When a public interest litigation came up before the Supreme Court of India earlier this week, the hearing took an unusual turn.
Chief Justice Surya Kant asked the petitioner whether he had drafted the plea himself. The man, who said he had studied only till Class 12 at Sanatan Dharm School in Ludhiana, insisted that he had.
The bench then pressed him on phrases used in the petition—“fiduciary risk to corporate donors” among them—asking him to explain what they meant. When the petitioner struggled, the bench pressed on and, eventually, the litigant admitted he had “searched AI tools” and that a typist had helped prepare the filing.
The Chief Justice remarked that the petitioner appeared to have “lent his shoulders to someone who has drafted a vague, wild petition,” before warning him against filing such PILs lightly.
The episode underscores a broader concern that courts are increasingly encountering: legal filings and judicial reasoning that rely on AI-generated content, sometimes containing fabricated case law or misleading legal language.
The apex court also recently issued a strong warning that reliance on AI-generated or non-existent judgments could amount not merely to an error of judgment but to judicial misconduct.
The bench of Justices PS Narasimha and Alok Aradhe, while taking cognisance of a trial court order that appeared to rely on purportedly AI-generated fake precedents, said that such practices strike at the integrity of the adjudicatory process.
Artificial Paralegal
Legal experts say the concern stems from the central role precedent plays in India’s constitutional framework.
“Under Article 141 of the Constitution of India, the law declared by the SC is binding on all courts in India. In that context, a judgment delivered by the SC is of seminal importance under our constitutional scheme as it is a binding precedent for all courts in India,” explains Kuriakose Varghese, Supreme Court advocate and Managing Partner at KMNP Law, in conversation with AIM.
“In such a context, AI-generated judgments/non-existent judgments that go into the legal reasoning, which ultimately result in the ratio (of the judgment), have the effect of poisoning the whole fruit,” he adds.
The apex court’s decision to distinguish reliance on AI for citations from a good-faith mistake signals a deliberate shift.
“By distinguishing it from an ‘error’ in good faith to a ‘misconduct’, the SC has raised the punitive bar where the intentionality standard is higher. The strict approach has perhaps been adopted by the SC to create a stronger deterrence, to a perceived higher use of AI in the near future,” Varghese observes.
Indian courts have already encountered plenty of instances wherein fabricated precedents surfaced in filings or orders.
The Delhi High Court, in 2025, flagged a petition that cited judgments which could not be traced in any legal database, prompting the petitioner to withdraw the plea.
In another instance before the Gujarat High Court this year, judges questioned citations in a GST adjudication order that lawyers argued resembled AI-generated outputs. The Income Tax Appellate Tribunal has also reportedly recalled an order after lawyers pointed out references to non-existent case law.
The growing use of generative AI tools in legal drafting has complicated the traditional research process, which historically relied on verified law reports and databases. While AI can assist with summarisation and drafting, courts have emphasised that the responsibility for verifying citations rests squarely with the lawyers who file them.
Who is Accountable?
Kuriakose says the ultimate accountability lies with the senior lawyer who signs off on the final filing.
“Any senior partner/lawyer ultimately has the responsibility for the final draft, which is being filed, even if the first draft may be made by his juniors. Anything which goes into a draft which is ultimately filed is to that extent a record for posterity in form or the other,” he notes.
He emphasises that verifying case law remains a fundamental part of legal practice. “Needless to state, case laws have to be verified from at least two or three sources before they find a place in a draft. A reasonably, seasoned lawyer, in my opinion, may be able to instinctively identify bogus citations and case laws.”
However, he notes that errors can sometimes appear subtly. “Sometimes the challenge is when this appears as an extra sentence or paragraph in otherwise well-known judgments. The veracity can only be checked by lawyers, much akin to how the reputed news media does a fact check.”
The Tricke-Down Effect
Despite these risks, many lawyers are already experimenting with AI tools in their workflows. According to the 2026 Wolters Kluwer Future Ready Lawyer Report, this number stands at 90% globally. India is still catching up, with over 40% of general counsels now testing AI-powered legal tech in beta, according to Gartner.
Clarence Anthony, Founder of Clarence & Partners, earlier told AIM that AI is well-suited to the legal sector due to its role in probabilistic analysis.
“We are trying to leverage what we believe are the biggest value drivers, which are AI and its immense capabilities for content generation, research, validation, accuracy, together with the muscle and capability that come with experienced lawyers, and a human sense of how to deal with difficult situations,” he added.
Kuriakose acknowledges that while he has not fully embraced the technology himself, its influence is becoming hard to ignore.
“I know many lawyers who are using it in liberating ways, which frees up their time, money, and resources,” he remarks.
AI is now also encouraging lawyers to pivot to boutique law firms.
Chinmay Bhosale, Co-founder of legal AI platform NYAI, explained to AIM in a previous interaction that AI has removed many of the structural disadvantages that once made scale essential. “When legal research, regulatory tracking, and compliance intelligence are structured and accessible, large teams are no longer a prerequisite for delivering depth,” he commented.
However, Kuriakose believes the effectiveness of such tools ultimately depends on the lawyer using them. “The more successful ones are the ones who have already finished some high-quality intellectual heavy lifting. This allows them to use AI to assist them rather than the other way around.”
For now, courts appear determined to draw a clear line between technological assistance and abdication of professional responsibility.
As Kuriakose put it, “Ultimately, the agility of mind and an analytical mindset capable of connecting the disparate dots are crucial in law. The thinking lawyer cannot be replaced by thinking AI, though even that day may not be far.”
Your reaction