AI-Generated Evidence And Judicial Decision-Making In India: Constitutional Limits Of Admissibility, Reliability, Human Oversight, And The Role Of Artificial Intelligence In Judicial Discretion

Abstract

The rapid integration of artificial intelligence into investigative and forensic processes has introduced a novel category of evidence, AI-generated outputs ranging from algorithmic analytics and facial-recognition matches to synthetic media and automated forensic reports, into Indian courtrooms. While the Bharatiya Sakshya Adhiniyam, 2023 (particularly Sections 61 to 63) treats such material as electronic records admissible upon certification, the statute provides no dedicated reliability standard, explainability requirement, or framework for scrutinising the underlying generative process. This article critically examines the constitutional limits on the admissibility and reliability of AI-generated evidence, locating the inquiry within Articles 14 and 21 of the Constitution, the right to a fair trial, and the non-delegable character of judicial discretion. It argues that the black-box nature of many AI systems, combined with risks of bias, hallucination and deepfake manipulation, threatens due process unless subjected to rigorous human oversight and a structured reliability inquiry that goes beyond formal certification. Drawing on the Supreme Court's insistence that justice requires human, not mechanical, interpretation, and on emerging judicial guidelines that prohibit AI from performing adjudicatory functions, the article contends that AI may assist but can never displace judicial discretion. It concludes by proposing a constitutionally grounded framework of mandatory disclosure, independent expert validation and an explicit human-in-the-loop requirement to preserve the integrity of Indian adjudication in the algorithmic age.