New Mexico Lawyer Fined for Filing AI-Generated Brief with Fake Testimony

Court sanctions attorney over AI hallucinations in criminal appeal

A New Mexico defense attorney has been fined $5,000 and held in contempt after submitting a legal brief containing entirely fabricated witness testimony and police statements generated by ChatGPT, according to a ruling from the state's supreme court.

Attorney Stephen Aarons was representing Oscar Renee Sandoval in an appeal of his murder conviction when he used OpenAI's chatbot to summarize trial proceedings. The resulting brief included what the court described as "false testimony from wholly fabricated witnesses," including fictional details such as statements that "the shooter was wearing dark pants and a white shirt."

The New Mexico Supreme Court issued its ruling Wednesday, criticizing Aarons for failing to verify the accuracy of his filing and demonstrating "a lack of remorse and a lack of concern for his client." The justices also referred Aarons to an attorney disciplinary board for further investigation.

Defense of 'honest mistake' falls flat

Aarons told the court during an August 21 hearing that he had fed a computer-generated transcript and case materials to ChatGPT, expecting it would generate "a bulletproof summary." He claimed he didn't fully understand the extent to which AI systems can "hallucinate" facts.

Justice Shannon Bacon expressed disbelief at Aarons' explanation during the hearing: "Counsel, do you watch the news? Do you listen to the radio? Do you read anything about what's going on in the world? Because the problem with lawyers relying on AI hallucinations is an 'above-the-fold' story every single day."

In a statement, Aarons acknowledged his error: "I am remorseful but hopeful that the disciplinary board takes into account it was an honest mistake. It is a lesson learned for all professionals who rely upon this powerful but sometimes unstable technology."

Why it matters

This case represents an escalation in AI-related legal misconduct. While dozens of lawyers have been sanctioned for submitting briefs with fabricated case citations, Aarons' filing appears to be among the first to include entirely invented witness testimony in a criminal appeal—potentially jeopardizing a defendant's constitutional right to effective counsel. The ruling underscores that professional responsibility standards haven't changed despite new technology, and attorneys remain personally accountable for every word filed with courts.

Growing pattern of AI sanctions

The New Mexico case joins a mounting list of incidents where judges have disciplined attorneys for inadequately vetting AI-generated legal documents. State and federal courts have sanctioned lawyers for briefs containing made-up case citations and misquoted law, though fabricated witness testimony in a criminal matter appears to represent new territory.

Sandoval's appeal remains pending. The case was reassigned on September 2 to Kim Chavez Cook, a New Mexico public defender. Sandoval had pleaded not guilty to murdering the mother of his children before being convicted and sentenced to life in prison last year.

OpenAI did not respond to requests for comment on the case, according to Reuters, which first reported these details.