Indian courts grapple with AI hallucination problem

India's Supreme Court has issued a ruling establishing zero tolerance for Artificial Intelligence (AI)-generated false legal precedents. The July 2, 2026, decision addresses a growing problem where various courts have relied on fabricated citations, non-existent judgments, or real judgments with invented paragraphs.

Over the past year, multiple instances have surfaced across tribunals, trial courts, and High Courts where AI-generated legal materials have led to erroneous decisions. One significant case involved a tribunal that based its order on six defective precedents. Three of these citations did not exist, and three real ones contained fabricated sections or incorrect case titles. Alarmingly, the tribunal sourced these fake precedents independently, bypassing scrutiny.

The Supreme Court has classified the use of such material as "misconduct" for advocates and a "serious lapse" for judges. It has directed the Bar Council of India to frame disciplinary norms and metaphorically compared hallucinated case law to "the release of methyl isocyanate in the province of law and justice."

This stance follows a series of related incidents. In February 2026, the Supreme Court declared a trial court's decision based on fake AI judgments as "misconduct" with legal consequences. Other High Courts, including the Bombay and Delhi High Courts, have fined litigants and officials for submitting AI-generated content. Even the Punjab and Haryana High Court engaged with ChatGPT in March 2023 to gain a broader perspective on bail jurisprudence, though it clarified this was for context and not an authoritative opinion.