What happened
Five Chinese central agencies (CAC, NDRC, MIIT, Ministry of Public Security, SAMR) jointly issued the 'Interim Measures for the Administration of AI Anthropomorphic Interaction Services' (issued April 10, 2026), which take effect July 15, 2026 — just after this reporting window. The rules ban virtual intimate/companion services for minors, mandate age verification, a 'minor mode,' anti-addiction measures, AI-disclosure requirements, and safety-assessment filings for providers exceeding 1 million registered users. ByteDance and Alibaba have already begun disabling affected persona/companion features in advance of the deadline.
Why it matters
This is a binding, imminently-effective rule directly regulating a fast-growing category of consumer AI products (companion chatbots/agents) with extraterritorial ambiguity for foreign providers serving PRC users, and sets a global precedent for scenario-specific (rather than general-purpose) AI regulation targeting emotional/anthropomorphic AI risks — an approach distinct from the EU AI Act's risk-tier model.
Action needed
AI companion/chatbot providers with PRC user bases should complete safety assessments and filings with provincial cyberspace administrations before July 15, 2026; implement age verification and minor-mode controls.