Fake AI Use in Courts Could Be Catastrophic, Supreme Court Warns
The Supreme Court set aside NCLT and NCLAT orders after finding that the NCLT had relied on fake and hallucinated judicial precedents in an insolvency case.
News
- Amaravati to Host One of India’s First Two IBM Quantum Computers
- HCLTech Wins $1.14 Billion AI Deal
- Fake AI Use in Courts Could Be Catastrophic, Supreme Court Warns
- Government Tells WhatsApp to Pause Username Rollout in India
- Meta Explores AI Cloud Push to Monetize Computing Capacity: Report
- OpenAI Floats 5% US Stake Plan
The Supreme Court on Thursday set aside orders passed by the National Company Law Tribunal (NCLT) and the National Company Law Appellate Tribunal (NCLAT) after finding that the NCLT had relied on fake and hallucinated judicial precedents, while the defect escaped scrutiny in appeal.
A Bench of Justices P.S. Narasimha and Alok Aradhe said courts must adopt a zero-tolerance approach to unverified AI-generated legal material. A decision based on such material, the court said, is “no decision in the eyes of the law.”
The case arose from insolvency proceedings against Essel Infraprojects Ltd. The NCLT’s Mumbai Bench had admitted Jammu and Kashmir Bank Ltd’s insolvency plea on 28 August 2024. The NCLAT upheld that order on 11 September 2025. The Supreme Court has now set aside both orders and restored the matter to the NCLT for fresh consideration.
During the hearing, senior advocate Madhavi Divan, appearing for the appellant, told the court that several precedents relied on by the NCLT were either non-existent or did not contain the passages attributed to them.
J&K Bank later said the disputed cases had not been cited by its counsel and appeared to have come from the tribunal’s own research.
The Supreme Court examined six decisions cited by the NCLT.
It found wrong citations, non-existent citations and passages that could not be traced to the judgments cited. In some cases, the citation existed but the paragraph relied on did not. In others, the citation itself was fake.
The Bench said fabricated AI-generated precedents were not a minor error.
It compared their use in adjudication to “the release of methyl isocyanate in the province of law and justice,” calling the effect “invisible, insidious, and catastrophic by the time anyone notices.”
The court said advocates who cite fake AI-generated judgments without verification commit misconduct.
Judges who rely on such material, it added, commit a serious lapse in the judicial process. Any decision in which even “an iota of fake or hallucinated material” enters the reasoning must be set aside, the Bench said.
The Supreme Court asked the Bar Council of India to constitute a committee to examine the issue and prescribe guiding principles to prevent lawyers from submitting fake or hallucinated material as legal precedent.
It also asked the council to consider disciplinary consequences for violations.
The Bench made clear that it was not ruling against the use of AI itself. Courts have absorbed technology before, it said, but AI is different because it can influence reasoning and decision-making.
Adjudication, the court said, must remain under “total and absolute control” of humans at every stage.
The warning comes weeks after the Supreme Court’s AI Committee published draft Regulations for Use of Artificial Intelligence in Courts, 2026, seeking comments from stakeholders and the public.
The draft framework seeks to govern AI use on principles including human primacy, transparency, accountability, data protection and judicial independence.
