The Supreme Court (SC) has asked the Bar Council of India (BCI) to constitute an expert committee on AI use in court proceedings after a trial court in Andhra Pradesh cited four non-existent AI-generated judgments in a property dispute. The BCI must submit a report. No formal order was passed. The next hearing is scheduled for May 26.
What the court said: Citing fake AI-generated judgments is misconduct with legal consequences, the bench of Justices PS Narasimha and Alok Aradhe said, and not a mere error. Parties cannot rely on AI-generated material and later apologise, it added. The court flagged accountability gaps on both sides: litigants submitting unverified AI citations, and judges using AI for research that surfaces hallucinated precedents. It called for a sovereign large language model (a state-built AI system) and an India-specific legal database, while making clear that it is not seeking to ban AI.
The regulatory gap
India has no binding rules governing AI use in courts:
Attorney General (AG) R Venkataramani said he would consult the Ministry of Electronics and Information Technology (MeitY) before placing the government’s position. The AG’s response will likely shape whether this moves toward binding regulation or remains within judicial self-governance.
Amicus curiae Senior Advocate Shyam Divan pointed the court to a white paper on AI in the judiciary published by the SC’s Centre for Research and Planning in November 2025. The paper explicitly warned about hallucinations and documented instances of trial court judges relying on AI-generated material containing non-existent precedents. Divan said it could serve as a starting point for the court’s directions.
What misconduct means in practice: The consequences differ sharply depending on who is in the dock. For lawyers, the BCI’s Disciplinary Committee can reprimand, suspend, or remove an advocate from the roll under Section 35 of the Advocates Act, 1961. For judges, the bar is far higher. No judge has ever been successfully removed in India under the constitutional impeachment process, which requires a two-thirds majority in both houses of Parliament.
Short of impeachment, the Chief Justice of India can initiate an in-house inquiry under a procedure the SC evolved in 1999, which can result in a judge being advised to resign or having judicial work withheld. No external oversight body exists. The SC’s framing of AI citation as misconduct is therefore significant precisely because the accountability mechanisms to act on it remain weak and untested in this context.
This is not an isolated case: MediaNama has been tracking this pattern since early 2025:
The paradox: The SC runs the Supreme Court Portal for Assistance in Court Efficiency (SUPACE) for AI-assisted research, the Supreme Court Vidhik Anuvaad Software (SUVAS) for judgment translation, the Transcription of Electronic Record and Speech (TERES) for real-time transcription, and the Legal Research Analysis Assistant (LegRAA), its own generative AI tool trained on Indian case law to reduce dependence on commercial AI systems. All this comes while over 5.29 crore cases remain pending across Indian courts. The judiciary’s own AI adoption makes the accountability question harder to dodge.
Background: The case stems from a property injunction suit. In August 2025, the trial court dismissed the defendants’ objections to an advocate commissioner’s report, citing four SC judgments that do not exist. The Andhra Pradesh High Court acknowledged that the citations were AI-generated but still dismissed the revision petition on merits, recording only a word of caution. The SC took up the matter in February 2026, stayed further reliance on the commissioner’s report, and called it a matter of “considerable institutional concern.”
Also read:
For You
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。