Konstruksi Pertanggungjawaban Pidana terhadap Konten Deepfake di Media Sosial Berdasarkan UU ITE dan Hukum Pidana Indonesia

Abstract

The rapid advancement of generative artificial intelligence has given rise to deepfakes, synthetic audio-visual content capable of convincingly imitating a person's face, voice, expressions, and gestures to the point of being indistinguishable from authentic recordings. This characteristic makes deepfakes prone to misuse for fraud, defamation, non-consensual pornography, and public information manipulation on social media. Indonesian criminal law does not yet comprehensively and explicitly regulate the term deepfake, so law enforcement against such acts relies on general provisions in the Electronic Information and Transactions Law, the National Penal Code, and the Personal Data Protection Law, none of which were historically designed to address the technical characteristics of AI-generated synthetic content. This study aims to analyze the adequacy of existing legal regulation of deepfake content and to formulate an appropriate construction of criminal liability for creators, uploaders, distributors, and parties who benefit from such content. The study employs a normative legal method with statute, conceptual, case, and comparative approaches. The findings reveal three fundamental problems: normative vagueness in the false-news offense, a legal vacuum concerning liability for automated generative AI use, and a mismatch between conventional actus reus elements and the technical mechanism of deepfake production. Based on these findings, this study formulates a tiered criminal liability model that differentiates perpetrators by role, form of fault, and purpose of deepfake use. The study concludes that a proportionate normative reformulation is needed to reach deepfake offenders without sacrificing legitimate freedom of expression, creative content, and satire.