China Regulates AI Digital Humans: What's Changing
China Regulates AI Digital Humans
China published comprehensive draft rules on April 3, 2026 targeting AI “digital humans” — virtual AI entities that simulate human personality and engage in emotional interaction. This is the first major regulatory framework for AI companions and avatars globally.
Last verified: April 2026
Key Requirements
| Rule | Details |
|---|---|
| Mandatory labeling | All digital human content must have prominent “digital human” labels |
| Minor protection | No virtual intimate relationships for users under 18 |
| Anti-addiction | Ban on addictive AI services for children |
| Data consent | Strict safeguards on personal data collection |
| Provider registration | AI digital human providers must register with regulators |
| Real identity | Services must verify user identity for age-gating |
What’s Covered
The rules apply to AI that:
- Simulates human personality traits
- Engages in emotional interaction through text, images, audio, or video
- Provides products or services to the public in mainland China
This includes:
- AI companions (Grok companions, Character.AI, etc.)
- Virtual influencers (AI-generated social media personalities)
- Customer service avatars
- AI chatbot personalities
- Digital streamers (virtual hosts on platforms like Douyin)
Why This Matters Globally
First-Mover Regulation
China is the first major economy to propose comprehensive rules specifically for AI digital humans. Europe’s AI Act touches on this but doesn’t have dedicated digital human regulations.
Impact on Global Companies
Any AI company operating in China must comply — this affects:
- OpenAI/ChatGPT (if operating in China)
- Grok companions (xAI)
- Character.AI and similar services
- Enterprise AI avatar products
Template for Other Countries
China’s AI regulations often become templates for other Asian markets and influence European and US policy discussions.
The Child Safety Focus
The strongest provisions protect minors:
- No virtual intimate relationships for under-18 users
- No addictive AI services designed for children
- Age verification required
- Parental controls mandated
This directly responds to growing concerns about children forming emotional attachments to AI chatbots — a trend that’s accelerated globally in 2025-2026.
Timeline
- April 3, 2026 — Draft rules published
- May 6, 2026 — Comment period ends
- Mid-2026 — Final rules expected
Last verified: April 2026