AI-Generated Evidence in the Judicial Proceedings

The rapid integration of artificial intelligence (AI) in judicial proceedings underscores the need to establish rules governing the use of its outputs as evidence to ensure the fairness of the trial. This paper analyses the concept of generative AI results as evidence to propose procedural safeguards to uphold the fairness of judicial proceedings. It establishes that generative AI systems draw conclusions similar to those of humans. Because of their inherent nature, their outputs are prone to bias, unexplainability and inaccuracy. The research applies case law, doctrinal, structural and analytical methods to determine that AI outputs may be submitted as evidence. They differ from traditional types of evidence since they are often untested and inaccurate and thus require additional safeguards when used. When such evidence is submitted, it should be corroborated by another independent piece of evidence, accompanied by an independent examination of its reliability, or it should allow a party the opportunity to question the decision-making process of the system.