The Supreme Court of India has struck down insolvency orders against Essel Infraprojects Ltd. after discovering that the lower tribunals relied on fictitious judicial precedents created using artificial intelligence. The apex court remanded the case to the National Company Law Tribunal (NCLT) for a fresh hearing based on verified legal citations and factual findings.
Case Background
Essel Infraprojects Ltd. acted as the corporate guarantor for a loan taken by another company. Jammu & Kashmir Bank initiated insolvency proceedings against the company. Pooja Ramesh Singh, the suspended director of Essel Infraprojects, filed a Section 7 petition challenging the admission of the company into corporate insolvency resolution. The NCLT Mumbai Bench initially admitted the insolvency petition on August 28, 2024, and the National Company Law Appellate Tribunal (NCLAT) confirmed that decision on September 11, 2025.
Fake AI-Generated Citations Exposed
Senior counsel Madhavi Divan, representing Essel Infraprojects Ltd., alerted the court to the issue. She pointed out that the NCLT had relied on several judgments that either did not exist or contained paragraphs not present in the original rulings. The Supreme Court conducted an independent verification and confirmed the problem. Some citations referenced actual Supreme Court rulings but included fabricated paragraphs, while others cited completely nonexistent judgments.
Supreme Court’s Strong Warning
A bench comprising Justices Pamidighantam Sri Narasimha and Alok Aradhe declared that the use of fake, AI-generated documents must be met with ‘zero tolerance.’ The bench emphasized that relying on hallucinated data without proper verification poses a serious threat to the integrity of the legal system. The case raises broader questions about the application of AI in legal decision-making.
The judges drew a clear distinction between using AI as a tool and allowing it to replace human judgment. They stated: “It is not only a device to help us in our work; it is an alternative to our very thoughts, reasoning, and even decision-making.” However, they clarified they are not opposed to AI usage in law practice or courts, adding: “While it can indeed be a tool in the courtroom, it must be supervised completely by human beings.” The court directed the Bar Council of India to form a committee to study the issue thoroughly.
Previous Instances of AI Errors in Indian Courts
This is not the first time Indian courts have flagged AI-generated errors. In January 2026, the High Court of Andhra Pradesh warned trial court judges after discovering that a lower court had referred to four fake case statutes generated by an AI software. In 2025, the Bombay High Court found a tax assessment order citing fake court precedents, and the Supreme Court itself encountered completely fabricated citations in a filed reply.
Draft AI Regulations Released
In response to the growing problem, the Supreme Court published draft regulations in June 2026. These regulations strongly prohibit the use of AI tools in judicial decision-making or determining case outcomes. However, they allow AI for administrative and research tasks such as legal research, writing, translation, transcription, and case management. The document specifically warns against the dangers of unreliable automated content, erroneous citations, and AI hallucinations.