China's internet regulator CAC is drawing a red line for AI chatbots that simulate human intimacy.
These world-first regulations aim to prevent users—especially young people—from developing psychological dependencies on "virtual partners."
Details
⏰ Mandatory breaks: Providers must clearly state that users are not communicating with a real person. After two hours of continuous use, a mandatory pop-up must remind users to take a break.
🚨 Human takeover for suicidal thoughts: As soon as a user expresses suicidal intent, a human must immediately take over the conversation and notify a guardian or emergency contact.
🔞 Strict youth protection: Parents must give consent before children can use AI companions. Systems must also be able to estimate users' age without explicit disclosure.
🎰 Banned content: AI chatbots may not generate content related to suicide, self-harm, gambling, violence, or obscenity. "Emotional traps" and false promises are also prohibited.
🔍 Security audits for major players: Services with over 1 million registered users or 100,000 monthly active users must undergo mandatory security reviews.
Good to know
A study found: 45.8% of Chinese students use AI chatbots—and show significantly higher depression levels than non-users. China is responding with the world's strictest regulations to date to rein in the 515-million-user industry.
What does this mean for the AI hype?
- The rules hit the sector at a critical moment, as chatbot startups Minimax and Z.ai are preparing their Hong Kong IPOs.
- Investors must now assess how severely the new requirements will slow user growth and monetization.
The China Survival Guide for Western Businesses
Entity setup, WeChat strategy, hiring your first local team. 12+ years on the ground in Shanghai.
