Lawyer fined $5K for using AI-generated fake witnesses in murder case
New Mexico Supreme Court sanctions attorney over ChatGPT-hallucinated evidence in appeal.
1The case
New Mexico's Supreme Court fined attorney Stephen Aarons $5,000 for submitting an appeal brief containing completely fabricated witnesses and false police testimony generated by ChatGPT. The brief was part of an appeal for his client's murder conviction. Justice C. Shannon Bacon held Aarons in contempt for failing to verify the AI-generated content, calling it a violation of professional responsibility in legal proceedings.
2The fabricated evidence
Aarons' brief included false testimony about the shooter's clothing and appearance, along with entirely made-up witness statements. The court filing explicitly states these were 'wholly fabricated witnesses' created by AI. This marks one of the most severe documented cases of AI hallucination affecting legal outcomes, potentially jeopardizing a murder case appeal.
3The judge's reaction
During an August hearing, Justice Bacon expressed astonishment at Aarons' lack of awareness about AI risks, asking if he followed current events. 'The problem with lawyers relying on AI hallucinations is an above-the-fold story every single day,' Bacon remarked. The judge referenced widespread media coverage about AI's tendency to fabricate information in legal contexts.
4Growing trend of AI misuse
This incident follows multiple cases where lawyers faced sanctions for AI-generated legal documents. In 2025, two law firms were penalized for submitting briefs with false citations, while MyPillow CEO Mike Lindell's attorneys were fined for similar AI fabrications. Legal experts note a 300% increase in such cases since ChatGPT's public release in late 2022.
5The lawyer's defense
Aarons claimed he believed ChatGPT would produce a 'bulletproof summary' of the trial and called his mistake 'honest' in a statement to Reuters. He expressed remorse but hoped the disciplinary board would consider his intentions. Legal ethics experts argue this defense highlights a dangerous misunderstanding of AI's limitations in professional contexts.
6Broader implications
The case raises critical questions about attorney competence in the AI era. State bars are considering mandatory AI literacy training for lawyers. The American Bar Association may update Model Rule 1.1 (competence) to explicitly address technological proficiency, potentially making such sanctions more common for AI misuse.
7AI in legal practice
While 68% of law firms now use some form of AI according to 2026 surveys, only 22% have formal verification protocols. Legal tech companies are developing specialized AI tools with built-in citation checking, but many small practices still rely on general-purpose chatbots. This case demonstrates the risks of unverified AI use in high-stakes litigation.
8The precedent
This ruling sets an important benchmark for AI accountability in legal professions. The $5,000 fine and contempt finding send a clear message about verifying AI outputs. Legal scholars predict it will accelerate professional standards around AI use, potentially preventing more serious miscarriages of justice in future cases.