# China’s Companion-AI Law: Emotional Chatbots Get Their Own Bucket

Times of AI Desk · 2026-07-15 · Policy

[https://timesof.ai/2026/07/china-anthropomorphic-ai-rules-take-effect](https://timesof.ai/2026/07/china-anthropomorphic-ai-rules-take-effect)

> China’s anthropomorphic AI interaction rules took effect July 15—targeting continuous emotional companions with youth and disclosure duties, while exempting ordinary customer service, Q&A, and work tools. Product taxonomy, not a blanket generative-AI ban.

If your UX optimizes for **relationship and affect**, you are in a different legal bucket than if you optimize for **tasks**. China’s July 15 effective date makes that taxonomy hard law for companion products—while productivity and research agents stay under older generative-AI and content rules.

China’s **Interim Measures for the Administration of Artificial Intelligence Anthropomorphic Interaction Services** came into force **July 15, 2026**. Issued **April 10, 2026** by multiple central authorities (including the **Cyberspace Administration of China**, **NDRC**, **MIIT**, **Ministry of Public Security**, and **SAMR**), the rules target **continuous emotional interaction** services that simulate a natural person’s personality, thinking, and communication style. Ordinary customer service, knowledge Q&A, work assistants, education, and research tools **without** continuous emotional interaction are **out of scope**.

## What the regime covers

From the published measures and English-language legal/tech analyses (China Law Translate, Geopolitechs, compliance briefs):

| Theme | Direction |
|-------|-----------|
| **In scope** | Public-facing services in China that deliver **sustained emotional / anthropomorphic** interaction via AI |
| **Out of scope** | Intelligent customer service, knowledge Q&A, work assistants, education/learning, scientific research **without** continuous emotional interaction |
| **Youth** | Heightened duties: age recognition, guardian consent patterns, usage limits, bans on designs that replace social interaction for minors (core draft/final themes) |
| **Content & security** | Prohibitions aligned with broader Chinese content law; provider responsibility for interactive content |
| **User disclosure** | Human-like agents expected to avoid deceptive “I am a real person” framing; psychological-dependence risks addressed in design duties |
| **Effective date** | **July 15, 2026** (Article 32 of the measures) |

The package is one of the world’s most **product-specific** AI rules: not “all generative AI,” but the **companion** surface regulators associate with mental health, addiction, and minor safety—after high-visibility companion apps in the Chinese market.

## Limits

- English summaries and legal translations are secondary to the PRC text; cite measures for compliance work.
- Exact enforcement practice (fines, takedowns) will show in post-effective cases, not the day-one statute alone.
- Out-of-scope productivity tools can still fall under older generative-AI and content regimes.

## Sources

- PRC multi-agency *Interim Measures for the Management/Administration of Anthropomorphic AI Interaction Services* (issued April 10, 2026; effective July 15, 2026).
- Geopolitechs / China Law Translate English analyses of the measures and draft history.
- Compliance summaries noting minor protections and emotional-interaction scope (e.g., Licentium, ITTC Network briefs on the July 15 effective date).