Imagine a courtroom in Delhi. A lawyer walks up confidently, armed with a meticulously prepared written submission. It cites several cases, each seemingly relevant and persuasive. The judge begins to read, pauses and then looks up.
“These cases,” the judge says slowly, “do not exist.”
It sounds like a nightmare scenario, but it is no longer far-fetched. Around the world, lawyers have faced real consequences for submitting AI-generated documents containing citations to judgments that never existed. Generative AI tools like ChatGPT can produce authoritative-sounding text, complete with plausible-looking case names and citations. But when lawyers fail to verify these citations and submit them blindly, courts have been unforgiving. India has not yet had a headline-grabbing incident of this kind, but when it happens, the legal framework already exists to address it.
In 2023, a case in the United States - Mata v. Avianca, Inc - sent shockwaves through the legal fraternity. Two lawyers filed a brief filled with six citations that looked authentic but were entirely fabricated. ChatGPT had generated the names, summaries and even fake quotations from judgments. When the court discovered the deception, it fined the lawyers $5,000 and publicly reprimanded them. The reputational damage was even worse than the financial penalty.
This was not an isolated incident. In Wyoming, in Wadsworth v. Walmart, lawyers cited eight non-existent cases. The court imposed sanctions under Rule 11 of the Federal Rules of Civil Procedure, fined the lawyers and stripped one of his right to appear before that court. In Australia, a Victorian lawyer faced disciplinary action in 2024 after filing submissions with fabricated AI-generated precedents. The court barred him from practicing as a principal lawyer, prohibited him from handling client trust funds and ordered him to work under supervision for two years.
Even large and established firms have been caught off guard. In Alabama, three attorneys from a well-known law firm were disqualified from a case after using AI-generated citations and the matter was referred to the State Bar for possible disciplinary proceedings.
Across jurisdictions, the message is consistent. Lawyers are expected to take personal responsibility for what they submit to courts. Blaming technology will not save anyone. Whether the citation came from a junior intern, a faulty database, or an AI chatbot, the lawyer who signs and files the document is accountable.
While Indian courts have not yet confronted an incident where AI-generated hallucinated citations were submitted, the legal system already has the tools to respond. Submitting fabricated or non-existent judgments, knowingly or recklessly, could expose a lawyer to criminal prosecution, contempt proceedings and disciplinary action by the Bar Council.
If a lawyer knowingly files fabricated citations, this could amount to giving false evidence under Section 193 of the Indian Penal Code, now mirrored in Section 229 of the Bharatiya Nyaya Sanhita, 2023. The punishment can extend up to seven years of imprisonment along with a fine. While this provision is aimed at witnesses and parties, lawyers are not exempt if they are complicit in misleading the court.
Even where there is no deliberate intent to mislead, Indian courts have the power to initiate proceedings under Section 340 of the Code of Criminal Procedure (Section 379 of Bharatiya Nagarik Suraksha Sanhita, 2023). This allows the court to direct an inquiry into whether false evidence was given and decide whether the matter should be prosecuted. The fact that the false citation came from an AI tool does not change the underlying duty of care owed to the court. Lawyers are expected to verify authorities independently before relying on them.
Submitting fabricated citations can also amount to contempt of court. Under the Contempt of Courts Act, 1971, any act that tends to obstruct the administration of justice qualifies as contempt. Deliberately misleading the court, or even acting with reckless disregard for the truth, could invite contempt proceedings. Punishment can include up to six months’ imprisonment, a fine of up to ₹2,000, or both. Indian courts have historically treated attempts to mislead them very seriously. Even when lawyers make inaccurate representations unintentionally, judges have often demanded unconditional apologies. The judiciary has made it clear that lawyers owe a duty of candour to the court and that their role is to assist in the administration of justice, not undermine it.
Aside from criminal and contempt proceedings, submitting false citations can also be treated as professional misconduct under the Advocates Act, 1961. Section 35 empowers State Bar Councils and the Bar Council of India to discipline advocates found guilty of misconduct. Penalties can range from a reprimand to suspension of licence or even permanent debarment from practice. Indian courts have used this mechanism before.
In a case before the Bombay High Court, lawyers were referred to the State Bar Council for making misleading submissions. The Court ultimately withdrew the referral after the lawyers offered unconditional apologies and explained that the mistake was inadvertent. That example shows that courts distinguish between deliberate misconduct and genuine errors, but the fact remains that the referral itself is damaging to a lawyer’s reputation and career.
Yes, intent matters a great deal. If a lawyer deliberately submits fabricated citations knowing they are false, the consequences can be severe. They could face prosecution for giving false evidence, contempt of court and disciplinary action by the Bar Council, all at the same time. But if the error is genuinely unintentional, courts are likely to adopt a more measured approach. In such situations, the response often depends on how the lawyer behaves once the mistake is discovered. If they proactively inform the court, withdraw the incorrect citations and apologise, courts have shown willingness to be lenient. However, if the error comes to light through the court’s own scrutiny and the lawyer is seen as careless, even an honest mistake can have serious professional consequences.
Indian judges are already aware of the risks posed by AI in legal practice. In 2025, Supreme Court Justice (now CJI) BR Gavai publicly warned that tools like ChatGPT can generate fake case citations and urged lawyers to verify everything before including it in court filings. The legislature has begun exploring guidelines on AI-assisted research, underscoring that while technology can be helpful, it does not reduce the professional responsibility of lawyers to ensure accuracy.
In other words, if a lawyer in India submits AI-generated hallucinated citations tomorrow, they cannot claim ignorance as a defence. The expectation of diligence remains unchanged.
Generative AI can be powerful, but it also hallucinates. For lawyers, that means a simple rule: verify everything. Courts will not care whether the mistake came from a junior associate, a database glitch or an AI chatbot. In the eyes of the law, responsibility always rests with the lawyer who signs and files the submission.
As AI becomes more common in the profession, lawyers in India have an opportunity to learn from global missteps. Technology can make us faster, but it cannot make us careless. A fabricated citation, no matter how convincingly written, can derail a case and damage a career. The safest path forward is to embrace the tools, but never surrender the diligence that defines good lawyering.
Mahi Agrawal is a third-year B.A. LL.B. (Hons.) student at Hidayatullah National Law University, Raipur.