When AI Hallucinates the Law: Rethinking Professional Misconduct and Judicial Responsibility in India's AI-Assisted Courts
Sristi Arya, Akshat BoseOct 31, 202610.5281/zenodo.23838229pp. 102–11716 pages
The paper highlights the issue of "hallucinations" of AI-aided Indian Judiciary, stressing on the landmark Supreme Court pronouncement in Pooja Ramesh Singh v. J&K Bank Ltd. (2026). The apex court has declared an absolute xero-tolerance standard for AI-produced unverified precedents, upholding those citations to be professional misconduct under Section 35 of the Advocates Act, 1961. The research paper examines the continuous tussle between the trap of efficiency of automatic legal research and the fundamental constitutional safeguards enshrined in Article 21, that includes the right of a fair trial and speedy justice. Comparing with international judicial pronouncements like the Mata v. Avianca, this study advocates that AI-produced unverified precedents divert the rule of law and vitiate the evidentiary efficacy of adjudication. At last, the paper argues for a strong regulation that makes human verification mandatory, protects judicial decision-making, and transforms judicial education to a antagonistic research pedagogy, guaranteeing that technology is a supportive mechanism that does not jeopardize truth and justice.