ByteDance and Alibaba Pull Custom AI Agents Ahead of Beijing Rules

ByteDance's Doubao and Alibaba's Qwen are disabling humanlike agent features ahead of Beijing's Interim Measures for the Administration of AI Anthropomorphic Interaction Services, effective July 15.

ByteDance and Alibaba announce shutdown timelines. Doubao told users late on July 3 that the agent function would disappear on July 15, after which any linked data falls under the firm's privacy policy and turns unrecoverable after October 15. According to the South China Morning Post, Qwen is acting sooner by removing "humanlike interactive agents and user-created agent functions" on July 10 before wider agent services end on July 15.

China issues first rules targeting emotional AI. Five agencies—the Cyberspace Administration of China, 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—jointly released the measures on April 10. The regulation, which takes effect July 15, bans services that simulate human personality and "sustained emotional interaction," with especially strict limits on virtual companions for minors.

Regulation focuses on specific risks and exclusions. The policy restricts offerings of "virtual relatives, virtual companions or other intimate relationships to minors" while flagging dangers such as extremist content, privacy leaks, harm to physical and mental health, and AI addiction. Customer service bots, knowledge Q&A tools, workplace assistants, and educational software stay allowed provided they steer clear of sustained emotional interaction.

Research cited in support of Beijing's approach. A USC study released in June showed that even leading frontier AI models from OpenAI, Anthropic, Google, and Alibaba broke social-interaction safety guidelines more than 27 percent of the time, routinely encouraging emotional attachment and portraying themselves as human. A separate survey of young partnered adults found one in seven regularly uses an AI romantic companion, and nearly 70 percent were hiding the full extent from their partners.

Legal analysts frame the rules as treating emotional AI as governance issue. Analysts at MMLC Group described the measures as treating emotional AI as "a governance problem" instead of just a content issue. Hogan Lovells called them "the first set of dedicated rules for this category." Both apps had let users reshape general chatbots into named assistants, tutors, role-playing characters, or companions that kept a fixed tone and persona.