Supreme court issues fine to lawyer for using ChatGPT in murder trial

The New Mexico Supreme Court recently held Santa Fe attorney Stephen Aarons in contempt after he submitted an appellate brief containing testimony and witnesses invented by ChatGPT. Last year, Aarons was defending a client convicted of murder when he used the chatbot to condense trial records, only to file the resulting document without any prior proofreading. The brief included statements attributed to witnesses who did not exist, along with fabricated remarks falsely credited to a real witness.

The court's response was blunt. Justice C. Shannon Bacon pointedly asked Aarons whether he followed the news at all, given how often AI "hallucinations" in legal filings have made headlines, and suggested he had either ignored the phenomenon or recklessly gambled with his client's case. Aarons, who has practiced law for over four decades, conceded his own "stupidity," further stating that he had wrongly assumed that the software would produce a reliable summary.

Beyond the rebuke, the court imposed a $5,000 fine, payable to a client protection fund, and referred the matter to the state's attorney disciplinary board. The case follows several other instances in which lawyers have faced sanctions or disciplinary scrutiny after submitting unverified AI-generated material in court (1, 2, 3).