A New Mexico Lawyer Was Fined $5,000 for Submitting AI-Invented Witnesses to the State Supreme Court
Stephen Aarons used ChatGPT to write an appeal in a murder case. The chatbot made up witnesses, invented testimony, and got clothing details wrong. The court was not impressed.
Key points
- New Mexico's Supreme Court fined attorney Stephen Aarons $5,000 on Wednesday, 10 September 2026, for submitting a brief containing AI-fabricated content.
- The appeal brief, written with ChatGPT, included false testimony from witnesses who do not exist and invented details about the shooter's appearance.
- Justice C. Shannon Bacon told Aarons that lawyers submitting fake AI citations is "an above-the-fold story every single day."
- Aarons said he believed ChatGPT would produce a "bulletproof summary" of the trial and called the incident an honest mistake.
Stephen Aarons is a New Mexico attorney whose client was appealing a murder conviction. To prepare the brief, a formal legal document presented to a court, he turned to ChatGPT, the AI chatbot made by OpenAI.
The chatbot hallucinated. That is the word the AI industry uses when a model generates confident, fluent text that is simply not true. Here, ChatGPT produced witnesses who never existed and invented testimony about what the shooter was wearing.
New Mexico's Supreme Court found out. On Wednesday, as first reported by The Verge citing Reuters, the court held Aarons in contempt and fined him $5,000 for failing to "verify the factual claims and legal authority in his AI-generated brief."
Why did the court react so sharply?
Because a lawyer's job is to verify what they submit. Aarons admitted at an August 2026 hearing that he used ChatGPT expecting it to produce a "bulletproof summary" of the trial. He did not check the output.
Justice C. Shannon Bacon did not hide her frustration. "Counsel, do you watch the news? Do you listen to the radio? Do you read anything about what's going on in the world?" she said, according to Reuters. "Because the problem with lawyers relying on AI hallucinations is an above-the-fold story every single day."
She has a point. Last year, a judge criticised two law firms for filing a brief packed with false, misleading legal citations and quotations. Lawyers for Mike Lindell, founder of bedding company MyPillow, were fined separately for submitting fake AI citations and misquotes. AI2Day first reported on AI hallucinations reaching courtrooms on 12 August 2026, and the cases have kept coming.
The pattern is consistent: a legal professional trusts the chatbot, skips verification, and a court catches invented facts.
What does this mean for ordinary people?
If your lawyer uses AI to draft documents affecting your case, those documents can contain invented evidence. You would not know unless a court or opposing counsel caught it.
Aarons said he is "remorseful" and called the episode an honest mistake. That may be true. A murder conviction appeal built on fabricated witnesses is not a minor paperwork error, though. The client's liberty was directly at stake.
Ask your lawyer whether they use AI tools and how they verify AI-generated content before it goes to court. A good lawyer will have a clear answer ready.
What should you watch for?
The problem here is not really ChatGPT. It's the uncritical trust some professionals place in any tool that sounds authoritative. Chatbots generate plausible text, not guaranteed facts, and that gap between "sounds right" and "is right" keeps landing people in proceedings for the wrong reasons.
If you are involved in legal proceedings, look out for citations in your case documents that seem oddly vague or that your lawyer cannot explain. Ask directly whether AI was used and how any output was checked. You are entitled to that answer.
