Abstract: AI face-changing is derived from deep forgery technology, which involves a large amount of facial information and has data attributes. AI face-changing has high fidelity and low operating threshold, and has been widely used. However, the abuse of this technology has led to higher risks of infringement of private rights such as copyright and personality rights, information security risks, and crime prevention and control risks. Currently, there are two main models of overseas legal regulation of AI face-changing technology: decentralized legislative regulation and unified legislative regulation. Our country has not yet established a systematic regulatory system. It should be improved from the following aspects in the future: First, it is clear that reasonable use under the premise of distinguishing application scenarios is the basic regulatory principle. Second, it is to build a preevent and in-process regulatory system to clarify the technical ethics of developers. and producers’ identification and statement obligations, protect the information subject’s right to know and consent, and strengthen the content review obligations of communication platforms. The third is to build a regulatory system with data and algorithms as the core; the fourth is to strictly pursue civil legal liability and strengthen Strong criminal law sanctions.