The development of artificial intelligence has significantly changed the production and dissemination of digital content. One form of its misuse is deepfake technology, which refers to AI-generated synthetic content capable of realistically imitating a person’s face, voice, or actions. This study examines criminal liability for the use of artificial intelligence in deepfake-related crimes, particularly the legal position of AI as an instrument of crime and the parties who may be held criminally liable. This research applies a normative legal method using statutory and conceptual approaches. The legal materials examined include the Indonesian Criminal Code, the Electronic Information and Transactions Law, the Personal Data Protection Law, the Pornography Law, the Sexual Violence Crimes Law, and relevant literature on criminal law, artificial intelligence, and deepfake technology. The findings show that AI cannot be regarded as a subject of c’iminal liability because it lacks free will, moral awareness, and the cap’city to be legally blamed. Criminal liability should instead be attributed to individuals or corporations that create, order, disseminate, control, or benefit from deepfake content. Indonesian criminal law remains fragmented in regulating deepfake-related conduct, making clearer rules necessary on synthetic content, digital identity manipulation, platform obligations, and victim protection.