AI Regulation Tracker / Court ruling
India's Supreme Court treats citing AI-invented precedent as advocate misconduct
In Pooja Ramesh Singh v. Jammu and Kashmir Bank Ltd. (2026 INSC 668), the Court set aside orders that rested on fabricated AI-generated case law, called citing such precedent misconduct, and directed the Bar Council of India to write disciplinary norms. The citation duty it enforces is universal.
This one should get a US attorney's attention even though it comes out of an Indian insolvency file. The docket facts are plain. The Supreme Court of India decided Pooja Ramesh Singh v. Jammu and Kashmir Bank Ltd., reported as 2026 INSC 668, on July 2, 2026, before Justices P.S. Narasimha and Alok Aradhe. The appeal came up from proceedings under the Insolvency and Bankruptcy Code. What turned it into a landmark was what the Court found in the record below: citations to judgments that do not exist, along with quoted passages that appear nowhere in any real decision. The pattern is the signature of a large language model inventing authority on demand.
What the Court held, as reported
I am working from the published reports of the judgment rather than the full text, so I will attribute the holding rather than state it as if I read every line. According to Verdictum's report of the judgment, the Court adopted a zero-tolerance posture toward producing, citing, or using AI-generated precedent without verification, and said citing such judgments is misconduct on the part of an advocate. The same report frames the consequence in stark terms. A decision that rests on fake or hallucinated material is, in the Court's words as reported, "no decision in the eyes of the law." The reporting notes the Court set aside the orders below and sent the matter back for fresh proceedings.
The line that operators should sit with is the one about how little contamination it takes. As Verdictum reports the ruling, such decisions are to be set aside even if an iota of fake or hallucinated material enters the decision-making process. Read that carefully. The test is not whether the fabricated citation changed the outcome. The presence of the fabrication is enough.
The direction to the Bar Council
Beyond the individual case, the Court looked at the system. Multiple outlets reporting the judgment note that it directed the Bar Council of India to constitute a committee and frame disciplinary guidelines for advocates who place fabricated or hallucinated precedent before a court. In plain terms, this is not being left as a one-off scolding. It is being pushed toward standing professional-conduct rules. I am not naming any sanctioned lawyer or a specific penalty here, because the reporting I relied on does not support stating either as fact.
Why a US professional should track this
The duty the Indian court enforced is not exotic and it is not foreign. In the United States you already carry it under Federal Rule of Civil Procedure 11 and its state analogs: when you sign a filing, you are certifying that the legal contentions are warranted. Courts here have already sanctioned lawyers who filed AI-invented citations. What India adds is a high court framing the same failure as professional misconduct at the national level, and pointing its bar regulator at written rules. That is the direction of travel. If you draft with an AI assistant, the workflow control is boring and non-negotiable: pull and read every case it cites before it goes in the document. A citation you cannot open and verify is a citation you do not file.
There is a second reason to watch it. India joins a growing line of courts across jurisdictions treating fabricated AI citations as a conduct problem, not a harmless mistake. When the rule shows up the same way in South Korea, the United Kingdom, and now the Supreme Court of India, it stops being a novelty and starts being the settled global expectation of anyone who files.
Questions professionals are asking
Does this ruling apply outside India?
No. 2026 INSC 668 is binding law in India, not in the United States. What carries across borders is the duty it enforces. Verifying every citation before you file is already required of US lawyers under Rule 11 and equivalent state rules, and courts here have sanctioned AI-invented citations.
What exactly did the Court say about AI citations?
According to Verdictum's report of the judgment, the Court held that citing unverified AI-generated precedent is misconduct on the part of an advocate, and that a decision resting even partly on fabricated material is no decision in the eyes of the law and must be set aside. We are attributing the holding to the reporting because the full text is not freely available.
What is the practical step for a lawyer using AI?
Treat AI output as an unverified draft. Before any citation goes into a filing, open the actual decision in a real reporter or database and confirm it exists and says what the draft claims. If you cannot verify it, do not cite it.
RELATED BRIEFINGS
Browse the full AI Regulation News tracker
Informational analysis for working professionals, not legal advice. Confirm how any rule applies to your situation with qualified counsel in the relevant jurisdiction.