China Clamps Down on AI Companions

China’s newly introduced “Interim Measures for the Administration of AI Anthropomorphic Interactive Services” officially took effect in July 2026, prompting major domestic tech companies to curb or shut down emotional companion features in their chatbot platforms. Major applications, including ByteDance’s Doubao and Alibaba’s Qwen, deactivated or altered user-created AI agent tools in the days leading up to the compliance deadline. The regulations establish strict operational boundaries designed to mitigate psychological dependency, mandate anti-addiction notifications, forbid virtual romantic companion services for minors, and require automated detection and intervention mechanisms for users showing severe mental distress.

As the world’s first comprehensive national framework targeting anthropomorphic AI interactions, the rules aim to address emerging societal risks associated with emotional AI engagement, such as algorithmic manipulation, privacy vulnerabilities, and acute psychological impacts. While the regulatory framework sets clear compliance standards to protect vulnerable groups, the sudden withdrawal of personalized AI companions has sparked widespread discussion among users who had come to rely on these virtual tools for daily companionship.