<![CDATA[The rapid advancement of artificial intelligence has given rise to deepfake technology, which employs deep learning through Generative Adversarial Networks (GANs) to realistically manipulate audio-visual content without the subject's consent. Misuse of this technology inflicts multiple layers of harm on victims, including violations of honor and privacy, alongside prolonged psychological trauma. Indonesia's existing regulations, namely the ITE Law No. 1 of 2024 and the PDP Law No. 27 of 2022, have not explicitly defined deepfakes as an independent criminal offense, creating a legal vacuum (rechtsvacuum) that leaves victims vulnerable to further harm. This research employs a normative legal method, using legislative, conceptual, and comparative approaches. The findings reveal that both regulations remain reactive and inadequate to address the complexity of AI-based digital identity manipulation. In response, a victim-oriented justice approach is proposed as a paradigm for legal reconstruction, positioning victims as the central subject of recovery. The proposed ideal reconstruction consists of three layers of protection: preventive, through the establishment of lex specialis regulation; repressive, through mandatory restitution; and rehabilitative, through guarantees of the right to be forgotten and psychological recovery. This study concludes that comprehensive special legislation oriented toward fulfilling victims' rights in digital crimes is urgently needed.]]>
