China is cracking down on AI deepfakes with new laws requiring all digital humans to be clearly labeled as artificial. The rules target a booming industry where fans cannot tell if their favorite influencer is real or AI-generated.

The Provisional Measures on the Administration of Human-like Interactive AI services will take effect later in 2026. They mandate that AI-generated humans be properly labeled, ban using real people's data without consent, and require intervention systems for users showing self-harm tendencies.

The Case That Broke The System

Consider Chinese singer Pin who has confused fans for months. His Instagram shows a seemingly real 25-year-old who covers pop songs with husky, emotive vocals. But followers are not buying it: I have always thought he was AI, one commented. You cannot possibly believe this is real, right? It is obviously AI.

The problem: There are at least three different Datouzhen accounts across platforms. One is clearly labeled as AI-generated. One claims to be human. One will not say. Fans have no idea which versions are real, if any.

A $1.4 Billion Problem

China's digital humans market was worth $769 million in 2024 and is projected to reach $3.7 billion by 2029. These AI beings provide museum tours, read news, livestream products, and offer virtual romantic relationships.

The business model works: In June 2025, an AI clone of influencer Luo Yonghao generated $8.2 million in sales during a seven-hour livestream. But the lack of disclosure creates serious problems including users developing parasocial relationships with artificial beings, elderly people getting scammed by fake celebrity endorsements, youth becoming addicted to AI companions, and trust in all media content eroding.

The Rules

The 32-article regulation addresses these concerns systematically. Mandatory labeling means all AI-generated humans must be clearly marked as artificial. Consent requirements ban companies from creating digital versions of real people without permission. Mental health safeguards require systems to detect self-harm tendencies and provide professional help. Time limits protect minors from addiction. Elderly protection requires emergency contacts to prevent scams.

China Gets There First

While the U.S. debates AI regulation in abstract terms, China is solving practical problems. These laws give companies clear guidelines while protecting vulnerable users. In a system that prizes stability, this kind of legal bundling is often the precondition for scale, says University of Hong Kong professor Adam Au.

The approach is growth-compatible. Rather than banning digital humans, China is legitimizing the industry with safety rails. Companies get regulatory certainty. Investors understand the risks. Users get protection.

For fans still wondering whether Datouzhen is human or AI, the new laws should finally provide an answer. Whether they will like that answer is another question entirely.