Supreme Court Sets Aside NCLT Judgment for Relying on AI-Hallucinated Citations

A Landmark Ruling on AI and Judicial Integrity

On July 2, 2026, the Supreme Court of India delivered a landmark judgment that sent shockwaves through the legal community. A bench comprising Justice PS Narasimha and Justice Alok Aradhe set aside orders passed by the National Company Law Tribunal (NCLT) and the National Company Law Appellate Tribunal (NCLAT) after discovering that both forums had relied on non-existent, AI-generated "hallucinated" judicial precedents while adjudicating an insolvency dispute related to Essel Infraprojects.

The ruling marks the first time India's highest court has directly confronted the dangers of artificial intelligence tools generating fictitious case law that finds its way into judicial proceedings. The court declared a "zero-tolerance" approach toward the production, citation, or reliance on AI-generated precedents that have not been verified for accuracy.

The Facts: How Fake Citations Entered the Record

The dispute arose in the context of insolvency proceedings involving Essel Infraprojects. During the course of adjudication, the NCLT relied on six judicial decisions that either did not exist or failed to support the legal propositions attributed to them. Senior Advocate Madhavi Divan argued before the Supreme Court that these citations were entirely fabricated, appearing to have been generated by an AI tool that "hallucinated" non-existent case law.

A critical aspect of the case emerged through an affidavit filed by Jammu & Kashmir Bank, which stated that the hallucinated judgments were not cited by its counsel. Rather, the NCLT itself had obtained these fictitious precedents through its own research, raising serious questions about whether the tribunal had used an AI tool without verifying the authenticity of the outputs. This finding was particularly alarming because it indicated that the problem was not limited to legal practitioners but had infiltrated the adjudicatory process at the tribunal level itself.

The Court's Observations: "Methyl Isocyanate in the Province of Law"

The Supreme Court bench used remarkably strong language to convey the gravity of the situation. Comparing AI-generated fake precedents to the deadly gas leak that devastated Bhopal in 1984, the Court observed that reliance on hallucinated AI material is "like the release of methyl isocyanate in the province of law and justice: invisible, insidious, and catastrophic by the time anyone notices."

This vivid analogy underscores the Court's view that AI hallucinations in legal proceedings are not merely technical errors but fundamental threats to the integrity of the justice system. When a tribunal relies on fictitious case law, the resulting orders lack legitimate legal foundation, potentially affecting the rights and obligations of parties in significant ways. The insolvency context made the consequences especially severe, given the far-reaching impact of NCLT orders on creditors, employees, and the corporate debtor itself.

BCI Directed to Constitute Expert Committee

Beyond setting aside the impugned orders, the Supreme Court issued a significant institutional directive. The bench directed the Bar Council of India (BCI) to constitute an expert committee to examine the issues arising from the use of artificial intelligence in adjudication. The committee is expected to frame guiding principles and norms that would regulate how AI tools are used by legal practitioners and judicial bodies alike.

This directive acknowledges that AI in the legal profession is not going away and that a regulatory framework is essential. Rather than banning AI use outright, the Court's approach focuses on accountability, verification, and transparency. The BCI committee is expected to address questions such as: What verification obligations should lawyers have when using AI-generated research? Should AI-generated content be disclosed to courts? What sanctions should follow the citation of unverified AI outputs?

Growing Global Concern Over AI Hallucinations in Courts

India is not alone in grappling with this problem. Courts around the world have encountered instances of lawyers submitting AI-hallucinated citations. In the United States, a federal judge sanctioned lawyers for citing fictitious cases generated by ChatGPT in Mata v. Avianca in 2023. The Indian Supreme Court's ruling, however, is notable because the AI-hallucinated content was introduced not by counsel but by the tribunal itself, raising the stakes considerably.

This judgment fits into a broader pattern of Indian courts and regulators engaging with AI governance challenges. The Digital Personal Data Protection Act and evolving IT Rules on deepfakes reflect a growing recognition that AI technologies require dedicated legal frameworks across different domains.

Implications for the Insolvency Framework

The case carries particular significance for India's insolvency and bankruptcy framework. NCLT orders in insolvency proceedings carry enormous consequences for all stakeholders, including creditors, employees, promoters, and the corporate debtor. When such orders are founded on fabricated legal authority, the entire process is compromised. The Supreme Court's intervention ensures that the matter will be reheard on a legitimate legal basis.

The broader implications for corporate governance and compliance are significant. Companies that are parties to tribunal proceedings now have additional grounds to scrutinize the legal reasoning in tribunal orders, particularly where the cited authorities are unfamiliar or difficult to locate in established databases.

What This Means Going Forward

The Supreme Court's ruling establishes several important principles. First, it places a clear verification duty on anyone who uses AI tools for legal research, whether lawyer or judge. Second, it signals that reliance on unverified AI outputs in judicial proceedings will be met with serious consequences, including the setting aside of orders. Third, by directing the BCI to frame norms, the Court has set in motion an institutional process that could result in formal rules governing AI use in Indian legal practice.

The earlier coverage of AI-generated fake precedents in Indian courts has highlighted that this is not an isolated incident. Multiple instances have surfaced across different courts, and the problem is likely to intensify as AI tools become more accessible and widely used.

Legal professionals must treat AI-generated research as a starting point that requires rigorous cross-verification against established legal databases. The convenience of AI tools cannot justify shortcuts that compromise the foundations of the justice system. As the Supreme Court has made clear, the consequences of failing to verify AI outputs in legal proceedings will be severe, both for the orders in question and for the professionals responsible.

As India develops its broader data protection framework and addresses the decriminalization of regulatory offences, the regulation of AI in judicial and legal proceedings will be a critical component of ensuring that technological advancement does not come at the cost of justice.