New Mexico Supreme Court Fines Lawyer $5,000 Over Fabricated Testimony in AI-Assisted Brief

The New Mexico Supreme Court has fined attorney Stephen Aarons $5,000 and held him in contempt after finding fabricated police testimony and fictional witnesses in a murder-appeal brief prepared with ChatGPT assistance. The court will also refer him for a disciplinary investigation.

The sanction concerns the appeal of Oscar Renee Sandoval, who was convicted of murder and sentenced to life in prison. That appeal remains pending. The court said Aarons’s primary brief contained false testimony from wholly fabricated witnesses, putting invented material into a filing in a criminal case.

How a generated summary entered the appeal

Aarons told the court that he fed ChatGPT a computer-generated transcript and other case materials. He said he expected what he called a “bulletproof summary” while preparing the appeal. But the court found the submitted brief included material that had no basis in the record.

The distinction is straightforward but consequential: supplying source material to a chatbot does not ensure that its prose remains tied to that material. Aarons was held in contempt for failing to verify the filing’s accuracy before it reached the court.

Not a citation error

The court said the filing appeared to include fictional statements that the shooter wore dark pants and a white shirt. Alongside the finding of fabricated witnesses, that detail makes this more than a dispute over an imprecise legal citation or paraphrase.

Aarons’s explanation

Aarons said he had not understood the degree to which AI could hallucinate facts. He characterized the episode as an honest mistake and said he was remorseful, while hoping the disciplinary board would take that explanation into account.

The unresolved professional question now moves to the disciplinary process. For Sandoval’s case, the immediate reality is separate: his life-sentence appeal is still active, even though the lawyer who submitted its primary brief has been sanctioned over its accuracy.