China's Cyberspace Administration just dropped draft regulations that would ban AI-powered "virtual intimate relationships" for anyone under 18. The rules also require prominent labelling on all digital human content, ban creating deepfakes of real people without consent, and prohibit using virtual humans to bypass identity verification.

Public comment is open until May 6. These will almost certainly become law.

Now here is the part that should make Western policymakers uncomfortable. Character.AI, Replika, and dozens of other AI companion apps are available to American teenagers right now with essentially zero regulatory guardrails. One of them already faces lawsuits after a teenager died by suicide. The US response? Congressional hearings. State-level patchwork. No federal rules.

China just wrote a comprehensive framework in the time it takes Congress to schedule a subcommittee meeting.

The draft regulations go further than just protecting kids. Service providers must intervene and provide professional assistance when users exhibit suicidal or self-harming tendencies. Digital humans cannot disseminate content that "endangers national security" or "undermines national unity." Providers must prevent sexually suggestive content, depictions of horror or cruelty, and content inciting discrimination.

Yes, some of those provisions are clearly about political control. The national security and unity clauses are classic Beijing censorship tools dressed up as AI governance. But the child protection measures are genuinely ahead of anything the West has implemented. That is the uncomfortable truth about Chinese AI regulation: it moves fast, it sometimes gets things right, and it often bundles good policy with authoritarian control.

The digital human market is exploding globally. AI companions, virtual influencers, deepfake customer service agents, synthetic celebrities. Every one of these applications raises questions that Western governments have not answered: When does an AI companion become emotionally manipulative? Who is liable when a virtual human causes harm? Should children be allowed to form parasocial relationships with software that is designed to be addictive?

Beijing answered those questions this week. Washington is still asking them.

First reported by Reuters. Draft regulations published by the Cyberspace Administration of China.