Artificial Intelligence, Machine Learning Outputs And Future Scientific Evidence Under The Bharatiya Sakshya Adhiniyam, 2023
Md Irtifa Ur Rahman, B.A.LL.B., School of Law Justice & Governance, Gautam Buddha University, Greater Noida, U.P.
ABSTRACT
The Bharatiya Sakshya Adhiniyam, 2023 (BSA), in force since 1 July 2024, modernised Indian evidence law by placing electronic and digital records on an equal footing with paper documents. Yet the statute was drafted for records that machines store, not for conclusions that machines generate. Outputs of artificial intelligence (AI) and machine learning (ML) systems, such as facial recognition matches, predictive risk scores, voice identification, automated transcription and synthetic media detection, are probabilistic inferences produced by models whose internal reasoning is often opaque. This article examines whether, and on what terms, such outputs can be admitted and weighed under the BSA. It analyses the definitional provisions on documents and evidence, the primary evidence explanations in Section 57, the certification regime in Section 63 as upheld by the Supreme Court in Pune Bar Association v. Union of India (2026), and the expert opinion provisions in Section 39. It argues that the Section 63 certificate secures the integrity of a record but says nothing about the reliability of an algorithmic inference. Drawing on comparative jurisprudence and the literature on explainable AI, the article proposes a layered approach combining authentication, a reliability inquiry modelled on scientific-validity standards, disclosure obligations, and safeguards for the accused under Article 21 of the Constitution. It concludes with recommendations for legislative rules, forensic standards, and judicial training so that future scientific evidence strengthens rather than distorts the search for truth.
Keywords: Bharatiya Sakshya Adhiniyam 2023; artificial intelligence; machine learning; electronic evidence; expert opinion; Section 63 certificate; algorithmic reliability; deepfakes; explainability; fair trial.
