China Removes 47,000 Posts in Crackdown on AI Celebrity Impersonation

China’s internet regulator says it has removed more than 47,000 pieces of illegal content and penalized 3,961 accounts in a crackdown on the use of AI to impersonate celebrities for profit.

The Cyberspace Administration of China (CAC) announced the figures this week, and the campaign has continued through the period. The targets are accounts that clone a public figure’s face and voice to run fake livestreams, promote products, and drive traffic — a gray market that has boomed as AI face-swapping and voice-cloning tools got cheaper.

The most visible cases have played out in public. On October 18, actress Wu Yue posted on Weibo to condemn an account that used AI to fake her image and voice while selling a facial serum. The account was banned the same day. On October 26, professor Yi Zhongtian refuted AI-generated videos that falsely claimed he had dementia.

The crackdown is part of a broader push. The regulator has separately launched a two-month campaign against health and science rumors, with a particular focus on AI-generated medical misinformation — including cloned voices of experts used to promote supplements.

The pressure is now falling on platforms. They are being told to improve detection of AI-generated content, tighten verification of accounts claiming professional or celebrity status, and remove fakes before they spread rather than after the fact.

For brands and marketers, the signal is clear: AI-generated endorsements using a celebrity’s likeness without authorization are now an explicit enforcement target, not a gray area.