This research analyzes Indonesia's legal construction in addressing the misuse of deepfake technology in the criminal offense of extortion, examining both substantive criminal law and evidentiary aspects. The findings indicate that the extortion provision under Article 482 paragraph (1) of the National Penal Code is inadequate to address deepfake-based threats, as the element of "violence" is still interpreted as physical violence, making the offense of intimidation under Article 483 of the National Penal Code and Article 27B paragraph (2) of the Electronic Information and Transactions Law a more precise legal construction. A single act of deepfake-based extortion may even fall within the scope of four overlapping legal regimes simultaneously, namely the National Penal Code, the Electronic Information and Transactions Law, the Personal Data Protection Law, and the Pornography Law, which in turn generates normative fragmentation and threatens legal certainty as conceptualized by Jan Michiel Otto. With respect to evidentiary matters, Law Number 20 of 2025 concerning the Criminal Procedure Code retains the negative statutory system of proof (negatief wettelijke bewijstheorie) while expanding the categories of admissible evidence to nine types, including electronic evidence and judicial observation, both of which are particularly relevant in assessing the authenticity of AI-generated content. Nevertheless, the practical implementation of this evidentiary framework remains constrained by the limited number of certified digital forensic experts and the uneven distribution of accredited forensic laboratories. This research concludes that the core challenge lies not in the absence of applicable norms, but rather in the urgent need for harmonization across legal regimes and the strengthening of law enforcement's technical capacity.