China issues draft rules for human-like AI: new controls

What happened On December 27, 2025, the Cyberspace Administration of China (CAC) published a draft regulation titled 人工智能拟人化互动服务管理暂行办法（征求意见稿） to govern AI systems that mimic human interaction. The draft requires operators to apply security measures, complete ethical reviews, ensure content and behavior align with “core socialist values,” and notify users when they are interacting with an AI system. (Reported by Reuters, Channel NewsAsia, The Straits Times and NBD.)

Key details The rules cover human‑like interactive services and set obligations for platform operators and service providers. Providers must clearly label AI interactions, run pre‑deployment reviews, and keep security and accountability procedures in place. The draft is open for public comment, indicating a consultation phase before any final rule. Multiple outlets reporting on the draft cite the same core requirements and the CAC’s request for public input.

Implications If finalized, the measures will increase compliance demands for vendors offering conversational or anthropomorphic AI in China. Organizations should track the consultation, prepare disclosure and audit capabilities, and implement content‑alignment and ethics review workflows to meet potential new obligations.

Why It Matters

• Platforms must implement labeling and logging to meet disclosure and accountability requirements.
• Dev teams should establish ethics‑review and security‑testing workflows before deployment in China to reduce compliance risk.
• Foreign vendors need localization and content‑alignment controls to conform with the draft’s 'core socialist values' requirement.
• The public consultation could change technical and legal obligations; monitor the process and submit comments if relevant.