3 September 2026 News

Marketing : Baidu Yijing (百度一镜) and Kekoukele (可口可乐) jointly created the “AI Fan Zhiyi Real-time Interactive Digital Human” which went live during the 2026 FIFA World Cup in the United States, Canada and Mexico. Users can interact with the digital human through voice or text chat about football via the Kekoukele (可口可乐) China mini-program “Kekoukele (可口可乐) Bar.” The case won the Top Innovative Marketing Impact (AI) award at the 2026 Busan International Advertising Festival, becoming the first Chinese real-time interactive digital human marketing solution recognized by the international advertising creative industry. Technically, it achieved response latency within 2 seconds, commercial deployment of the first dialect-level IP voice cloning, real-time synchronization of competition data leveraging RAG mechanisms, and resolved the stiffness problem of traditional virtual human interaction through a four-dimensional proactive dialogue control system. Reports from Frost & Sullivan and IDC show Baidu Yijing (百度一镜) has ranked first domestically in both AI digital human market share and comprehensive product strength for two consecutive years. Previously, Luo Yonghao’s digital human livestream achieved 55 million yuan GMV in 6 hours. Anta Jituan (安踏集团) released the “AI 365 Strategy,” with digital human livestreaming as one of the implementation results in retail operations. Throughout 2025, the group’s digital humans accumulated over 100,000 hours of livestreaming, exceeded 30 million views, and contributed over 300 million yuan in GMV (a significant increase from 120 million yuan in 2024). Digital humans contributed approximately 6% of business on shelf e-commerce during night hours (0 to 7 a.m.), and achieved 24-hour cross-timezone localized livestreaming across over 20 stores in six Southeast Asian countries. By early 2026, over 80 stores were operating digital human livestreams, with the Tmall FILA FUSION official flagship store becoming an industry benchmark case. Digital human assets were also migrated to short dramas and livestream scenarios for coordinated operations by brands such as La Mer. An investigation into the gray industry of virtual artists revealed numerous “idols that never collapse” accounts proliferating on social platforms. For example, virtual artist “Jin Daoyu K-Dowoo” operator “AIGC Tengda Yunying (AIGC腾达运营)” claimed to generate virtual human “experience” promotional videos of beauty devices without needing to send samples, only using product images. In this case, the contact lens brand “Jianguo Li (坚果力)” (Anhui Huaxin Qiye Guanli (安徽花信企业管理有限公司)) faced intervention from market supervision authorities due to the virtual idol “Fang Taozi” posting non-compliant experience-type advertisements. Virtual human “Mengke·Cora” under Qianhang Wenhua (千珩文化) charges 70,000 yuan per product placement video, while “Baibian NORA” charges 55,000 yuan. Regarding virtual human “Liu Yexi,” CEO Liang Zikang of the operator Chuangyi Technology stated that brands value the controllable persona of virtual IP and the sense of safety in placements without collapse risks.

Cultural Heritage : The dance drama “The Thousand-Li Tea Road” AI derivative series was created by executive director Meng Yan and her team, transforming the lead performers into AI digital avatars and reorganizing the narrative while continuing the cultural core of the original production. The first three episodes will premiere on Migu Culture Technology platform, aiming to convert the physical narrative on stage into more plot-driven audiovisual expression, building a distribution pathway of “online understanding, offline experience, and post-viewing extension” for audiences, while also opening up a new artistic space for actors in digital character co-creation. Shanghaishi Zhishichanquanju (上海市知识产权局) issued an intellectual property registration certificate for the “Luo Tianyi audio dataset” to Shanghai Bili Bili Keji (上海哔哩哔哩科技有限公司), marking the first time a completely virtual human audio data product has obtained intellectual property confirmation, providing a practical model for the protection and rights defense of virtual singer digital assets.

Public Sector and Governance : The 2026 National Cybersecurity Publicity Week will be held in Jinan, Shandong from September 14-20, featuring a Cybersecurity Technology Application Experience Zone (September 13-17, Shandong International Convention and Exhibition Center) that will organize interactive experiences such as “Cybersecurity Clinic” and “AI-Replicated Digital Humans,” promoting the inclusive sharing of cybersecurity technology achievements through physical displays and human-machine interaction. This activity is jointly hosted by Zhongyangxuanchuanbu (中央宣传部), Cyberspace Administration of China and nine other departments. The new national standard GB/T 47699-2026 in the blockchain field officially takes effect on September 1, for the first time incorporating “virtual digital humans” and “AI agents” into a comprehensive OID object identifier encoding system, reserving a standardized framework for national-level identity identification of virtual objects. Pilot cases include the “Jia Qi Xin Rong” commemorative digital collectible issued by Zhongguogongshangyinhang (中国工商银行)’s private banking division, which has already adopted this identification system to achieve on-chain confirmation of rights and cross-platform recognition. Heping District in Shenyang has integrated the “Smart Heping” agent into multiple livelihood and industrial activities in the jurisdiction, including at the fall 2026 opening ceremony of Heping No. 1 Primary School, where the digital human “He Zhi Xuan” welcomes students at the school entrance and sets up a science and innovation experience zone. The agent also sets up an offline experience point at the Shenyang Artificial Intelligence Application Supply and Demand Docking Conference for attendees to interact with the digital human “He Zhi Xuan.” Zhao Yongming, director of Heping District’s Industrial and Information Technology Bureau, positions it as an important practical achievement in implementing the “AI Plus” initiative. Wuhan is distributing “Wuhan Qingnian Daxue Xinsheng Zhinengka (武汉青年大学新生智能卡)” to all 2026 cohort university freshmen arriving in the city, which is the nation’s first personally exclusive urban intelligent agent featuring the finless porpoise mascot as a virtual image, integrating eight intelligent agent capabilities including campus guidance, multilingual city touring, academic planning, mock interviews, and startup policy services, aimed at reconstructing the connection between youth and the city of Wuhan through digital-intelligent means. Chengdu Qizhong Chuzhong Fushuxiaoxue (成都七中初中附属小学) relied on artificial intelligence technology at the opening ceremony to accurately restore virtual images of revolutionary predecessors, building an AI temporal-spatial dialogue platform spanning a century, allowing students to engage in immersive question-and-answer interactions with virtual predecessors, breaking the static presentation mode of traditional red education. Zhengzhou Hangkonggang Keji Yijie Xiaoxue (郑州航空港科技一街小学) launched an AI interactive terminal named “Haoqi Chuang (好奇窗)” in the September 2026 opening season, with built-in virtual images covering AI-personified characters of more than ten domestic and foreign scientists and cultural celebrities including Einstein, Li Bai, Lu Ban, and Leonardo da Vinci, leveraging the multimodal large language model “Qimeng · Xinghe (启梦·星核)” to respond to students’ cross-disciplinary questions in a Socratic questioning manner, and designing cognitive buffering mechanisms such as “three-second pauses” to guide deep thinking.

Governance and Ethical Contexts : Cyberspace Administration of China announced the second-phase progress of the “Clear the Air: Remedying AI Application Chaos” special campaign, cumulatively removing over 5.61 million pieces of illegal and non-compliant information, investigating and handling over 49,000 accounts, and disposing of over 2,400 non-compliant websites and applications, naming Doubao (豆包), Yuanbao (元宝), Qianwen (千问), Wenxin Yi Yan (文心一言) and other platforms to strengthen source material review and enforce identification of AI-generated synthetic content implementation. Typical cases reported include seven categories of problems such as using AI “face-swapping” and “voice-swapping” to impersonate public figures and generate vulgar content for profit, and creating AI digital humans for false advertising and goods promotion that mislead elderly people. Nanfang Dushibao (南方都市报) investigation revealed the gray industry chain of virtual artists, with virtual artist operators (such as MCN institutions including “AIGC Tengda Yunying (AIGC腾达运营)”, “Xiaoyu Hudong (小雨互动)”, “Qianhang Wenhua (千珩文化)”) knowingly violating platform prohibitions on AI digital humans demonstrating product usage effects by still promoting goods through “underwater transactions” with virtual artists conducting “experiential” promotions. The involved virtual artist account “I’m Really AI” released videos containing advertisements for prescription drug “Aishi Dakou Rongmo (艾时达口溶膜)” treating male erectile dysfunction. Virtual idol “Fang Taozi” sparked controversy by promoting contact lens product “Jianguo Li (坚果力)”, with the company Anhui Huaxin Qiye Guanli (安徽花信企业管理有限公司) to which the product belongs already under investigation by local market supervision departments. Chen Zekun, lawyer at Guangdong Legal Shengbang Law Firm, pointed out that virtual humans lack civil subject capacity but operators and related subjects cannot be exempted from legal responsibility. Tian Feng, director of the Kuaisi Manxiang Research Institute, believes platform penalties for virtual artist unauthorized advertising are “account-level” while illegal gains are “advertiser-level”, creating a responsibility mismatch issue. AI short drama face-theft and voice-theft incidents occur frequently. Voice actor Ma Zhengyang issued a statement claiming popular drama “Nine Gates” spin-off drama and inserted advertisements allegedly misappropriated his voice. The involved production company “Chengdu Wanhou (成都顽猴)” responded that the advertisement used Ji Meng AI (即梦AI) tools for automatic generation and failed to determine the AI voice source, resulting in similarity to his voice. Fudandaxue Kexin Jushen Zhineng Yanjiuyuan (复旦大学可信具身智能研究院) junior researcher Ma Xingjun and Zhongguokejidaxue Wangluoanquan Xueyuan (中国科技大学网络安全学院) vice-president Zhang Weiming pointed out that AI foundation model training materials may capture publicly available network photos, videos, recordings and other unauthorized content, and biometric features such as gait and iris that are difficult to change may also be incorporated into AI training use in the future, requiring multi-party collaborative governance from legislation, judiciary, regulation, technology and industry self-discipline. Actress Qi Wei sparked “selling faces” controversy by authorizing her digital alter ego “Han Qingxia” to star in an AI manga series. She responded emphasizing this is a permitted use with “time limits, scope limits, review and accountability” rather than rights transfer, and reminded colleagues not to easily sign permanent unlimited scope sub-authorization contracts for a single work opportunity. The article also mentioned the phenomenon of ordinary people “selling faces” for monetization, such as a bank employee being purchased 7 times in 3 days through a facial authorization platform, earning 3,500 yuan in income. The facial image authorization trading platform built by Shanghai Puguangxing Yi Wenhua Keji Gongsi (上海浦光星奕文化科技公司) has over 1,000 people registered. Intellectual property lawyer You Yunting reminded that after authorization, personal facial usage scope is difficult to fully control, carrying risks of difficult rights protection and evidence collection. Cyberspace Administration of China released in April the “Digital Virtual Human Information Service Management Measures (Draft for Public Comment)” requiring digital virtual human service providers not to engage in false advertising, malicious consumer inducement and other illegal activities, and must continuously display conspicuous identifying labels with the words “digital human” in display areas. The child companion robot market boom has raised concerns over minor protection. Article 14 of the “Temporary Measures for Managing AI Anthropomorphic Interactive Services” explicitly stipulates that providers of anthropomorphic interactive services must not provide virtual relatives, virtual companions and other virtual intimate relationship services to minors, yet reporter investigation found some merchants circumvent regulatory red lines through “setting virtual humans like parents” edge-case designs. Meanwhile, the measures require obtaining guardian consent before providing other anthropomorphic interactive services to minors under fourteen years old. China saw an AI matchmaking fraud case, with Chengdu police busting a 28-member fraud gang operating under the name “Sichuan Wanhe Xitang Marriage Service Co., Ltd.”, using AI-generated false images of “perfect matchmaking prospects” (such as a self-identified 67-year-old “Li Ge” with annual income of 6 million yuan) specifically to defraud middle-aged and elderly women into paying service fees. The gang defrauded over 1 million yuan in less than a year, with 28 suspects apprehended.