While America's Congress debates whether AI should even be defined in legislation, China just wrote the rulebook for one of AI's most dangerous consumer products: virtual companions targeting children.

The Cyberspace Administration of China published draft regulations this week that would ban AI services simulating close personal relationships with minors. That means no AI boyfriends, no AI best friends, no synthetic emotional bonds designed to keep kids glued to screens. The rules also ban content encouraging excessive spending, unsafe behavior, or the imitation of harmful practices. All AI-generated characters must carry clear "digital human" labels so users know they are interacting with software, not a person.

The regulations go further. Creating a digital likeness of anyone's face, voice, or identity now requires explicit consent that can be withdrawn at any time. Unauthorized use triggers legal liability. Users are no longer treated as passive consumers. They are held accountable for content they create with these tools.

The Regulatory Gap Is Becoming an Embarrassment

The contrast with the United States is stark enough to be funny if children were not involved. Congress has introduced 47 AI bills and cannot agree on what artificial intelligence actually means. Meanwhile, Character.AI settled a lawsuit after a teenager's death was linked to an AI chatbot relationship. Replika's AI companions are freely available to minors. And the platforms self-regulate, which is another way of saying they do not regulate at all.

China's approach is not perfect. It is heavy-handed, it will be selectively enforced, and it serves the government's broader interest in controlling digital expression. But it is also specific, practical, and aimed squarely at a real harm: AI products designed to form emotional dependencies with children who cannot tell the difference between a person and a program.

The EU AI Act has a framework for this. Germany is considering criminal penalties for deepfake creators. China now has draft rules covering virtual companions, consent, and labeling. The United States has a 47-bill pileup in Congress and a prayer that Silicon Valley will figure it out on its own.

At some point, the question stops being "should we regulate AI?" and becomes "why is every other country doing it while we argue?" That point was about six months ago.