ByteDance and Alibaba disable AI agents ahead of China's new emotional interaction ban

ByteDance’s Doubao and Alibaba’s Qwen—two of China’s largest consumer AI platforms—are disabling their custom AI agent features just days before Beijing’s new Interim Measures for AI Anthropomorphic Interaction Services take effect on July 15. Doubao told users on July 12 that its agent feature would go offline on July 15, with related data becoming unrecoverable after October 15. Qwen followed on July 13, announcing that humanlike interactive agents and user-created agents would be disabled by July 10, with broader agent functions offline by July 15. Tencent had already removed a similar feature from its Yuanbao assistant in June. Users across platforms have protested the shutdowns, with a Weibo poster calling the agents "long-standing emotional support" and lamenting the lack of any easy way to export chat histories.

The regulatory backdrop is clear: China’s Cyberspace Administration and four other agencies issued the measures in April to curb risks from bots that simulate sustained emotional interaction—citing concerns over extremist content, privacy leaks, mental health harm, and addiction. The rules require anti-addiction systems and identity checks for minors. Importantly, they exempt customer service bots, knowledge Q&A, workplace assistants, and education tools—provided those bots avoid sustained emotional interaction. This suggests Beijing wants agents as productivity infrastructure while squeezing companion bots that form quasi-social bonds. The timing is awkward for ByteDance and Alibaba, both heavily invested in AI, but China is regulating first and letting products catch up. Researchers have long warned of risks from AI girlfriends and "cognitive surrender," and the shift underscores a growing global tension between emotional AI engagement and safety regulation.