Lawyer fined $5K over AI-hallucinated witnesses in a murder case
New Mexico’s Supreme Court fined attorney Stephen Aarons $5,000 and held him in contempt after finding that an appeal brief he submitted relied on AI-generated content that included fabricated witnesses and false police testimony. The court faulted Aarons for not verifying the factual claims and legal authority in the brief, which mischaracterized details about the shooter’s clothing and appearance.

Why It Matters
The ruling highlights legal and ethical risks as attorneys increasingly use generative AI tools without adequate verification. Courts are beginning to impose sanctions when AI-produced material introduces fabricated facts or authorities into filings, signaling stricter scrutiny of AI-assisted work in courtrooms.
Key Facts
- Court: New Mexico Supreme Court
- Attorney: Stephen Aarons
- Penalty: $5,000 fine and held in contempt
- Reason: Failed to verify factual claims and legal authority in an AI-generated appellate brief
- Fabrications: Brief contained fabricated witnesses and false testimony about the shooter’s clothing and appearance (per the court filing)
New Mexico’s highest court sanctioned attorney Stephen Aarons after determining an appellate brief he filed relied on AI-generated material that included invented witnesses and inaccurate testimony. The court imposed a $5,000 fine and held Aarons in contempt, saying he did not adequately confirm the factual assertions or legal authorities cited in the document.
The filing specifically identified false testimony attributed to wholly fabricated witnesses and noted incorrect statements about details such as the shooter’s clothing and appearance. The court’s action followed disclosure that Aarons used an AI tool to help produce the brief and did not sufficiently verify the output before submitting it in a murder-appeal matter.
At an August hearing, the court questioned Aarons’ reliance on generative AI after he acknowledged using ChatGPT and expecting it to produce a reliable summary of the trial. Justice C. Shannon Bacon pressed Aarons on his awareness of the well-publicized risks of AI hallucinations, noting that problems with AI-generated inaccuracies have been widely reported.
The decision comes amid a growing number of incidents in which courts have discovered fabricated citations or misquoted material produced by AI. In recent years judges have faulted law firms for false legal citations and fined other attorneys — including counsel for MyPillow’s Mike Lindell — for submitting briefs containing misquotes and invented authorities. Aarons told Reuters he was remorseful and hoped disciplinary authorities would take into account that the errors were an honest mistake.
Keep Reading

The debate over AI ‘doomsday’ warnings

US lawyer cites fake witnesses in murder case, blames ChatGPT

US legislators push AI safety laws amid human extinction warnings
ChatGPT-using lawyer punished for citing fake testimony from made-up witnesses
Original source: The Verge