ARTIFICIAL INTELLIGENCE AND FORENSIC TOXICOLOGY: LEGAL CHALLENGES IN THE ADMISSIBILITY OF AI-ASSISTED TOXICOLOGICAL EVIDENCE IN INDIA

Abstract

Forensic toxicology occupies a pivotal place in the Indian criminal justice system, providing scientific answers to questions of poisoning, drug-facilitated crime, and cause of death. The discipline is now being reshaped by artificial intelligence (AI) and machine learning (ML) tools that promise faster compound identification, improved interpretation of mass-spectrometry data, and predictive modelling of toxicity. Yet Indian evidence law—rooted in Section 45 of the erstwhile Indian Evidence Act, 1872, and now Section 39 of the Bharatiya Sakshya Adhiniyam, 2023—was not designed with opaque, self-learning algorithms in mind. This paper examines how AI-assisted toxicological evidence fits within the Indian law of expert opinion and electronic evidence, drawing on comparative experience from the United States and the European Union, particularly the due-process controversies generated by proprietary risk-assessment algorithms such as COMPAS. It argues that the "black box" nature of many AI systems creates tension with foundational fair-trial guarantees—the right to cross-examine, the right to a reasoned judicial opinion, and the constitutional right to a fair investigation—and proposes a framework of algorithmic disclosure, human-in-the-loop verification, and judicial gatekeeping suited to Indian conditions.