ByteDance and Alibaba are disabling key features of their AI chatbots in China that allowed users to create their own human-like AI companions and chat with them. The move is a direct response to new regulations from the Chinese government that will take effect on July 15, 2026.
Doubao, Qwen, and Yuanbao shut down
As reported by the South China Morning Post, ByteDance informed users of Doubao – China’s most used chatbot with over 300 million monthly users – on Saturday night that the agent feature would go offline on July 15 due to “product function adjustments.” From October 15, related data would then be handled according to the company’s privacy policy and would no longer be viewable or recoverable in the app. Alibaba’s Qwen announced the same weekend that it would disable its “human-like interactive agents and user-created agent features” as early as July 10, with further agent features and services following on July 15. Tencent’s Yuanbao had already discontinued similar features in June. All three services had previously allowed users to turn a general chatbot into a named assistant, learning companion, role-playing character, or companion with a fixed personality and tone.
The new Chinese regulation in detail
The basis for the shutdowns is the “Interim Measures for the Administration of Artificial Intelligence Anthropomorphic Interaction Services” jointly issued by five Chinese authorities in April 2026, primarily published by the Cyberspace Administration of China (CAC) together with the National Development and Reform Commission, the Ministry of Industry and Information Technology, the Ministry of Public Security, and the State Administration for Market Regulation. According to SCMP, the guidelines affect AI services that “simulate human personality traits, thought patterns, and communication styles to enable sustained emotional interaction.” Pure customer service bots, knowledge Q&A systems, work assistants, and educational and research tools are explicitly exempted, provided they do not offer sustained emotional interaction. Among the risks addressed by the authorities are extremist ideas, privacy violations, harm to physical and mental health, as well as dependency and addiction.
A global pattern: California and the USA
Similar regulatory approaches are now also emerging outside of China. In California, the already signed Senate Bill 243, which has been in effect since January 1, 2026, requires providers of companion chatbots to clearly inform minors about the AI nature of the conversation, maintain protective protocols against suicidal content, and refer affected users to crisis facilities. The law also provides for a private right of action for affected individuals.
Regulatory pressure in the USA is also fueled by specific court cases: Google and Character.AI reached an out-of-court settlement in January 2026, according to CNN, in several lawsuits in which families accused the companies of contributing to the suicide or severe mental harm of minors through their chatbots – without the companies admitting any wrongdoing. OpenAI is also facing similar lawsuits in which families establish a connection between the use of ChatGPT and the suicide of teenagers.
Classification
That China is the first country to implement nationwide binding regulations for human-like AI interaction may be surprising, but it fits into an internationally observable pattern: The more AI chatbots position themselves as emotional companions rather than mere tools, the clearer the regulatory and legal pressure becomes on their providers. While Beijing relies on a centrally issued regulation with clear deadlines, a comparable effect in the USA is currently emerging more through state laws like California’s SB 243 and through court cases, the outcomes of which create immediate costs and reputational risks for the companies involved.
