Back
Supreme Court Sounds Alarm on AI's Fake Judgment Menace While Granting Rare Relief
In a striking order on , a bench of the comprising Justice Rajesh Bindal and Justice Vijay Bishnoi expunged critical observations made by the against The company had approached the apex court via Special Leave Petition (SLP) No. 3090/2026, challenging remarks in paragraph 22 of the High Court's , order in Writ Petition No. 8390/2009. The respondent was Deepak s/o Shivkumar Bahry . While granting the expungement as an act of indulgence, the Court issued a stern global warning on the rising peril of AI-generated phantom judgments clogging courts.
From Rent Dispute to AI Citation Scandal
The saga traces back to a dispute under the , where was the appellant before the . In its written submissions filed in February and April 2025, the company allegedly referenced a non-existent case: . No citation or judgment copy was provided, prompting the High Court and its law clerks on an exhaustive—but fruitless—search.
Justice M.M. Sathaye, in the
, lambasted the submissions for hallmarks of AI generation, such as
"green-box tick-marks, bullet-point-marks, [and] repetitive submissions,"
pointing fingers at tools like ChatGPT. The High Court imposed a Rs. 50,000 cost on the appellant, payable to the
, and cautioned that such lapses waste "precious judicial time." It stressed that while AI can aid research, parties bear the "great responsibility" to verify outputs, threatening stricter action like Bar Council referrals for repeat offenders.
The SLP arose solely to scrub these remarks, with the company's director appearing in-person via video conference.
Petitioner's Plea: Denial and Limited Focus
Representing himself, the director of narrowed the SLP to paragraph 22, vehemently denying he cited the phantom judgment. He sought only expungement of the remarks branding his submissions as AI-generated folly, without challenging the underlying rent control ruling.
Respondent's Side: Upholding Judicial Scrutiny
For the respondent, Senior Advocate led a team including Advocates , , , and AOR . Though the hearing transcript doesn't detail their oral submissions, the High Court's order underscores their position: unverified AI dumps hinder justice and demand accountability.
Court's Balancing Act: Mercy Meets Caution
The Supreme Court sidestepped the factual dispute—
"Though, he tried to explain that he never cited that judgment, however, at the present we are not going into that issue"
—opting for expungement
"
."
Yet, it pivoted to a broader indictment, declaring the practice a
"menace...rampant in all Courts now, not only in India rather throughout the world. Everyone needs to be careful about this."
This echoes ongoing Supreme Court concerns; another bench is seized of the issue after a trial court judgment peddled AI-forged citations. Justices have repeatedly flagged unchecked AI in pleadings, blending innovation with peril.
No precedents were cited in the order, as the SLP focused narrowly on expungement rather than substantive law.
Pivotal Quotes from the Bench
-
On the indulgence granted
:
" , we expunge the remarks made in the aforesaid paragraph."
-
Global warning
:
"However, the fact remains that this menace is rampant in all Courts now, not only in India rather throughout the world. Everyone needs to be careful about this."
-
Judicial oversight
:
"In fact, this Court is already on judicial side."
Relief with Ripple Effects
The SLP stands disposed, with pending applications closed. Permission for in-person argument via VC was granted. While the company escapes the stigma, the order reinforces verification mandates, potentially stiffening costs or sanctions in future AI-slip cases. As courts worldwide grapple with generative AI's double edge, this ruling spotlights the urgency: wield tools wisely, or risk judicial wrath.
