Burhan Doğuş Ayparlar

← All decisions

India · 27 February 2026

Interim ruling

Supreme Court of India takes up a trial court's reliance on AI-generated fake judgments and calls decisions based on them misconduct

A trial court relied on four fake, AI-generated rulings; the Supreme Court called such decisions misconduct with legal consequences and took up the issue.

Court / authority
Supreme Court of India
Date
27 February 2026
Case / decision no.
Gummadi Usha Rani & Anr. v. Sure Mallikarjuna Rao & Anr., SLP (C) No. 7575/2026, order of 27 February 2026
Status
Interim ruling
Supreme Court of India, Gummadi Usha Rani & Anr. v. Sure Mallikarjuna Rao & Anr., SLP (C) No. 7575/2026, order of 27 February 2026, 27 February 2026 — https://www.burhandogusayparlar.com/ai-decisions/gummadi-usha-rani-v-sure-mallikarjuna-rao-ai-fake-judgments/
  1. Facts
  2. Question
  3. Decision
  4. Why it matters

Facts

The petitioners were defendants in an injunction suit over property in Andhra Pradesh. The trial court had appointed an Advocate Commissioner to note the property's physical features and, on 19 August 2025, dismissed the defendants' objections to his report, relying on four purported Supreme Court decisions, including ‘Subramani v. M. Natarajan (2013) 14 SCC 95’. In revision, the defendants argued that these judgments did not exist. The High Court of Andhra Pradesh found that they were AI-generated, recorded a word of caution and, deciding on the merits, dismissed the civil revision petition on 21 January 2026. The defendants then sought special leave to appeal.

Question

The merits concerned an ordinary property dispute; the real question was what follows when a judge, not a party, grounds an order on authorities that do not exist. The High Court had dealt with the problem through a word of caution and upheld the result on its own reasoning. The Supreme Court had to decide whether the trial court could keep acting on the order while it reviewed the case, how reliance on fabricated precedents should be characterised, and where responsibility lies when AI-generated material enters adjudication. It also had to settle how to examine the wider questions about AI use by judges and lawyers.

Decision

On 27 February 2026 a bench of Justices Pamidighantam Sri Narasimha and Alok Aradhe issued notice and directed that, pending the petition, the trial court must not proceed on the basis of the Advocate Commissioner's report. Taking cognisance of the trial court deploying AI-generated, non-existent alleged judgments, it declared that a decision based on them ‘is not an error in the decision making’ but ‘would be a misconduct and legal consequence shall follow’. It issued notice to the Attorney General, the Solicitor General and the Bar Council of India and appointed Senior Advocate Shyam Divan to assist the Court. The order makes no finding on the trial judge's personal responsibility.

Why it matters

Most AI citation cases concern lawyers; here the fabricated authorities appeared in a judicial order. By describing reliance on them as misconduct with legal consequences, the Court signalled that a judge's failure to verify authorities may be treated as more than an error to be corrected on appeal. The order is interim, and no ruling on accountability had been made at the time of checking. At the 5 May 2026 hearing the Court orally asked the Bar Council to form an expert committee, without passing a formal order, and according to LiveLaw said it was not seeking to prohibit AI.

Related stages

High Court of Andhra Pradesh: CRP No. 2487/2025, judgment of 21 January 2026 (not reviewed for this entry). According to BBC reporting of that judgment, the trial judge told the High Court that this was a first use of an AI tool and that the citations had been believed genuine, and the High Court accepted that the error was made in good faith; LiveLaw described the judgments as cited by a litigant. The Supreme Court's orders do not say how the citations entered the order. On 26 May 2026 the Court added the information technology and law ministries as respondents and listed the case for 1 September 2026; no later order was found. In March 2026 another bench, in a separate matter, called the citation of AI-generated fake judgments a ‘menace’ rampant in courts worldwide (LiveLaw, 26 March 2026).

Related articles on this site

Similar decisions