<![CDATA[Artificial Intelligence (AI) is increasingly integrated into digital forensic and evidentiary processes, raising unresolved doctrinal questions in criminal procedure law. In Indonesia, although electronic evidence is formally recognized, the law does not yet provide specific admissibility standards for AI-based materials, particularly regarding authenticity, methodological reliability, process traceability, explainability, and accountability. This study examines the admissibility of AI as electronic evidence in Indonesia and compares it with legal approaches in the United States and Japan. This study employs a normative juridical method using statutory, conceptual, and comparative approaches to analyze the evidentiary frameworks of the three jurisdictions. The findings show that the United States emphasizes expert gatekeeping and digital authentication, while Japan adopts a softer regulatory model centered on traceability, documentation, and actor accountability. By contrast, Indonesia, still lacks specific procedural standards for assessing AI-generated outputs beyond the general recognition of electronic evidence. This article argues that the key legal issue is no longer whether electronic evidence is admissible in general, but how AI-based evidence should be evaluated in a legally reliable and accountable manner. The scientific contribution of this study lies in proposing a five-parameter evaluative model for AI admissibility—covering authenticity and integrity, process traceability, model performance, identity verification, and accountability. This model is offered as a normative reference for future reform of the Criminal Procedure Code and the Electronic Information and Transactions Law, while safeguarding legal certainty and justice.]]>
Copyrights © 2026
