How Fake AI Research Just Blew up a Major Supreme Court Corporate Insolvency Case

In a major ruling on judicial accountability, the Supreme Court has ordered that any legal judgment relying on fabricated, artificial intelligence-generated research is void and must be thrown out entirely.

Establishing a strict "zero-tolerance" policy for both the Bench and the Bar, a bench of Justices P S Narasimha and Alok Aradhe declared that any judicial order contaminated by even an "iota" of hallucinated research violates the core tenets of adjudication.

The landmark directive has effectively quashed the insolvency proceedings against Subhash Chandra-promoted Essel Infraprojects Ltd. The apex court discovered that lower tribunals had inadvertently built their entire case around completely non-existent precedents generated by AI.

The bench issued an unusually blunt warning on how generative AI tools can quietly compromise the legal system, drawing a parallel to the Bhopal gas tragedy.

"...for those in the province of adjudication and determination of disputes, this by-product of AI, i.e., the production of fake, non-existent, and hallucinated material and its utilisation as precedents in law, is like the release of methyl isocyanate in the province of law and justice: invisible, insidious, and catastrophic by the time anyone notices," the bench observed.

Remarkably, the fake citations did not originate from the advocates arguing the matter. Instead, the National Company Law Tribunal (NCLT) panel found the hallucinated precedents themselves during independent research. The appellate body (NCLAT) subsequently failed to catch the fabrications on appeal.

Challenges with AI

While acknowledging that technology has seamlessly aided court administration for years, the bench warned that generative AI poses an entirely different threat because it acts as an unregulated alternative to a judicial officer's own thinking and decision-making.

The ruling shifts immediate accountability directly onto the courtroom floor. For judges, relying on unverified AI materials is now classified as a "serious lapse." The Supreme Court clarified that any decision using fake precedents is unsustainable in law, regardless of whether the fake material directly or indirectly altered the final outcome.

For practicing lawyers, citing unverified, AI-fabricated judgments now officially constitutes professional misconduct. The Supreme Court has ordered the Bar Council of India (BCI) to immediately form a committee to frame regulatory guidelines and outline strict disciplinary actions for future violations.

The court clarified that the order does not bar the legitimate use of AI as an efficiency or administrative tool, but draws a strict boundary against presenting machine hallucinations as binding law.

The Case

The crisis originally came to light during a long-running financial dispute involving Jammu & Kashmir Bank and Essel Infraprojects.

In 2013, the bank sanctioned a ₹2 billion working facility to Pan India Utilities Distribution Co. Ltd., secured by a corporate guarantee and a land mortgage from Essel Infraprojects. Following a default by Pan India, the bank triggered the guarantee and moved the NCLT to initiate insolvency proceedings against Essel Infraprojects for an outstanding amount of ₹874.32 million.

The NCLT admitted the insolvency plea, and the NCLAT later upheld the move.

However, the case imploded when Pooja Ramesh Singh, a suspended director of Essel Infraprojects, moved the apex court. Her legal team exposed that the key case citations the NCLT used to justify its order simply did not exist in any law report.

Court reporters and legal researchers tracking the daily cause list at Tilak Marg had quietly whispered for weeks about inconsistencies in the lower tribunal's written text, but few expected the apex court to issue such a sweeping, system-wide reckoning.

Setting aside the flawed insolvency orders, the Supreme Court directed the NCLT to hear Jammu & Kashmir Bank's plea afresh—this time, completely free of AI fabrications.