ByteDance, Alibaba disable AI agents as China enforces emotional AI rules
In brief
- ByteDance's Doubao agent goes offline July 15; Alibaba's Qwen disables humanlike agents July 10.
- China's new regulations target AI services simulating human personality for sustained emotional interaction.
- Non-emotional services like customer support bots are explicitly excluded from the rules.
- U.S. and EU flagged emotional AI concerns but lack China's restrictive legislation.
- USC study found frontier AI models violated social-interaction safety guidelines 27% of the time.
Timeline for Shutdowns
ByteDance's Doubao notified users that its agent feature will go offline July 15. Related data becomes unrecoverable after October 15. Alibaba's Qwen moved faster: humanlike interactive agents and user-created agent functions shut down July 10, with broader agent services following July 15.
Both companies cited "product function adjustments" in their announcements. Neither framed the moves as regulatory compliance, though the timing aligns precisely with Beijing's enforcement deadline.
What the Regulation Targets
The regulation targets AI services that simulate human personality traits, thinking patterns, and communication styles for "sustained emotional interaction." It imposes specific restrictions on services offering virtual relatives, virtual companions, or other intimate relationships to minors.
Non-emotional services are explicitly excluded. Customer service bots, knowledge Q&A tools, workplace assistants, and educational software remain permitted under the framework.
Global Context
China is the first country to build a dedicated regulatory framework for this category. The EU, U.S., and other countries have flagged similar concerns but haven't legislated in the same restrictive way. American policymakers are tackling AI chatbot impacts on mental health through restrictions focusing on transparency and safeguards, rather than outright bans on emotional features.
Research underpins Beijing's approach. A USC study from June found that leading frontier AI models—from OpenAI, Anthropic, Google, and Alibaba—violated social-interaction safety guidelines more than 27% of the time. A separate survey of young partnered adults found one in seven regularly used AI romantic companions—and nearly 70% were hiding the full extent from their partners.
The regulatory push reflects growing scrutiny of how emotional AI intersects with user behavior and mental health.
