Reconstructing Validity and Reliability Standards for AI-Based Evidence in Indonesian Criminal Proceedings

Abstract

The use of artificial intelligence in criminal investigations and court proceedings creates new challenges for Indonesian evidence law. AI can be used for facial recognition, data analysis, voice identification, and detecting fake media. However, Indonesian criminal procedure does not yet provide clear rules on whether AI-based evidence is reliable, transparent, and acceptable in court. This study aims to develop a normative framework for evaluating the admissibility and reliability of artificial intelligence-based evidence in Indonesian criminal proceedings by integrating evidentiary principles, algorithmic accountability, and defendants’ procedural rights. The research employs a normative legal method using statutory, conceptual, and comparative approaches, examining Indonesian evidentiary rules alongside selected foreign legal frameworks governing the admissibility and scrutiny of artificial intelligence-based evidence. Primary legal materials include Indonesian criminal procedure legislation and regulations concerning electronic information, while secondary materials consist of legal scholarship on digital evidence, algorithmic accountability, and fair trial guarantees. The study finds that classifying evidence as electronic evidence is not enough to establish its admissibility. AI-based evidence may raise additional concerns about its source, reliability, transparency, and the ability of the defence to challenge how it was produced and interpreted. Its admissibility should depend on how the evidence was obtained, its integrity and accuracy, the transparency of the methods used, expert verification, and the defence’s ability to examine and challenge it. The study concludes that Indonesia needs a clear legal framework for AI-based evidence to ensure that AI results are properly examined and that the use of technology does not undermine due process.