In what may prove to be one of the most consequential judicial pronouncements on the use of artificial intelligence in the Indian legal system, the Supreme Court has set aside the judgments of the National Company Law Tribunal (NCLT) and the National Company Law Appellate Tribunal (NCLAT) after discovering that the adjudicatory process had relied upon non-existent, AI-generated “hallucinated” judicial precedents. Describing such fabricated authorities as an existential threat to the administration of justice, the Court observed that the use of AI-generated fake judgments is “like the release of methyl isocyanide in the province of law and justice” an invisible yet catastrophic danger capable of eroding the very foundations of judicial integrity. Simultaneously, the Court directed the Bar Council of India (BCI) to constitute a committee of experts to comprehensively examine the legal, ethical and professional challenges arising from the increasing use of artificial intelligence in legal practice and adjudication.
The judgment was delivered by a Bench comprising Justice P.S. Narasimha and Justice Alok Aradhe while hearing an appeal arising out of insolvency proceedings involving Essel Infraprojects. During the course of the proceedings, the Supreme Court was informed through an affidavit that several judicial authorities relied upon by the NCLT in support of its reasoning could not be traced in any recognised legal database. The impugned orders referred to purported Supreme Court decisions, including State Bank of India v. Shree Ram Urban Infrastructure, Everest Kento Cylinders v. Union of India and ICICI Bank v. Urban Infrastructure Real Estate. However, upon verification, the Court found that these authorities simply did not exist. They were neither reported in official law reports nor available in authenticated legal databases, exposing them as fabricated AI-generated citations or “hallucinations.” Accepting the submissions, the Supreme Court concluded that judicial decisions founded upon imaginary precedents could not be permitted to survive and accordingly quashed both the NCLT judgment and the appellate judgment of the NCLAT which had affirmed it.
The Court’s observations constitute perhaps the strongest judicial criticism of unverified artificial intelligence in legal adjudication delivered anywhere in India till date. Justice Narasimha observed that one of the defining limitations of current generative AI systems is their tendency to produce entirely fictitious yet convincingly worded legal material. While acknowledging that engineers and computer scientists may eventually resolve such technological deficiencies, the Court emphasised that the judiciary cannot afford to wait for technological perfection when the integrity of adjudication is at stake. Comparing AI hallucinations to the release of methyl isocyanide a reference evoking one of the gravest industrial disasters in Indian history the Court observed that fabricated precedents silently infiltrate judicial reasoning and often remain undetected until substantial damage has already been caused. The analogy reflects the Court’s recognition that misinformation generated by artificial intelligence poses systemic rather than merely individual risks to the justice delivery mechanism.
Perhaps the most significant aspect of the judgment lies in the Supreme Court’s declaration of a “zero-tolerance” policy towards the citation, production or judicial reliance upon AI-generated authorities without independent verification. The Court categorically held that an advocate who cites AI-generated judgments without verifying their authenticity commits professional misconduct. Equally, the Bench observed that a judge or adjudicating authority commits a serious institutional lapse if such fabricated material is relied upon while determining disputes. In unequivocal terms, the Court declared that any decision founded, even partially, upon fake or hallucinated precedents is “no decision in the eyes of law.” It further held that such decisions are liable to be set aside even if only “an iota” of fabricated material enters the judicial reasoning process because the legitimacy of adjudication depends fundamentally upon the authenticity of legal authorities cited in support of judicial conclusions.
Recognising that the challenge extends beyond an isolated case, the Supreme Court directed the Bar Council of India to establish an expert committee for examining the implications of artificial intelligence within the legal profession. The Court observed that the legal fraternity requires comprehensive guiding principles governing responsible AI usage, mandatory verification standards and disciplinary consequences for violations. It emphasised that the Bar Council must treat the issue with utmost seriousness and formulate a regulatory framework capable of preserving professional ethics while simultaneously allowing legitimate technological innovation. The direction reflects the Court’s understanding that artificial intelligence is no longer a future possibility but an active participant in legal drafting, legal research and courtroom preparation, necessitating immediate institutional safeguards.
The judgment does not represent the Supreme Court’s first engagement with the issue of AI hallucinations. Earlier this year, while dealing with another matter involving fabricated AI-generated citations, the same Bench had taken suo motu cognisance of the broader problem and sought responses from the Attorney General, Solicitor General and the Bar Council of India regarding appropriate regulatory mechanisms. Subsequently, the Supreme Court also released draft guidelines governing the use of artificial intelligence in legal proceedings and invited feedback from members of the Bar. The present judgment therefore represents the culmination of an evolving judicial concern regarding the unchecked integration of generative AI into legal practice.
From a jurisprudential perspective, the decision marks an important evolution in the doctrine of judicial integrity. Traditionally, appellate courts interfere with judgments on grounds such as perversity, jurisdictional error, violation of natural justice or misapplication of legal principles. The present ruling introduces another dimension by recognising reliance upon fabricated legal authorities as an independent ground for invalidating judicial decisions. In effect, the Supreme Court has elevated authenticity of legal sources into an essential component of procedural fairness and adjudicatory legitimacy. The implication is profound: irrespective of whether the ultimate conclusion reached by a court may otherwise appear legally sustainable, the decision itself becomes vulnerable if its reasoning rests upon imaginary precedents masquerading as genuine judicial authorities.
The ruling also carries significant implications for advocates practising before constitutional courts and specialised tribunals. Generative AI tools such as ChatGPT, Gemini, Claude and other large language models have rapidly become commonplace for legal research, drafting pleadings and preparing case summaries. While these technologies substantially improve efficiency, they remain susceptible to “hallucinations” confidently generating citations, quotations and legal propositions that have never existed. Internationally, several courts have already sanctioned lawyers for relying upon fabricated AI-generated authorities, most notably in the United States following the widely discussed Mata v. Avianca proceedings. The Indian Supreme Court’s decision aligns domestic jurisprudence with this emerging global consensus by affirming that technology cannot dilute a lawyer’s professional duty of verification.
Equally significant is what the judgment does not prohibit. The Bench was careful to clarify that the Court is not opposed to the legitimate use of artificial intelligence in legal practice. AI may continue to assist in document review, legal research, summarisation, translation, analytics and administrative efficiency. The Court’s objection is directed not against technology itself but against blind dependence upon unverified AI-generated outputs, particularly where such outputs are presented before courts as authentic legal precedents. This distinction reflects a technologically balanced approach that encourages innovation while reaffirming that professional responsibility cannot be delegated to algorithms.
The decision also arrives at a time when Indian courts are actively integrating artificial intelligence into judicial administration. The Supreme Court has already experimented with indigenous AI tools such as SUPACE (Supreme Court Portal for Assistance in Court Efficiency), LegRAA (Legal Research Assistance Application) and AI-assisted e-filing scrutiny mechanisms. Simultaneously, the Court’s Centre for Research and Planning has issued guidance requiring independent human verification of AI-generated summaries and precedent lists before judicial reliance. The present judgment complements these institutional initiatives by reinforcing that AI may function only as an assistive technology and never as an autonomous substitute for judicial reasoning or professional legal research.
Beyond the immediate insolvency dispute, the ruling may well become a constitutional milestone in India’s emerging jurisprudence on artificial intelligence and the administration of justice. By declaring fabricated AI-generated precedents fundamentally incompatible with the rule of law, the Supreme Court has reaffirmed that technological advancement cannot come at the cost of judicial credibility. The decision places equal responsibility upon the Bench and the Bar to maintain the authenticity of legal sources, recognising that public confidence in the justice delivery system depends not merely upon correct outcomes but upon transparent, verifiable and intellectually honest reasoning. As artificial intelligence becomes increasingly embedded within legal ecosystems, the judgment establishes a clear constitutional principle: technology may assist adjudication, but it can never replace the indispensable human obligation of verification, accountability and fidelity to the law.
