Criminal law responses to AI-driven deepfake biometric information in China
Year: 2026
Type: article
Abstract: Biometric information is inherently immutable, unique and irrevocable. The abuse of generative AI-driven deepfake technology has given rise to complex criminal risks such as identity theft, property loss, and reputational damage, transcending traditional data security hazards and thus emerging as a global issue. Based on a survey of 327 valid respondents across various industries and age groups, this reveals strong public concern about deepfake-related harms and an urgent demand for legal protection of digital identity. A comparative analysis of legislative and regulatory practices in China, Europe, and the United States reveals that China's existing criminal law framework, which focuses on “data control behaviors”, struggles to regulate new crime patterns like lawful-acquisition deepfake synthesis. In contrast, the EU and the US have formed targeted regulatory paths by expanding legal interests, establishing cross-field identification standards for rights infringement, and adopting precondition-based regulatory mechanisms. To address this gap, this paper proposes a paradigm shift in China's criminal law response: At the legislative level, prioritize adding the crime of “Illegal Forgery of Identity Information” or modifying the applicable boundaries of existing charges as an alternative. At the judicial level, establish a cross-field hierarchical evaluation mechanism for the degree of criminal harm and clarify the tripartite liability of developers, platforms, and users. At the governance level, construct a full-chain, cross-border prevention and control system through the connection between administrative law enforcement and criminal justice, integrating technical measures such as digital watermarking and liveness detection. This study aims to provide a systematic and operable legal solution for safeguarding digital identity security in the AI era while balancing technological innovation and risk prevention.
Authors: Lushasha Xu, Zeqiang Guo
Language: English
Subfield: Law
Field: Social Sciences
Domain: Social Sciences
Open Access status: hybrid
Estimated APC paid: $3,380
OpenAlex ID: W7212261965