The Impact of Artificial Intelligence on Evidence in Criminal Trials

Abstract

The study has explored the dynamic nature of artificial intelligence on the presentations and the examination of evidence during criminal trials in the Pakistani judicial system. The study used a mixed-methods design to examine 150 criminal trials of district and sessions courts in Karachi, Lahore, Islamabad and Peshawar in the years 2020-2024 with 45 legal practitioners participating in structured interviews. The analysis showed that the question of admissibility and reliability assessment of the AI-generated evidence is often difficult, and judges proved to have different degrees of acceptance and awareness of AI technologies. Major outcomes were that although AI-enhanced forensic evidence was more accurate in 73 cases, judicial reluctance towards new technological evidence generated uneven use of the admissibility criteria. Findings of the study revealed that there are critical gaps in the legal frameworks that regulate AI evidence, and current Evidence Act provisions are insufficient to solve the issue of algorithmic transparency and bias. Findings showed that effective adoption of AI evidence was strongly linked with the quality of judicial training and quality of expert testimony. The findings of the study were that the criminal justice system in Pakistan needs a complete overhaul of the legal system, standardization of procedures in the evaluation of AI evidence, and an improved sphere of judicial education in order to utilize the potential of AI and preserve the principle of fair trial and the protection of the right to due process.