<![CDATA[The rapid development of Artificial Intelligence (AI) has introduced deepfake technology, which is capable of manipulating audiovisual data with high realism. While this technology offers creative utility, its misuse for synthesizing non-consensual pornography, political disinformation, and character assassination severely violates an individual’s right to reputation. This study examines the civil and criminal liabilities of AI deepfake platform providers under the Indonesian legal framework, specifically Law No. 1/2024 (Second Amendment to the ITE Law) and Law No. 27/2022 on Personal Data Protection (UU PDP). Using a socio-legal research method, this study analyzes how textual legal norms interact with the digital behavior of platform providers and the psychological and social vulnerabilities of victims. The results indicate a significant normative gap: current regulations primarily target the individual creators or distributors of deepfake content, while platform providers escape liability by hiding behind the "safe harbor" doctrine. Sociologically, this legal vacuum fosters a culture of impunity and exacerbates digital gender-based violence. This paper argues for a paradigm shift from passive moderation to active algorithmic accountability. Platform providers must be held civilly liable under Article 1365 of the Indonesian Civil Code for systemic negligence, and criminally liable under the corporate fault doctrine for failing to implement mandatory digital watermarking and proactive content filtration systems.]]>
