Deepfake, portrait rights and identity protection: a comparative analysis of Chinese and Brazilian law

Abstract

The rapid development of generative artificial intelligence has transformed the ways in which human identity can be represented, reproduced, and manipulated. Deepfakes, in particular, enable the creation of highly realistic synthetic representations of individuals’ faces, voices, and gestures, often without their knowledge or consent. This technology challenges traditional concepts of personality rights and raises complex questions concerning image, voice, privacy, dignity, reputation, and the allocation of liability among individuals, platforms, and technology providers. Against this background, this article examines the legal protection of identity against deepfakes through a comparative analysis of Chinese and Brazilian law. It first examines the Chinese framework, focusing on the 2021 Civil Code and its express protection against technologically generated forgery, as well as administrative regulations governing deep synthesis and generative artificial intelligence. It then analyzes the Brazilian framework, based on constitutional guarantees, personality rights under the Civil Code, the General Data Protection Law, the Marco Civil da Internet, and emerging judicial responses to synthetic content. The comparative analysis identifies two distinct regulatory approaches: a more preventive and technologically specific framework in China and a more reactive and jurisprudential model in Brazil. The article argues that, despite their different institutional structures, both systems face unresolved challenges concerning consent, platform responsibility, rapid content removal, freedom of expression, and the protection of identity in artistic, commercial, political, and personal contexts. It concludes by examining the extent to which existing personality rights can accommodate synthetic identity and whether additional legal and institutional mechanisms are necessary to provide effective protection in an increasingly synthetic information environment.