Deepfake scams impersonating celebrities spark calls for AI regulation in China
Deepfake videos impersonating Chinese actor 靳东 and others have been used in scams targeting elderly people and the general public in China. The actor and tech entrepreneur 雷军 raised the issue at 2025 parliamentary sessions, calling for legislation to regulate AI face-swap and voice-synthesis technology used to deceive people.
What happened
Deepfake videos impersonating actor 靳东 (Jing Dong) have been used to scam people online. In November 2024, an elderly woman in Yichun, Jiangxi, was nearly defrauded into applying for a 2 million yuan loan after viewing an AI deepfake video of the actor at a bank. In 2020, a 61-year-old woman in Jiangxi left home based on an online romance with a "fake 靳东."
Actor 靳东 stated on March 5, 2025, at a Chinese People's Political Consultative Conference (CPPCC) group discussion that audiences of his films have been "scammed very badly" by AI deepfake videos, calling the behavior "extremely heinous" and urging the establishment of better rules. Xiaomi founder 雷军 (Lei Jun), speaking as a National People's Congress delegate, stated he is a "principal victim" and called for urgent legislation to crack down on the problem.
Reported impact
Deepfake scams targeting elderly people have increased in scope and effectiveness. Young people are also vulnerable to AI-generated misinformation, with rumors such as "accumulated death rate of Chinese people born in the 1980s is 5.20%" spreading widely before being debunked by a Tsinghua University professor. Stock prices of companies like iFLYTEK have fluctuated due to AI-generated false information.
Experts warned that AI-generated content is increasingly difficult to distinguish from authentic material. Detection technology lags behind generation capabilities, and small platforms lack resources to invest in detection systems, leaving a governance gap.
Index analysis
The incident reflects a gap between rapid AI technology development and regulatory frameworks. Existing regulations such as the "Internet Information Service Deep Synthesis Management Regulations" have been enacted, but cross-platform coordination and the speed of technology development remain challenges.
Detection of AI-generated images remains behind generative model development. Small and medium platforms struggle to afford the high costs of detection technology, and text-based content governance relies on detecting duplicative content, which is insufficient to fully address misinformation.
Governance controls
Tsinghua University's School of Journalism and Communication Vice Professor 陆洪磊 warned that failure to manage AI rumors effectively could seriously undermine social confidence and stability. Experts called for further refinement of laws and regulations to ensure AI-generated content aligns with human ethical norms.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.