and automated short-video generation, aimed at reducing repetitive editorial work. Hunan Jingshi TV (湖南经视) deployed AI anchors “Shengsheng” (声声) and “Shuangshuang” (双双) for round-the-clock, multilingual news broadcasting, and separately used AI restoration to revive classic variety-show footage (recreating scenes with performers Wang Han and Ma Ke) for a new AI-driven variety show, “Cesheng Yongbulao” (策神永不老). Anhui TV aired an AI-produced drama series, “Taohuatan Ji” (《桃花潭记》), explicitly labeled as AI-generated content. Jining Radio and Television (济宁广播电视台) similarly deployed AI anchor “Xiaoting” (晓婷). Xinjing Bao (新京报/The Beijing News), per its president Mao Xiaogang, has developed AI digital humans “AI Xiaobei” and “AI Xiaohui” and produced AI-based anti-fraud short drama “Lao Bei You Dianr Bei” (《老贝有点儿"背"》) and an AI-assisted documentary on the Tokyo Trials, alongside a generative ESG large-model tool (“Beike Caijing · Zhiyin ESG”) co-developed with Ant Group. In Zhejiang, Zhejiang TV’s emotional talk show “Xiexie Ni Lai Le” (《谢谢你来了》) grew its cross-platform matrix to 41.17 million followers with 2025 e-commerce sales reaching RMB 340 million, up 200% year-on-year, illustrating how digital-human and content-driven monetization has become integral to provincial broadcaster transformation strategies; the Beijing Municipal Radio, Television and other reports noted over 10,000 broadcast-affiliated short-video accounts nationwide, with digital-human-driven content increasingly used to modernize aging state-media production pipelines.

Healthcare and Insurance Virtual Assistants : China Telecom’s Fujian branch has deployed the “Xiao Ji” (小吉) smart medical digital human at Fujian Provincial People’s Hospital, guiding patients (including elderly users conducting dialect-based inquiries) through registration and triage, while Xiamen’s municipal medical insurance bureau operates a dedicated AI insurance robot. Jiujiang University’s affiliated hospital (九江学院附属医院) in Jiangxi launched a WeChat-based “AI digital-double” customer service function, explicitly scoped only to appointment scheduling, department information, and visit-process guidance (not diagnosis or prescriptions), providing 24/7 responses and automatically deflecting roughly 70% of routine inquiries away from human staff, escalating complex cases to a human hotline. Zhongyin Insurance (中银保险) rolled out AI digital-human figures across several provincial branches (Shenzhen, Jiangxi, Jiangsu, Shanghai) during its 2026 “7.8 National Insurance Public Awareness Day” campaign, integrating digital humans into underwriting and claims explainer videos and disaster-prevention outreach; separately, Dajia Property Insurance (大家财险) introduced AI digital human “Lingze” (灵泽) across 36 affiliated institutions and over 1,000 branch outlets nationwide to deliver flood-season risk-reduction education. Yantai’s municipal government response to a local legislator’s proposal confirmed deployment of a “digital human guide” (数字人导服员) system within its unified government AI large-model service platform to handle citizen inquiries in a more approachable, conversational format, part of a broader push to apply AI in elder-care and tourism-related government services.

Retail, Marketing and Livestreaming : Chinese regulators have moved to police AI virtual hosts in e-commerce livestreaming: Beijing’s municipal market regulation bureau issued a consumer advisory requiring livestream sellers using AI-generated hosts to conspicuously and continuously label them (e.g., “AI digital human host,” “AI-generated”) throughout product demonstrations under the “Measures for Labeling AI-Generated Synthetic Content” and “Regulations on Live E-commerce Supervision,” and warned that beautified or color-corrected AI-rendered product images may misrepresent real merchandise (texture, sizing, fabric, defects), urging consumers to request unretouched photos/videos from customer service and to preserve evidence (screenshots, payment records, chat logs) for complaints via the 12315 hotline if deceived; Huaxia Net republished the identical advisory. Separately, Chinese e-commerce platforms have seen a wave of AI-generated “buyer” avatars and deformed AI apparel models (reports of six-fingered hands, extra limb segments in product photos) appearing in fashion listings, exacerbating returns in the women’s apparel sector; commentary (via Sohu/China News Service) argued that operators, brands, and technology vendors — not the AI avatars themselves — bear legal responsibility for such misleading content, and criticized weak enforcement of the mandatory AI-content labeling rule that took effect September 1, 2025. Elsewhere, dubious AI-generated “patients” and “experts” have appeared in advertisements for therapeutic patches (e.g., brands “Mingjia Laopu” 明家老铺 and “Meng’aoshen” 蒙奥神) marketed via Chinese short-video platforms, falsely dramatizing pain-relief testimonials using unlabeled or barely labeled AI-generated avatars, prompting China Consumer Association-affiliated investigative reporting; regulators noted these Class II medical devices’ promotional claims exceeded approved usage scope. In the home-goods sector, Jinpai Kitchen & Bath (金牌厨柜家居科技股份有限公司), based in Xiamen, Fujian, created “Jin Chanzi” (金婵紫) in 2022, described as the furniture industry’s first branded virtual digital human, as part of a “metaverse” marketing push targeting younger consumers, and remains tagged under China’s “virtual digital human” stock concept sector; similarly, telecom-equipment and other listed firms such as Sanxia Energy, Chuanwang Media (川网传媒), Bojui Communications (博瑞传播), and Hengyin Technology (恒银科技) are all classified under China’s A-share “virtual digital human” (虚拟数字人) investment concept sector, reflecting the breadth of listed Chinese companies now tied to digital-human technology narratives for market purposes.

Identity Verification and Governance : Haizhou District (海州区) in Lianyungang, Jiangsu, on August 5, 2026, launched “Yuanlian Chacha” (元脸查查), described as China’s first blockchain-based facial/voiceprint authorization verification public service platform, developed through a strategic partnership between Lianyungang Public Security Group (连云港安防集团) and platform operator Yuanlian FacesMarket (元脸FacesMarket), enabling creators to upload authorization documents generating a public verification link, allowing the public and regulators to check the authenticity of face/voice usage rights, with all authorization data stored immutably on-chain and a one-click reporting channel for suspected infringement; the launch was paired with China’s first Global Public Security-themed AIGC creation competition, and the district framed the platform as closing a “quick-verification loophole” wherein AI face-swap, virtual dubbing, and digital-human broadcasting tools proliferate faster than legal enforcement, aiming to build Jiangsu’s first standardized, nationally emulable AIGC compliance zone. Separately, a Sina Finance opinion piece by Nankai University’s Chen Bing, writing on the Ministry of Industry and Information Technology’s draft “Measures for the Administration of Digital Virtual Human Information Services,” outlined China’s move toward a full-chain accountability system dividing responsibility among “technology supporters,” “service providers,” “service users,” and “distribution platforms,” banning unauthorized use of a specific natural person’s likeness or voice to build a digital-human model, and mandating content moderation and traceability obligations, particularly for politically sensitive, government, financial, healthcare, and educational applications of virtual humans in China. Additionally, Xinjing Bao’s Mao Xiaogang, in a cross-strait media forum speech, disclosed that the state-affiliated newspaper had produced multiple AI-driven public-interest and cultural products using its “AI Xiaobei” and “AI Xiaohui” digital humans, framing responsible AI content governance as central to cross-strait media cooperation between mainland and Taiwan outlets.

06 August News

Healthcare : Insurance and health-management platforms in Taiwan are examining AI-driven “health broker” concepts, with Dentsu Taiwan’s creative team and Tamkang University’s Strategic Foresight Center proposing that insurers, medical groups, and retail platforms could integrate personal health data, diet, exercise, and medical resources into personalized digital health plans, potentially deploying AI companions and data-driven coaching tools to sustain behavior change. Separately, Beijing-based Zheyuan Technology (哲源科技) has built a “digital drug trial ground” using digital twin models to simulate virtual patients (“虚拟病人”) for drug testing, working with Peking University Cancer Hospital to run virtual clinical trials in which digitally modeled patients received simulated drug treatment, with computer predictions matching real trial outcomes for a pancreatic cancer drug candidate now in Phase I trials. In the insurance sector, Dachild Insurance Service (大童保险服务) executive Guo Mengwen advocated an “AI-enhanced human” model rather than unmanned AI sales, warning that AI-generated virtual host livestreaming lacks the credentials to serve as licensed agents, and that AI should handle standardized risk assessment while licensed human brokers retain oversight of complex family financial planning and final client communication.

Entertainment : AI-generated virtual actress Fang Taozi (方桃子), the lead of the AI short drama “The Girl Who Got Laid Off” (被裁掉的女孩) produced by Guangzhou-based MCN My Film and its investor Shisan Culture, became a major case study in AI virtual celebrity commercialization; the drama surpassed 200 million views on a single platform, and Fang Taozi’s independent Douyin account amassed over 400,000 followers within a month, commanding advertising rates of RMB 168,000 for 1-20 second videos, RMB 208,000 for 21-60 seconds, and RMB 258,000 for videos over 60 seconds—rates exceeding many real influencers with millions of followers. Her co-star Zhou Yiheng also built an independent account, and their “CP” (couple) storyline trended on Weibo. Fang Taozi’s team pursued an operating model mimicking real celebrities, posting daily vlogs of cooking, tennis, and shopping alongside scripted drama content to build a “living” persona; she later filmed a colored contact lens (美瞳) advertisement claiming personal comfort after “wearing them all day,” which triggered controversy over AI’s inability to genuinely test physical products, given that colored contacts are classified as Class III medical devices in China subject to strict advertising rules under the Advertising Law. Legal experts including Zhao Liangshan and Beijing Dacheng law firm partner Li Weihua noted that AI virtual figures are not legally recognized “endorsers” under Chinese advertising law, so liability instead falls to the operating company (in this case an MCN) and the brand advertiser if false claims are made; Guangdong regulators and commentators called for AI content to be explicitly labeled and for operators to avoid fabricating usage experiences. Beyond Fang Taozi, other Chinese AI virtual idol projects have emerged: Maiya Media’s (麦芽传媒) “12 Star Training Competition” (12星练赛) and a rival AI idol survival show “Xingguang 108” (星光108) both simulate real K-pop-style trainee competitions entirely with AI-generated contestants, while Bilibili ran an “AI Idol Debut Season” (AI偶像出道季) crowdsourcing contest that drew over 160 million views, with AI singer “Yuri,” created by a Tsinghua-trained developer, gaining recognition from state broadcaster CCTV and securing brand deals with automaker Voyah (岚图). Legacy Chinese entertainment company Yuehua Entertainment (乐华娱乐), which previously acquired virtual idol group A-SOUL from a ByteDance unit, launched a new AI-driven idol group “HeyDream,” debuting at ChinaJoy with AI-generated singing and dancing performances comparable to human idols. Elsewhere, actress Joey Wong (Wang Zuxian, 王祖贤) licensed her younger image to NetEase Games for an AI-generated short film “Silhouette” (倩影) recreating her classic film roles (Nie Xiaoqian, Bai Suzhen), and actors Wu Qihua (吴启华) and veteran actor You Benchang (游本昌, 93) licensed younger AI likenesses for new AI-generated dramas, prompting debate about digital “resurrection” of past personas, appropriate boundaries for licensing (single-use versus perpetual rights), and posthumous digital revival ethics as discussed by entertainment law scholar Song Yao. AI short dramas overall have surged: Chinese online video industry data shows AI-generated micro-dramas made up over 95% of newly released short dramas in Q1 2026, with hundreds of thousands of AI dramas produced monthly on platforms like Douyin, though fewer than 1% achieve breakout popularity; Anhui TV aired China’s first fully AIGC-produced costume drama “Peach Blossom Pool Chronicle” (桃花潭记), crediting an “AIGC director,” while Zhejiang TV aired the AI drama “Chinese Legend: White Snake” (中国传说·白蛇). Platform Hongguo Duanju (红果短剧) issued guidelines banning duplicate “AI faces” across multiple dramas to address “uncanny valley” backlash and homogenized AI-generated facial features, a phenomenon linked to biased training data (skewed toward young, light-skinned, heavily beautified faces, with Korean drama-influenced aesthetics) and cost-driven use of free AI tools. The economic disruption to human short-drama actors has been severe: top-tier actor Ke Chun’s (柯淳) daily rate reportedly fell from RMB 100,000 to around RMB 60,000, mid-tier actors saw rates fall from RMB 5,000-8,000 to under RMB 3,000 daily, and background extras in Hengdian (横店) earn as little as RMB 80-150 per day, with some actors selling facial likeness rights to AI digital-human databases for as little as RMB 500; actor Xu Peng, laid off amid the AI short-drama boom, pivoted to founding a media company in Qingdao planning original human-cast dramas and a training camp for amateur actors, while actor Zhou Rui described Hengdian’s stratified survival crisis across four tiers of performers. Animation scholar Liu Shuliang of the Communication University of China, author of the book “Virtual Idols, AI Digital Humans, and Cyber Life,” analyzed the “uncanny valley” effect driving public discomfort with AI faces and argued Chinese AI drama production remains stuck in this technical valley, unable to convincingly replicate human micro-expressions and lived emotional nuance.

Marketing : Fang Taozi’s success has catalyzed broader interest in AI virtual influencers for advertising in China, with agencies like PureblueAI Qinglan (清蓝), BlueFocus (蓝色光标), and Guangdong Zhihui Zhaoneng (智慧兆能, in partnership with Xiaoice/小冰) positioning AI-generated content and digital human marketing as core service offerings; advertising firm Guangyuan Zhineng (光元智能) and Ziguang Huazhi (紫光华智) in Guangzhou similarly offer AI short-video production, digital human customization, and Cantonese-language AI dubbing for regional brand campaigns. Chinese state media outlet China Culture Daily (中国文化报) deployed AI news anchors “Zhong Xiaowen” and “Zhong Xiaomei,” digital avatars modeled on real reporters, to deliver daily cultural news broadcasts as part of a broader push toward “short, real, new” (短实新) multimedia storytelling. Elsewhere, apparel manufacturer Handan Chenhan Clothing (邯郸晨涵服装) in Hebei’s Shexian county adopted a “real person + AI” hybrid livestreaming model, using a custom AI virtual digital human to deliver round-the-clock product livestreams synced to pre-recorded scripts, automatically responding to comments and streaming continuously without human staffing, which the company credited with boosting online exposure and expanding sales channels to Henan, Shandong, and Jiangsu. Chinese digital exhibition company Fengyuzhu (风语筑) has used Baidu’s Wenxin Yiyan large language model to power AI-driven multimodal interaction for digital human tour guides in immersive exhibition spaces and metaverse environments, while listed company Zhongshi Media (中视传媒) is integrating AIGC-generated virtual digital humans and AR interactivity into physical tourism sites such as the Wuxi film and TV base, combining 4K/8K ultra-high-definition production with AI to create immersive cultural tourism experiences.

Gaming : To mark the successful UNESCO World Heritage inscription of “Jingdezhen’s Handicraft Porcelain Industry Heritage,” Jingdezhen authorities and a game developer jointly launched “Digital Jingdezhen: Porcelain Capital Craftsman” (数字景德镇·瓷都小匠), the world’s first heritage-nomination mini-game, which incorporates an AI-driven digital human “porcelain master” (“瓷器老师傅”) built on a multimodal knowledge base of over 1,900 real porcelain artifacts, using AI-driven facial expression, lip-sync, and voice generation to let players interact conversationally with a virtual craftsman character while also using AI image-to-3D generation so players can design their own “digital porcelain” pieces from an uploaded photo. The game’s producers stated the AI digital human and generative 3D technology aimed to make cultural heritage “visible, experiential, and creatable” for younger, especially international, audiences, and reported that international heritage experts praised the immersive interactive approach at the UNESCO World Heritage Committee’s 48th session announcement.

Enterprise : Chinese fintech software firm Xinchen Technology (新晨科技) disclosed plans to deploy blockchain-based service platforms (BaaS) with applications spanning metaverse digital humans, VR/AR wearables, and cloud rendering for financial-sector clients, part of its broader digital-currency and cross-border payments business. Startup incubator and media company Chuangye Heima (创业黑马) has run a “metaverse digital human accelerator” initiative supporting SMEs in adopting AI applications, in partnership with Huawei, Zhipu Huazhang (智谱华章), and Xiamen Suanneng, combining large-model infrastructure with AI training and enterprise services for small businesses. Listed education company ST Kaiyuan (开元教育), through its Tianhu Education subsidiary, has built a “digital human laboratory” for producing and filming instructional digital humans and developed “metaverse” course content covering 3D modeling, VR scene creation, and character animation. Jiangsu Cable (江苏有线), a state-owned broadcasting network operator, deployed an AI “digital employee” named “Jiang Xiaochan” (江小婵) for exhibition-based tourism consultation and public service interaction in Nanjing, complete with branded merchandise. Listed company Digital Human (数字人), a Beijing Stock Exchange-listed firm specializing in digital anatomy education products used in over 500 medical schools nationwide, restructured its board to pursue a “one body, two wings” strategy encompassing its core medical-education business, overseas expansion (with products piloted in over 40 countries), and a newly established cloud-services subsidiary, Yichuang Cloud (易创云), targeting SaaS-based digital human cloud platform commercialization; its “Digital Human STEM System (Cloud Edition)” has been adopted by 364 colleges with nearly 400,000 registered users. Guangdong bank Guangfa Bank (广发银行) launched “Shuqitong” (数企通), a digital ecosystem platform for SMEs integrating digital HR, tax, and welfare management modules, alongside blockchain-based cross-border payment tools tied to China’s CIPS system. In manufacturing and retail contexts, Chinese steel trading platform Zhaogang.com (找钢网) built an AI “digital employee” trained on historical WeChat, DingTalk, and internal chat logs to replicate veteran trader decision-making in steel trading workflows, deploying multi-screen human-AI collaborative dispatch systems for order intake and quoting while reserving high-stakes deal approval for human staff.

Cultural : A Suzhou-based mechanical engineer named Zuming created an AI-generated guofeng (national aesthetic) music video “Su Embroidery Dream” featuring a fictional AI character “Su Liniang” (苏丽娘), blending traditional Suzhou embroidery (Su embroidery) techniques and Kunqu opera aesthetics (referencing the character Du Liniang from “The Peony Pavilion”) with AI-generated visuals depicting gravitational wave patterns rendered as embroidery motifs, winning a prize at the “Suzhou Intangible Heritage AI Wonder Day” competition; the project has since been extended into an AR interactive exhibit at Suzhou Museum’s western branch and has reportedly driven a rise in demand for Su-embroidery-branded products among younger consumers in Pingjiang Road shops. The Zhejiang-based “AIHangzhou City Living Room” exhibit in Hangzhou’s Wensan Digital Life District showcased Chinese-developed technologies including a bio-inspired “cyber fruit fly” digital brain model developed by a Zhejiang University-affiliated team (reportedly praised by Elon Musk), a “MusicMol” molecule-to-music AI interactive model created by Zhejiang University researchers linking chemical structures to musical composition, and a humanoid robot “Wanda” from Chinese robotics company Youli Qi (优理奇), alongside 3D-scanning “digital twin” avatar generation booths allowing visitors to create personal 3D virtual models in 0.3 seconds. In Wuzhen, the 2026 Wuzhen Youth Science and Technology Carnival featured VR/AR digital human research courses letting teenagers build personal digital avatars, part of a broader youth-oriented science education push blending gaming, robotics, and AI-generated content across the ancient town.

Ethical Contexts : The rapid commercialization of Chinese AI virtual actress Fang Taozi has intensified debate over whether AI virtual “performers” can lawfully serve as advertising endorsers in China; legal experts (Zhao Liangshan, Li Weihua, Xiao Sa, Qiu Zhimiao, Meng Xiaodong) generally agree that under China’s Advertising Law, AI-generated digital humans lack legal personhood and cannot be classified as endorsers, meaning liability for false or exaggerated claims—such as Fang Taozi’s assertion that she “wore” a contact lens comfortably—falls on the operating MCN company and brand advertiser rather than the AI character itself, and that China’s regulations (including AI-generated content labeling rules effective since September 2025, and regional guidelines from Hainan and Ningbo) require explicit disclosure that content is AI-generated and prohibit fabricated usage testimonials, particularly for medical devices like Class III colored contact lenses. Separately, a broader wave of Chinese AI-driven health-product livestreams has drawn scrutiny: investigative reporting found multiple “miracle oil” and pain-relief product livestreams (including a “Hong Kong Chinese Medicine” brand actually manufactured in Jiangxi, and a “Tibetan intangible-heritage scorpion venom” product actually made in Henan) using AI-generated “recovery” testimonial scenes alongside real hosts to simulate healing effects, with some videos carrying only faint, easily overlooked “AI-generated” disclaimers, and false claims about non-existent intangible cultural heritage certification and misrepresented insurance backing (misusing outdated China Pacific Insurance branding). Additionally, deepfake and unauthorized AI likeness cases have affected multiple Chinese celebrities, with actors Wang Jinsong, Wen Zhengrong, and others publicly reporting AI-generated deepfake videos and livestreams using their voices and faces without consent for unauthorized product endorsements, prompting several studios to issue public statements condemning the infringement of image and voice rights.

05 August News

Entertainment : The AI virtual being landscape in China’s entertainment industry has centered on AI actor “Fang Taozi” (方桃子), star of the AI short drama “The Girl Who Was Laid Off” (被裁掉的女孩), produced by a small team led by director “Feifeifei.” Since debuting on Douyin in June, the drama surpassed 200 million views, and Fang Taozi’s independent social media account amassed over 400,000 followers within 30 days, engaging audiences with lifelike Vlogs depicting mundane activities like making sandwiches, buying flowers, and commuting. Her male co-star Zhou Yiheng, with over 120,000 followers, formed a fan-celebrated “CP” (couple) pairing dubbed “Tao Zhi Yi Heng” that trended on Weibo. Fang Taozi’s advertising rates, published via Juliangxingtu (巨量星图) data, reached up to 258,000 yuan for videos over 60 seconds—surpassing many real influencers with millions of followers—and she completed her first paid endorsement for a colored contact lens (meitong) brand, though this drew controversy over whether an AI could credibly claim to have “worn” the product, raising Advertising Law compliance questions about false endorsement claims for medical device-classified eyewear. Alongside Fang Taozi, Yuehua Entertainment (乐华娱乐) launched Asia’s first AI-driven real-person idol group “HeyDream” in March, blending AI-generated promotional content with real member stage performances, and previously acquired the virtual idol group A-SOUL from ByteDance’s Nice-Time Studio for 30 million yuan. The AI virtual singer “Yuri,” created by a Tsinghua-trained developer named Zhao Hanqing, became China’s first officially certified virtual idol, receiving a “digital artist card” from the Beijing Economic-Technological Development Area, participated in the Boao Forum for Asia, and secured brand deals with Voyah and other companies; her debut single “Surreal” surpassed 12 million views. Other studios, including Yaoke Media (耀客传媒) and Yuxiao Media (聿潇传媒), have signed AI virtual actors like Qin Lingyue and Lin Xiyan, or licensed real influencers’ likenesses (such as Han Anran and Xu Zhibin) to create reusable AI actor assets for short dramas. Veteran actors including Joey Wang (Wang Zuxian), Wu Qihua, and You Benchang have licensed AI reproductions of their youthful likenesses for commercial use, such as Wang’s appearance in a NetEase game promotional AI short film titled “Qian Ying.” iQIYI has built an “AI Actor Library” with 117 contracted actors offering digital likeness rights via 3D scanning and motion capture, and its AI agent platform “NaDou Pro” supports AI-driven film and TV production. AI idol talent competition shows have also emerged, including Douyin’s “Xingguang 108” (星光108), billed as the world’s first AI virtual idol competition selecting 9 winners from 108 contestants, and “12 Xing Lian Sai” (12星练赛) by Maiya Media, which replicates real idol audition formats using AI contestants and has drawn over 545,000 Douyin followers; Bilibili’s “AI Idol Debut Season” attracted contestants generating over 160 million views, with virtual idol “ENA” (恩娜) topping voting. Separately, Hunan TV announced China’s first all-AI singer variety show “Phantom Voice Singer” (幻音歌手) featuring nine virtual AI singers, while snack brand Qiaqia (洽洽食品) launched its AI brand ambassador “Qiakeer” (洽可儿) for retail, livestreaming, and e-commerce applications. AI-generated short dramas have also entered traditional broadcast television: Anhui TV aired the first fully AI-generated drama “Peach Blossom Pond Chronicle” (桃花潭记), depicting Ming Dynasty papermaking heritage, while Zhejiang TV broadcast the AI comic drama “Chinese Legends: White Snake” and Dongfang TV aired the sci-fi AI series “When the Sun Falls” with 35 million viewers. Concerns about AI-generated faces achieving troubling homogeneity (“AI face” phenomenon) prompted platforms like Hongguo Duanju to restrict reuse of identical AI faces across multiple productions.

Marketing and Advertising : AI virtual beings are increasingly deployed in Chinese digital marketing and advertising campaigns, exemplified by Fang Taozi’s high-value contact lens endorsement and the AI comic drama “Yu Ni Shen Qing Qin Ru Yu Sheng” (予你深情侵入余生), whose male lead “Duan Yan” promoted a matching bracelet that sold over 2,000 units for total sales exceeding 2.5 million yuan. Baijiu producer Luzhou Laojiao (泸州老窖) deployed an AI “digital wine connoisseur” named “Lulin’er” (泸麟儿) serving over 30 million members, while food company COFCO and Qiaqia have adopted similar branded AI spokesperson strategies through iFlytek’s (科大讯飞) newly launched “hyper-realistic real-time digital human generation technology,” which reduces digital human creation from hours-long video capture sessions to a single photograph, enabling more natural expressions and complex actions like drinking or dancing; this technology underpins iFlytek’s marketing digital human solutions serving enterprise clients. Douyin’s short video platform introduced mandatory AI labeling requirements effective August 10, requiring creators to explicitly mark AI-generated content in e-commerce livestreaming videos, including AI-generated digital human hosts, product demonstrations, and voiceovers, to prevent misleading claims about product effects. Legal experts have noted that AI virtual beings cannot serve as legally recognized advertising endorsers under China’s Advertising Law since they lack natural or corporate legal personhood, meaning liability for false advertising claims falls to the operating company, brand, or MCN agency behind the AI character rather than the AI itself; this issue was highlighted specifically regarding Fang Taozi’s contact lens endorsement, given that contact lenses are classified as Class III medical devices under Chinese regulations, and an AI cannot genuinely demonstrate product wearing experience, raising risks under Advertising Law Article 28 regarding false experiential claims.

Digital Human Platforms and Enterprise Applications : Chinese firms are rapidly commercializing enterprise-facing digital human and virtual employee platforms. Nanjing-based Bapoint Digital Technology (南京八点八数字科技) developed the digital human generation model XMEN.AI, capable of real-time driving 2D, 2.5D, and 3D digital humans with lip-sync latency as low as 0.16 seconds, powering its product “Yihua” (亿话) which has attracted over 10,000 software users, and has since launched AniShort, a one-stop AI film and television production collaboration platform featured at ChinaJoy 2026, integrating script generation, storyboard planning, and automated video generation into a unified workflow that reportedly triples short-drama production efficiency; clients include Yitong Production (壹同制作), Ruosen Digital (若森数字), and Chongqing Bola (重庆博拉). Shenzhen Qinghong Data Technology’s “Qinghong AI” (青虹AI) positions its Agent Harness platform as an enterprise-grade “AI digital employee system” distinct from mere chatbots or RPA scripts, combining large language models, RPA automation, and enterprise agent architecture to handle tasks like invoice reconciliation and ERP data entry, with deployments at a Shenzhen 3C manufacturer reducing after-sales repair processing from 40 minutes to 2 minutes. In banking, Chinese regional banks including Chengdu Bank, Guiyang Bank, Shaoxing Bank, and Guangzhou Rural Commercial Bank have issued tenders for virtual digital employees, with Sichuan Rural Credit Union investing 1.7 million yuan in a digital human customer service product including avatar customization and interaction configuration; CITIC Bank’s credit card center has repeatedly recruited AI intelligent training managers to support digital human customer service infrastructure. Ping An Insurance Group unveiled its recruitment AI digital human “Chengxiao’an” (橙小安), designed to provide job recommendations, transparent recruitment process visibility, and 24/7 Q&A support for approximately 6,000 annual graduate hires, explicitly designed to deliberately exclude bias-inducing labels like gender, region, or school tier. In Jilin, Zhonglin Education Technology (钟霖教育科技) in Changchun’s Nanguan District Science and Innovation Park specializes in digital human and AI agent R&D, partnering with agricultural technology firms to co-develop digital human livestreaming applications. National Healthcare Security Administration launched a public design competition for an official “medical insurance digital human” mascot to embody trustworthy, professional public service branding infused with Chinese cultural elements.

Cultural Heritage and Public Sector Applications : Chinese cultural institutions are deploying virtual humans and AI-driven digital reconstructions to preserve and internationally showcase heritage. The Jincheng Museum (晋城博物馆) in Shanxi Province created the “Tiangong Yingxiang” (天宫映像) digital exhibition of the Yuhuang Temple’s Yuan Dynasty painted sculptures and murals, exhibited in Ho Chi Minh City, Vietnam, using VR technology allowing visitors to virtually walk through the temple hall and interact with sculptures, contributing to China’s digital culture “going global” strategy and Sino-Vietnamese cultural exchange. Zhongguo Chuban Chuanmei (中国出版传媒) and its subsidiary Gulian (古联公司) have built a large-scale digitized ancient text corpus of over 66,249 titles and 11 billion characters, using digital human explainers alongside short videos and audiobooks to make classical Chinese texts more accessible to younger audiences. In Guangdong, the Guangdong Science and Technology Vocational College has proposed leveraging AIGC and digital avatar tour guides, AR heritage route recommendations, and metaverse-based Lingnan opera and lion dance training curricula to strengthen cultural identity among Hong Kong and Macau youth as part of Guangdong-Hong Kong-Macao Greater Bay Area integration efforts, framing digital avatars as tools to deepen “xin lian tong” (heart-to-heart connectivity) between mainland and special administrative region residents.

Ethical, Legal, and Regulatory Contexts : Chinese regulators and courts are actively grappling with AI virtual being governance. The Beijing Internet Court ruled in a landmark case involving the short drama “X Trace” (X迹) that AI-generated faces need only be “recognizable” by the public as resembling a specific real person to constitute portrait rights infringement, rejecting the defendant Hangzhou company’s “technological neutrality” defense and establishing that alleged infringers bear the burden of proving their AI generation process did not use identifiable personal data; the court also ruled that copyright licensing does not exempt distributors from portrait rights liability. The Cyberspace Administration of China’s “Qing Lang” campaign shut down over 1,840 illegal livestreaming accounts for exploitative content, while draft “Digital Virtual Human Information Service Management Measures” released in April 2026 establish that any use of real individuals’ faces or voices for virtual human modeling requires individual consent, extending platform regulatory responsibility across the entire digital human service chain. Following controversy over Fang Taozi’s contact lens endorsement, lawyer Zhao Liangshan clarified that AI virtual humans cannot be legal advertising endorsers under Chinese Advertising Law, placing compliance responsibility on the operating company and brand; consumers retain recourse via consumer protection complaints or civil lawsuits against the responsible human/corporate parties, not the AI itself. Chinese state media and industry associations have also raised concerns about AI-generated short dramas’ content risks, including copyright infringement via “face-stealing” and “voice cloning,” script plagiarism, and content harmful to minors (promoting materialism, distorted views on relationships, or violence); a June 2026 government crackdown (“Jianwang 2026”) targeted AIGC copyright violations, and China’s first criminal case involving illegal recording of over 1,700 pirated AI short dramas concluded with convictions. Separately, a Vanderbilt University professor’s research, reported by Japanese media, alleged that a Chinese company called “GoLaxy” (referred to in Chinese reports as 中科天玑) developed AI tools using hundreds of trained “virtual personas” to influence public opinion in Taiwan, Hong Kong, and Japan on behalf of Chinese government and intelligence entities, representing a cognitive warfare application of AI virtual beings that Western analysts have flagged as a national security concern.

04 August News

Cultural : Content creation increasingly incorporates AI-driven virtual humans across China. At the 2026 World AI Conference in Shanghai, digital human kiosks and AI agents are being deployed for public services such as registration and ticketing in old telephone booths repurposed as “AI companion booths,” while Shanghai positions itself as a testbed for city-scale virtual human deployment. In Yunnan, the “Folk Songs Grand View” television special commemorating the 50th anniversary of the Yunnan-Tibet Highway used a virtual digital host modeled on Yu Shaolong, a bilingual broadcaster from Diqing Tibetan Autonomous Prefecture, created by a team at Yunnan University of Finance and Economics led by Professor Yi Hong; this virtual host supports multiple Southeast Asian languages and accompanied AI-generated historical imagery reconstructing scenes from the highway’s construction. In Shanghai’s Huangpu district, the memorial site of the Central Committee’s former political bureau office upgraded its immersive digital exhibition using film-grade digital character modeling to recreate historical figures and events from 1928-1931 for visitors. In Xi’an, the metro system partnered with Alipay for a “Tang Poetry Capital” campaign allowing users to obtain digital human avatars styled after Tang poetry themes as an incentive for participation. Jiangxi’s cultural policy discussions emphasize using AI virtual tour guides, such as an “AI guide Wang Bo” deployed at Nanchang’s Tengwang Pavilion, alongside three-dimensional modeling and virtual light effects to animate heritage sites. Guangxi’s Nanning has emerged as a hub for AI-generated short dramas and animated films, leveraging local ethnic and non-heritage cultural resources; productions like “Legend of Qi Xia” use AI-digital humans and multilingual dubbing, and the “Feitutu FLYTOTO AI Short Drama Co-creation Platform” launched by Guangxi Publishing Media Group integrates script generation, digital human production, and multilingual translation for export to Southeast Asian markets.

Entertainment and Gaming : At ChinaJoy 2026 in Shanghai, Tencent Games unveiled an “AI full set” of technologies spanning AI-native content, including “Digital Jingdezhen: Porcelain City Craftsman,” the world’s first UNESCO heritage-themed AI interactive game featuring AI 3D generation and intelligent digital human interaction that lets players role-play as ancient kiln workers; “Cat Immortal Amulet” (猫仙札), which uses AI agents to create independent NPC personalities with dynamically generated dialogue and branching storylines; and internal AI tools including Craft, LAP, MagicDawn, and MagicLomix for game asset production. NetEase Games similarly used digital human and AI technology to bring actress Joey Wong (王祖贤), retired for 22 years, back to screens via an AI digital replica of her 20s-to-30s appearance, licensed for the game “Tianxia” in an ad short film called “Reflection” (倩影), which surpassed 2.98 million views on Bilibili; other veteran Hong Kong actors including Jackie Chan, Wu Qihua (Michael Wu), and You Benchang have similarly licensed youthful or classic digital likenesses for AI-generated films and microdramas, such as You Benchang’s AI-assisted microdrama “Ji Gong: The Fake Dragon Subduer” which recreated his 1980s Ji Gong character. SenseTime showcased a “Future Life” zone at WAIC 2026 featuring AI products including Seko, described as the world’s first integrated AI video creation agent with over 1 million creators, and a robotic retail concept store, SenseMart Go, deployed simultaneously in Shanghai’s World Expo Museum, Pudong Expo Tiandi, and West Bund International Convention Center, where robots autonomously handle product retrieval and restocking. AI-generated short dramas have become a major industry force: Tencent-linked data shows AI microdramas comprised over 95% of new titles in Q1 2026, with real-life drama production in Hengdian dropping roughly 75% and studio vacancy reaching 78%, prompting actors to take up street vending or tourism gigs; conversely, a notable success story is the AI short drama “The Girl Who Got Laid Off,” whose AI virtual lead character “Fang Taozi” attracted over 400,000 followers within 30 days and secured advertising rates of 258,000 yuan for 60-second-plus videos, exceeding many real influencers due to brand demand for “controllability” and lack of scandal risk. Guangxi’s Nanning also positions itself as a national leader in AI short-drama production and export, with companies like Fengzhi Culture and Yuguangtongchen Culture using AI-assisted scriptwriting, digital human performers, and multilingual dubbing to cut production costs to roughly one-third of live-action filming, targeting ASEAN markets via platforms like Stardust TV.

Marketing and Brand Ambassadors : iFlytek (科大讯飞) unveiled an upgraded intelligent interactive digital human platform in Beijing on July 30, introducing “hyper-realistic real-time digital human generation technology” that creates a fully interactive digital human from a single photograph, eliminating the previous two-hour video capture requirement; the platform integrates Xinghuo (Spark) voice large language models for long-term memory and vertical-domain knowledge, and supports complex actions like drinking water, singing, and dancing. iFlytek’s marketing digital human lineup includes “Lulin’er” (泸麟儿), an AI tasting ambassador for liquor brand Luzhou Laojiao serving over 30 million members, and “Qiake’er” (洽可儿), the first digital brand ambassador in the snack food industry, created for Qiaqia Food (洽洽食品) with a triple persona as “Chief Happiness Officer,” “Sunflower Town Seed-Roasting Heritage Keeper,” and “All-Around Brand Artist,” debuting via a themed single and music video called “The World Together with Qiaqia.” iFlytek also launched training digital humans (讯飞智聘, 讯飞培训, 星火陪练) covering recruitment, onboarding, and industry-specific practice scenarios across automotive, insurance, customer service, and dining sectors, completing over 10 million simulated training sessions annually; enterprise deployment includes mobile digital humans with obstacle avoidance and multilingual support across 37 languages used in overseas contexts such as a Kenyan telecom showroom and Bangkok’s airport. Guangdong’s new consumer protection regulations, effective October 1, 2026, explicitly regulate AI digital human livestream sales, requiring platforms to verify streamer identities and mandate clear AI-generated content labeling to prevent misleading consumers. Tianyu Digital Technology (天娱数科), whose stock surged following strong earnings, operates the virtual guofeng (Chinese-style) digital human “Tianyu” with over 5.5 million followers across platforms, an early example of culturally themed IP achieving overseas engagement, alongside AI video generation tools MetaMind and CineART supporting advertising placement on Douyin, Kuaishou, and Xiaohongshu. Other listed companies referencing virtual digital human business lines include Guomai Culture (国脉文化), which developed virtual spokespersons “Irene” and “Xiaolin” themed around company leadership, and Tianyoude Liquor (天佑德酒), whose Qinghai-based virtual character “Li Linger” promotes highland barley liquor blending regional heritage with digital marketing.

Robotics and Physical AI Integration : At WAIC 2026, SenseTime demonstrated embodied intelligence robots powered by its Kairos world model capable of autonomously performing household tasks like breakfast preparation, alongside its “Da Xiao Robot” (大晓机器人) subsidiary and an ACE 2.0 environmental data collection system using tactile gloves for robot training. At ChinaJoy 2026, Unitree Robotics displayed its G1 humanoid robot performing combat maneuvers and dance routines while autonomously recovering from being pushed, and Magic Atom (魔法原子) showed consumer robots for entertainment and guided-tour applications; robotics companies increasingly overlap with the digital-human and gaming ecosystem as physical embodiments of AI characters. Shanghai-based Meijue Technology (美爵科技), operating under the “Jinwuxie” brand, positioned itself in the humanoid robot licensing space by offering standardized portrait rights clearance for converting influencer, celebrity, and virtual digital human likenesses into physical humanoid robot forms for retail, exhibition, and brand ambassador use, addressing legal risks around unauthorized likeness replication in China’s growing embodied AI robot market.

Enterprise and Public Sector Applications : iFlytek’s digital human platform extends into enterprise “digital employee” applications, including financial advisors, virtual anchors, and tour guides across media, tourism, finance, education, and healthcare sectors, with roundtable participants from China Merchants Capital and Qiaqia Food discussing how large-model-driven digital humans are becoming “interactive, executable digital employees,” particularly in finance. Yidu Tech (医渡科技) partnered with Nanchang University’s First Affiliated Hospital to launch two medical AI agents—an emergency-teaching agent and “Peng Uncle’s Cardiac Insights” (彭叔心视界), described as China’s first cardiology-focused public-education agent—both integrating digital human capabilities from JD Health with hospital knowledge bases to support medical education and patient health literacy. In Guangdong’s Pingyuan County, teachers from Guangdong Post and Telecommunications Vocational and Technical College trained rural grassroots officials in using AI agent tools including Doubao’s AI-avatar (“AI fenshen”) video creation feature to enhance rural promotional content alongside other office AI tools. Qingdao’s Mengdou Technology (檬豆科技) built the OPIE (Qingdao Single-Person AI Entrepreneur) platform in partnership with the Chinese Academy of Sciences’ Institute of Computing Technology, supporting one-person AI startups across sectors including digital human and AI comic/animation ventures, having incubated 126 registered one-person companies within its first half-year of operation.

Ethical and Regulatory Contexts : Guangdong’s newly passed Consumer Protection Regulation and Greater Bay Area Cooperation Regulation, effective October 2026, directly address AI digital human commerce, requiring livestream platforms verify streamer identity information and mandating that AI-generated persona images and videos used in livestream sales be clearly and continuously labeled to prevent consumer deception; violations carry specific penalties under the new rules. Legal experts including attorney Wang Xiaoqi have noted that portrait rights licensing for AI digital doubles of actors—such as those granted by Wang Zuxian, Jackie Chan, and Michael Wu—require explicit, scenario-specific written consent detailing usage duration, context, and platform, alongside mandatory AI-generation labeling, with particular concern around deceased performers whose likenesses lack living rights-holders to authorize or contest use. Commentary around AI actors replacing human performers, sparked partly by unauthorized “face-blending” of AI virtual actors (such as Yaoke Media’s Qin Lingyue and Lin Xiyan, whose faces were found to closely resemble real actors including Zhai Zilu, Zhao Jinmai, and Zhang Zifeng), triggered widespread public backlash and boycott campaigns generating over 4.2 billion reads across social platforms, centered on unresolved portrait-rights and legal ambiguity for AI-generated composite likenesses in China, with historical AI-virtual-human infringement cases showing under 30% success rates for plaintiffs. Separately, American academic commentary alleges that a Chinese company, GoLaxy, has developed AI virtual personas trained on public social media data to conduct influence operations targeting Japan, Hong Kong, and Taiwan on behalf of Chinese state and intelligence interests, deploying “virtual profiles” designed to appear as friends or influencers while embedding ideological messaging.

03 August News

Entertainment : The Chinese gaming and digital entertainment sector showcased extensive virtual being applications at ChinaJoy 2026 in Shanghai, where Tencent Games (“腾讯”) unveiled an “AI full suite” spanning AI-native games, creation tools, and digital human technology. Tencent’s flagship title “Digital Jingdezhen: Porcelain Capital Craftsman” (数字景德镇·瓷都小匠), the world’s first heritage-themed AI interactive game tied to Jingdezhen’s UNESCO World Heritage bid, integrates AI digital humans, AI 3D generation, and procedural content generation to let players simulate historical porcelain-making processes across different dynasties. Tencent also demonstrated “Cat Immortal Tale” (猫仙札), a narrative RPG featuring AI Agent-driven characters with independently evolving personalities and dialogue, and four self-developed AI tools (Craft, LAP, MagicDawn, MagicLomix) covering game development, rendering, and art asset production. Century Huatong (世纪华通) hosted the second “Shulong Cup” Global AI Innovation Competition, awarding projects including Shengjing Technology’s “Senbox” AI 3D space-creation game and supporting Haiyi Interactive Entertainment’s three-platform matrix comprising Haiyi AI, overseas short-drama platform Moreshort, and the Seasoul digital human platform for original IP companion interaction. At the Qualcomm Snapdragon pavilion, visitors could generate personalized 3D digital avatars from a single photo using on-device processing that keeps data local, alongside an AI game coach providing real-time tactical guidance for titles like “Honor of Kings” without requiring network connectivity. Virtual VTuber Yongchu Tafei (永雏塔菲), a domestic virtual streamer who debuted in 2021 and built a large fanbase on Bilibili’s gaming channel, saw fan-made decorative content dominate Steam’s points shop trending page, demonstrating the crossover appeal of Chinese virtual idols beyond their original platforms. AI-generated idol survival shows also emerged as a major trend, with Douyin’s “Star Light 108” (星光108) featuring 108 AI-generated virtual trainees competing for 9 debut positions through a format mirroring traditional idol survival shows, complete with fan voting, though the project drew controversy over unauthorized use of real celebrities’ facial features in training data. Baidu App partnered with Yuehua Entertainment on the “AI Star Program” (AI星企划), an AIGC virtual idol competition where creators submit AI-generated characters for public voting on Baidu’s platform. Bilibili likewise launched an “AI Idol Debut Season” featuring 30 finalist AI trainees judged partly by celebrity mentors, while debates persist over whether AI performances can replicate the emotional authenticity of human idols.

Media and Digital Human Platforms : Chinese media and technology companies are deploying AI digital humans as core products across newsrooms and enterprise services. Science and technology firm iFlytek (科大讯飞) unveiled an upgraded intelligent interactive digital human platform in Beijing, enabling real-time digital human generation from a single photograph rather than lengthy video training, integrated with its Xinghuo (星火) voice model for long-term memory and domain-specific knowledge, supporting complex actions like singing, dancing, and drinking; the company has deployed branded digital humans for clients including Luzhou Laojiao (“泸麟儿”) and Qiaqia Food (“洽可儿”). Qiaqia Food (洽洽食品) launched “Chacky” (洽可儿), the snack food industry’s first digital brand ambassador, featuring an original theme song and music video, positioned as a persistent, reusable digital asset intended to reduce the costs and reputational risks associated with human celebrity endorsements while enabling deeper AI-driven interactive engagement with Gen Z consumers. Heilongjiang Daily Press Group (黑龙江日报报业集团) reported building a “Longtou Zhichuang” (龙头智创) digital human production platform in partnership with Tsinghua University and Harbin Institute of Technology, generating multiple award-winning media products as part of its broader “Longjiang Zhihui Cloud” AI infrastructure. Alibaba’s Taotian Group (阿里巴巴淘天集团), together with Nanjing University, developed “TaoMate,” an anchor-guided persistent memory framework enabling stable multi-minute digital human video generation with synchronized audio, achieving real-time streaming interaction for use cases including virtual hosts, livestreaming, and customer service avatars, with demonstrated roles such as household explainer, tech host, business consultant, and health advisor. Chengdu museums have deployed specialized digital humans including Wuhou Shrine’s “Xiao Zhuge” (小诸葛), built on Three Kingdoms historical records to provide multilingual guided tours, and Du Fu Thatched Cottage’s hyper-realistic “AI Du Fu,” which allows visitors to interact via glasses-free 3D displays to discuss the poet’s works and life in Sichuan.

Cultural and Heritage Preservation : Digital humans are increasingly used to revitalize China’s intangible cultural heritage (非遗) and traditional culture. At the 2026 “Xinjiang is a Good Place” intangible heritage exhibition in Shihezi, Shaanxi Provincial Cultural Center unveiled “Qin Qiaoqiao” (秦瞧瞧), a dialect-speaking AI digital human wearing traditional opera costume that provides Shaanxi-dialect interaction and cloud-based tour-guide services, described by its creators as the center’s fourth flagship digital human IP and a “full-capability digital employee” making heritage “touchable, sensible, and conversable.” The Oriental Yiyun Technology Group deployed mobile “intangible heritage caravans” equipped with 8K ultra-high-definition displays, AI, and extended reality (XR) technology touring heritage sites. In Yunnan, the Honghe Prefecture Culture and Tourism Bureau launched its own AI digital human tourism ambassador, “Honghe Niuniu” (红河牛牛), unveiled at a Beijing promotional event co-organized with People’s Daily Online to promote summer tourism and the region’s UNESCO-recognized agricultural and irrigation heritage sites. In Xi’an, the metro system partnered with Alipay for a “Tang Poetry Capital” campaign at Zhonglou Station, where users completing poetry quizzes through Alipay could win Tang poetry-themed digital human avatar designs alongside transit vouchers and other prizes, linking cultural heritage promotion to civic green-commuting initiatives. In Chongqing, Xiaba Middle School incorporated an AI digital human “raindrop” character into Chinese-language classes to guide students through poetic and humanistic content alongside AI-driven tools in physics, mathematics, biology, and health education.

Virtual Beings in Film and Celebrity Licensing : A significant trend has emerged in which retired or semi-retired Chinese and Hong Kong celebrities license their peak-era likenesses to AI companies for digital resurrection in advertising and film. Actress Joey Wong (王祖贤), retired from acting for 22 years, licensed her 20s-to-30s-era likeness to NetEase’s game “Tianxia” (天下) for its 18th-anniversary campaign; the resulting AI-generated short film “Reflections” (倩影), depicting her classic roles Nie Xiaoqian and Bai Suzhen “speaking” directly to players, surpassed 2.98 million views on Bilibili and was described as China’s first game-industry narrative short fully built on AIGC technology with complete star authorization. Hong Kong actor Lawrence Ng (吴启华), 62, similarly licensed his 20-year-old likeness to an AI film production company to star as the lead in an AI-generated movie without needing to appear on set, calling the compensation “quite substantial.” Veteran actor Tang Guoqiang (唐国强), famous for portraying Zhuge Liang, authorized his AI likeness to ByteDance’s Doubao education app, appearing as a virtual “Chinese language teacher” powered by ByteDance’s unreleased Seedance 2.5 video generation model. Jackie Chan (成龙) had previously used AI in the 2024 film “The Old Man” to recreate his 27-year-old appearance, though the technology at the time drew criticism for unnatural facial expressions. Elderly actor You Benchang (游本昌), 92, known for playing Ji Gong, partnered with Hongguo Duanju to produce the AI-assisted short drama “Ji Gong: The Fake Dragon Subduer,” recreating his 1980s television character; his family’s company co-produced the project, giving him unusual creative control compared to typical licensing arrangements. This trend has sparked debate over its impact on working actors: reports indicate some AI short-drama companies are purchasing five-year likeness rights from lesser-known performers for as little as 500 yuan, while background actors face declining opportunities as AI increasingly substitutes for extras. Legal experts including Beijing-based entertainment lawyer Wang Xiaoqi noted that portrait rights licensing requires explicit, written consent detailing scope, duration, and use cases, with AI-generated content requiring clear labeling to distinguish it from real performances; disputes remain over whether AI-generated content that fans deem “vulgar” (such as Wong’s nine-tailed fox transformation) could constitute reputational rather than portrait-rights infringement.

AI Actors and the Short-Drama Industry : China’s short-drama (微短剧) industry is undergoing rapid transformation driven by AI-generated virtual actors, with dramatic cost reductions reshaping the labor market for performers. Industry data cited in multiple reports shows AI short dramas comprised over 95% of the roughly 128,000 titles released online in Q1 2026, with production costs for AI dramas running approximately one-tenth those of live-action productions (roughly 50,000 yuan vs. 500,000 yuan per project), and some productions costing as little as 3,000 yuan. The AI short-drama and comic-drama market reportedly exceeded 22 billion yuan in the first five months of 2026, on pace to approach 400 billion yuan for the full year. AI-generated drama “The Girl Who Got Laid Off” (被裁掉的女孩) produced virtual actors “Fang Taozi” and “Zhou Yiheng” who trended on Weibo, with Fang Taozi’s standalone social media account amassing over 330,000 followers and 25 million likes by posting lifestyle content indistinguishable in tone from human influencers. Anhui TV aired “Peach Blossom Pool Chronicle” (桃花潭记), a fully AI-produced 20-episode drama centered on Xuan paper craftsmanship in Jingxian, ranking among the top ten primetime programs nationally, while Zhejiang TV’s “Chinese Legends: White Snake” became China’s first satellite-television AI drama, ranking second nationally in its broadcast slot. This has coincided with a reported 75% year-on-year drop in live-action short-drama production starts and over 80% studio vacancy rates at Hengdian, China’s major filming base, with background actor daily wages falling from 1,500-2,000 yuan to under 100 yuan; some displaced performers have shifted to selling vegetables or driving for ride-hailing platforms, while others have pivoted to producing AI dramas themselves. On July 31, 2026, the National Radio and Television Administration issued Order No. 16, the “Micro-Short Drama Development Management Measures,” establishing China’s first dedicated regulatory framework requiring classification-based content review, mandatory AI-content labeling, and copyright verification for AI-generated material, effective September 1, 2026; broadcasting institutions are now required to build end-to-end content risk-control systems spanning script pre-review, filing management, and production oversight, with differentiated controls for self-produced, co-produced, and imported dramas. Separately, streaming platform Youku’s president Wu Qian stated publicly that AI cannot replace real actors, emphasizing that Youku would continue cultivating human performers despite the cost pressures from AI-generated content.

Museums and Public Cultural Institutions : Beyond dedicated heritage campaigns, Chinese museums are integrating digital human guides into permanent visitor experiences as part of broader digital transformation efforts, exemplified by Chengdu’s Wuhou Shrine Museum’s “Xiao Zhuge” and Du Fu Thatched Cottage Museum’s “AI Du Fu,” both of which combine large-scale historical knowledge bases with real-time multilingual interactive capabilities for visitors, reflecting a citywide trend of differentiated digital strategies across Chengdu’s major cultural institutions.

Healthcare : Chinese medical institutions are integrating digital human technology into traditional Chinese medicine (TCM) practice, exemplified by Zhengzhou Guancheng Hospital of Traditional Chinese Medicine (郑州市管城中医院), which operates a “Famous Doctor Digital Human Heritage Studio” alongside a 24-hour AI TCM consultation room and an “AI Brain Platform for TCM” that supports clinical decision-making through data integration. During an inspection visit on August 1, 2026, Chinese Academy of Engineering academician Fan Daiming (樊代明), a leading authority in integrative oncology and president of the World Integrative Oncology Association, toured the digital human studio and praised the platform for enabling “ancient Chinese medical wisdom to display new vitality” through AI-driven data integration and clinical decision support, during a broader visit examining the hospital’s integration of TCM with Western oncology treatment protocols.

02 August News

Marketing : Chinese brands are increasingly deploying AI digital humans as branded spokespersons and always-on marketing assets. Snack maker Chacha Food (洽洽食品) unveiled Chacky (洽可儿), described as the first digital spokesperson in China’s snack food industry, complete with a custom debut single and music video, and roles spanning livestream hosting, e-commerce guidance, customer service, and short dramas. Beauty brand Qiaoyi Zhishui used sensory-focused salon experiences blended with digital content, while marketing agencies Lanse Guangbiao (蓝色光标), Xuanya International (宣亚国际), and Liou Group (利欧股份) were flagged on Chinese exchanges under “virtual digital human” and “AI marketing” investment concepts due to their AI-driven advertising, digital human livestreaming, and metaverse content operations; Xuanya International specifically operates a “Julang Metaverse Joint Laboratory” producing virtual humans and XR livestreaming for corporate clients. In Guangdong, China Unicom’s (中国联通) Guangdong branch partnered with Diancheng district’s agarwood industry association to deploy AI digital human livestreaming assistants alongside 5G networks to support product livestreaming, automated copywriting, and multi-camera broadcast setups for small rural agarwood merchants. In Vietnam-bordering Guangxi, an AI cross-border e-commerce competition showcased a “digital employee” e-commerce platform by Shanghai Humi Technology, integrating chat, accounting, and operational functions with digital human livestreaming for outbound trade with ASEAN markets.

Cultural and Heritage Preservation : AI digital humans are being used extensively to animate and popularize China’s cultural and ethnic heritage. In Xinjiang, the Akesu intangible heritage exhibition in Shihezi incorporated XR immersive experiences and AI digital humans into displays of Zhuang embroidery, Daur folk painting, and other traditions to make heritage more interactive for visitors. In Guangxi’s Baise and Donglan revolutionary base areas, a virtual guide named “Mangguo Ge” (芒果哥) delivers narrated histories of the Red Army’s Baise Uprising to tourists via QR code, supported by a nearby Baise intelligent computing center, while local schools use “5G+AI” platforms for education. Xinjiang’s Bingtuan region deployed a digital human called “Qin Qiaoqiao” (秦瞧瞧) at a heritage exhibition to interact with visitors in Shaanxi dialect and offer cloud-based travel companionship, part of a broader push using 8K, XR, and mobile “non-heritage trucks” to bring traditions to the public. In Yunnan’s Red River Prefecture, tourism authorities unveiled an AI digital human tourism ambassador “Honghe Niuniu” (红河牛牛) to promote seasonal travel themes and cultural attractions including Hani rice terraces and Jianshui ancient town. Hunan’s provincial social science popularization week featured AI digital hosts “Tan Xiaobi” and “Tan Xiaoquan” reading award announcements, alongside immersive light-and-shadow displays and NPC-style costumed performers roaming a themed market. Zhejiang’s overseas Chinese community delegation toured AIGC and digital virtual human demonstration centers in Hangzhou as part of efforts to promote cultural export via short dramas and digital content.

Grassroots Political Communication : Local Chinese Communist Party propaganda efforts have adopted AI digital hosts to modernize ideological outreach. In Haian, Jiangsu, the “Bugu Xiangyin” (布谷乡音) theoretical dissemination initiative introduced digital lecturers named “Xiao Bu” and “Xiao Gu” to conduct interactive, scenario-based explanations of party theory online, with livestreams and short videos reportedly drawing over 100,000 views, complementing a network of human “grassroots celebrity” lecturers trained in partnership with Nanjing Normal University.

Public Services and Government Applications : Chinese public sector bodies are integrating digital humans into citizen-facing services. Xi’an’s traffic police bureau, working with Baidu, launched a digital human named “Yong’an Jun” (永安君) that answers real-time questions about traffic regulations, violation records, and accident procedures, logging over 450,000 interactions and serving more than 200,000 people to date. Yunnan’s Kunming Information Port featured a digital human legal-education segment as part of an anti-fraud and legal-awareness video series. In Guangxi’s revolutionary old areas around Baise, Donglan, and Duan counties, China Mobile’s Guangxi branch has embedded AI large-model tools into hospital diagnostic support and county-level educational platforms alongside its heritage-related digital human deployments.

Corporate Recruitment and HR : Ping An Insurance Group (中国平安) unveiled an AI recruitment digital human named “Chengxiao’an” (橙小安) at a July 31 university-enterprise cooperation conference in Shenzhen, designed to provide 24/7 job recommendations, application status queries, interview tips, and alumni connections for graduating students; the launch was tied to Ping An’s broader 2026 recruitment push offering over 6,000 new graduate positions and partnerships with 42 universities including Tsinghua and Peking University.

Financial Sector Applications : Chinese asset managers are patenting AI-driven interface technology for internal digital workplace tools. E Fund Management (易方达基金) obtained a design patent for an “AI Digital Office” graphical user interface enabling human-machine interaction for operating an AI-powered virtual office environment, reflecting fund management firms’ interest in embedding conversational AI agents into enterprise software design, though the filing indicates internal tooling rather than a public-facing digital human product.

Elder Care and Companionship : A Heilongjiang consumer expo for elder-focused technology showcased an AI-powered digital photo frame by Shenzhen-based Shenying Digital (深颖数字), which converts static photographs and voice recordings of deceased or distant family members into interactive digital humans capable of real-time conversation, medication reminders, and companionship for elderly users; the product is paired with a smart ring for continuous health monitoring including blood oxygen, heart rate, and sleep tracking, and has been demonstrated as addressing emotional isolation among China’s rapidly aging population, particularly in Heilongjiang province where the elderly population share exceeds 27%.

Gaming and Entertainment : At ChinaJoy 2026 in Shanghai, Tencent Games unveiled “Digital Jingdezhen: Porcelain Capital Apprentice” (数字景德镇·瓷都小匠), billed as the world’s first UNESCO heritage-tied mini-game, integrating AI digital human narrators, AI 3D generation, and AI-plus-procedural content generation to recreate historical porcelain-making scenes tied to Jingdezhen’s newly designated World Heritage status; the game was developed jointly with Jingdezhen municipal authorities. Tencent also showcased other AI-native titles including “Cat Immortal Talisman” (猫仙札), which combines AI agents with narrative gameplay, and demonstrated a full AI creation toolchain spanning “Codename Craft” for natural-language game generation, “LAP” for art asset production, and “MagicDawn” and “MagicLomix” for rendering and 3D engine development. Separately, Perfect World’s (完美世界) open-world RPG “Yihuan” (异环) achieved over 1.4 billion RMB in global cumulative revenue since its April launch, with company executives describing AI as an efficiency tool embedded across script generation, digital human video, and 2D-to-3D asset conversion via its proprietary AIGC platform, while emphasizing that AI has not yet displaced core creative and design work. At the same event, Qualcomm’s Snapdragon booth demonstrated on-device AI digital-human generation tools allowing attendees to create real-time 3D “cyber avatars” from a single photo using the fifth-generation Snapdragon 8 Elite platform.

Livestreaming, Short Video, and Regulatory Response : Chinese regulators and industry associations have moved to address controversial digital human and livestreaming practices. In Guangzhou, a newly formed automotive self-media association explicitly prohibited the use of AI-generated digital humans or AI-fabricated content to disparage competing brands, mandating disclosure of AI usage in vehicle reviews. In Chongqing, a newly established Network Performance Branch under the local performing arts association issued a self-discipline pact covering livestreaming conduct, explicitly referencing regulation of “digital human performances” and immersive livestreaming as emerging technology applications requiring oversight alongside bans on sexualized content and fan wars. Separately, a Guizhou tourist attraction’s controversial paid “companion rafting” service, which offered tiered physical-contact packages with hosts, was reported using a “Huiplayxing digital human” (慧播星数字人) for its news video production, illustrating how AI-anchor tools are now routinely used by Chinese media outlets for churning out short-form news content, including in an unrelated report about Taiwanese politician Ko Wen-je’s wife rendering aid to a child on a mainland high-speed train.

Livestreaming Consumer Protection : A Chinese consumer complaint case handled by Jimu News involved a Shandong woman defrauded of 10,000 RMB through a livestreamed “AI operations” course sold via a knowledge-payment livestream; though the underlying fraud was human-perpetrated, the incident highlights the proliferation of AI-training and “AI digital human e-commerce” courses being marketed through Chinese livestream channels, sometimes fraudulently, prompting calls from legal experts for stricter platform oversight of knowledge-based livestream sales.

Enterprise and Industrial Applications : Chinese industrial and technology conferences highlighted digital human integration into broader AI ecosystems. At the Jiangsu AI Industry Alliance launch event in Nanjing, Nanjing-based Aihuashen Technology (爱化身科技) presented digital human and metaverse applications across multiple business scenarios as part of a showcase of AI service providers. Wuxi-based Wuchan Jinlun (物产金轮), a textile and steel wire manufacturer, disclosed a minority stake in Beijing Lingban Instant Intelligent Technology (北京灵伴智能科技), a company specializing in AI voice interaction, AIGC content generation, and digital humans, though the companies stated no current synergy exists between the investment and Wuchan Jinlun’s core manufacturing business.

01 August News

Marketing : Hubei Jingruida Electronic Technology Co., Ltd. (湖北晶锐达电子科技有限公司) is promoting its comprehensive digital human solutions for exhibition hall interactive scenarios, covering digital human design, custom development, and operation and maintenance, with its core products being interactive digital human systems and digital human exhibition hall construction used for welcome guidance, information inquiry, and brand interaction. In a separate development within the snack food industry, Qiaqia Food (洽洽食品) launched its AI digital spokesperson “Chacky” (洽可儿) in Beijing, making it the first digital spokesperson in the leisure snack sector; built with multimodal interaction and emotional perception capabilities, Chacky is designed for e-commerce guidance, live-streaming interactions, and offline displays, and the company plans to extend its use to brand short dramas and flash mob performances. A video report on a controversial rafting service in Guizhou province was noted to have been produced using a Huiboxing (慧播星) digital human, and a similar note about Huiboxing digital human usage was attached to a news video concerning a rape case in Shaanxi province and another report on dismissals at Dongfeng Motor Corporation.

Enterprise : iFlytek (科大讯飞) launched its hyper-realistic digital human real-time generation technology, upgrading its suite of interactive service, marketing, and training digital human solutions, enabling a single photo to generate a digital human capable of real-time interaction, complex actions like fetching objects and dancing, and supporting 7x24 live streaming with consistent identity and smooth multimodal expression coordination. At the 2026 ChinaJoy event, Tencent Games showcased its AI-powered games including “Digital Jingdezhen: Porcelain Capital Young Artisan” which uses AI digital humans and 3D generation, while the Snapdragon Theme Pavilion featured a Digital Avatar Workshop where attendees could generate a 3D digital twin. China Ping An (中国平安) released its campus recruitment AI digital human “Cheng Xiao’an” (橙小安), which offers 7x24 intelligent job recommendations, recruitment progress inquiries, and a new alumni communication function. Kuaidi 100 (快递100) launched the logistics industry’s first full-brand integrated “AI Courier” (AI快递小哥), an AI assistant designed to act as a “digital doppelganger” for couriers, handling customer service, intelligent order processing, and automated reconciliation. The China International Game Developers Conference (CIGDC) featured a presentation from Epic Games on Unreal Engine 5.8, highlighting the Mesh to MetaHuman feature for converting any character into a fully rigged digital human. An A-stock analysis included “virtual digital human” among actively traded market concepts.

Cultural and Tourism : Guizhou University’s Qingniao practice team developed a “Red Army Digital Human” (红军数字人) based on AI large models for Bikeng Village, enabling cross-generational dialogue and serving as a core tool for digital red culture dissemination. Yunnan’s “Folk Song Spectacular” program used a virtual digital host modeled after a real person, Diqing prefecture media professional Yu Shaolong, generated by a team led by Professor Yi Hong from Yunnan University of Finance and Economics, with AI technology used for multi-language broadcasting including Southeast Asian languages. The Honghe Culture and Tourism Bureau (红河州文化和旅游局) officially launched its AI digital human “Honghe Niuniu” (红河牛牛) at a global promotion event in Beijing. Within the smart tourism sector in Jiangsu province, an AI digital human is employed at the Hongshan Site Museum in Wuxi to provide detailed artifact narrations for a “borderless museum” experience.

Public Service and Governance : The Xi’an Public Security Traffic Management Brigade deployed a digital human named “Yong’an Jun” (永安君), the nation’s first intelligent service system in the public security traffic management field to utilize large model technology, which was recently upgraded with the DeepSeek open-source large model for dual-model switching and has answered over 450,000 inquiries since its public beta. The Hunan Women’s Federation introduced an upgraded AI host named “Xiangxiang” (湘湘) at the 2026 Hunan Women’s New Media Conference, who will anchor a series of short videos on topics like policy interpretation and family education under a new engagement plan.

Gaming and Entertainment : Century Huatong (世纪华通) featured the Seasoul digital human platform from Haiyi Interactive Entertainment at its ChinaJoy AI+LAB exhibition area, enabling interactive AI companion functions for original IPs. The AI creative platform AniShort (八点八数字科技), developed by Nanjing-based 8.8 Digital Technology, launched a streamlined workflow for adapting domestic hit short dramas for overseas markets through features like one-click style transfer and face swapping. In the realm of AI-generated content and short dramas, industry analysis noted that the hit AI short drama character Fang Taozi (方桃子) accumulated nearly 400,000 followers on social media within 30 days of her debut, with her persona extending beyond the show into daily vlogs, dance videos, and advertising placements, mirroring the full operational playbook of a real celebrity. The Red Fruit Short Drama platform issued a public notice to regulate AI character creation, aiming to reduce the overuse of “high-frequency AI faces” and standardize character differentiation in AI short dramas. A class action stock analysis report for Guomai Culture (国脉文化) mentioned that the company has developed digital virtual human IPs named “Irene” and “Xiaolin,” based on its general manager and an employee representative.