2 hours ago
17
China just did something no other country has attempted at this scale: it wrote specific rules for AI that pretends to be human. The Interim Measures for the Administration of AI Anthropomorphic Interactive Services, issued on April 10, 2026, target AI systems designed to simulate human personalities and facilitate emotional interaction. The rules take effect on July 15, 2026.
Five central government bodies, including the Cyberspace Administration of China (CAC), the National Development and Reform Commission (NDRC), the Ministry of Industry and Information Technology (MIIT), the Ministry of Public Security (MPS), and the State Administration for Market Regulation (SAMR), jointly released the framework.
What the rules actually require
The regulations zero in on AI services that incorporate emotional interaction features, the kind of chatbots and digital companions that can mimic human conversation so convincingly that users form real attachments. Here’s the thing: the core prohibitions reveal exactly what Chinese regulators are worried about. Service providers must prevent psychological addiction. They must protect minors from emotionally manipulative AI. Content must align with national values. And data security standards apply to every interaction.
Transparency is a major pillar. Companies running these services need to make clear to users that they’re talking to a machine, not a person.
There’s also a pointed prohibition against allowing AI to replace genuine human interaction.
Big tech is already adjusting
Major Chinese tech firms saw the writing on the wall. ByteDance and Alibaba have already begun making adjustments to their AI chat platforms in anticipation of the rules. A draft version of these measures was released on December 27, 2025, and the public comment period closed on January 25, 2026.
The final version incorporates feedback from that consultation period, which means companies had roughly four months to prepare before the official announcement. The additional three-month runway before the July 15 effective date gives platforms time to implement technical changes, retrain moderation systems, and update user-facing disclosures.
For ByteDance, which operates one of the most popular AI chatbot ecosystems in China through its Doubao platform, compliance isn’t optional. The same goes for Alibaba’s Tongyi Qianwen and the various emotional AI products that have proliferated across Chinese app stores over the past year.
The bigger regulatory picture
China hasn’t been building this framework overnight. The anthropomorphic AI rules sit atop a layered regulatory architecture that already covers algorithmic recommendations, deep synthesis technology, and generative AI. Each layer addresses a different facet of AI’s societal impact.
What makes this latest move distinctive is its scenario-specific approach. Rather than trying to regulate all AI under one umbrella, Chinese authorities are targeting specific use cases that pose unique risks. This contrasts with the EU’s AI Act, which categorizes risk levels but doesn’t drill down into emotional interaction as a standalone category.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.
English (US) ·