Lawyer Fined Used ChatGPT to Invent Witnesses in Murder Appeal

A New Mexico lawyer has been fined $5,000 after submitting a murder appeal containing fabricated witnesses and police testimony generated by ChatGPT, without checking whether the information was accurate.

The New Mexico Supreme Court found attorney Steven Aarons in contempt after his appeal filing included false testimony from entirely fictional witnesses and statements that did not appear in the trial record.

Aarons admitted that he prepared the filing with the assistance of artificial intelligence.

According to the court, the filing contained “false testimony from entirely fabricated witnesses”, raising serious concerns about the lawyer’s conduct and his failure to properly represent his client.

The court imposed the $5,000 fine and referred Aarons to the state bar’s disciplinary board.

Aarons, who practises in Santa Fe, told Reuters that he initially used ChatGPT to summarise the trial after taking on the case. He said he did not understand the extent to which the AI system could produce what are commonly known as “hallucinations” — false or fabricated information presented as fact.

Aarons said he regretted what happened and hoped the disciplinary board would recognise it as an honest mistake.

“It is a lesson learned the hard way for all professionals who rely on this powerful but sometimes unstable technology,” he said.

OpenAI did not respond to a Reuters request for comment.

Growing scrutiny of AI use in US courts

The case comes as state and federal judges across the United States increasingly impose disciplinary measures on lawyers who submit AI-generated material without adequately checking it.

Some judges have also faced criticism for using AI tools themselves, while numerous lawyers have faced sanctions after artificial intelligence misinterpreted or fabricated legal authorities and other information.

The growing number of cases has highlighted the risks of relying on generative AI for legal work without independent verification.

Murder defendant’s appeal

Aarons represented Oscar Rene Sandoval in his appeal. Sandoval maintains that he did not kill the mother of his children.

A jury found Sandoval guilty last year and sentenced him to life imprisonment. His appeal remains pending, and a new attorney, public defender Kim Chavez Cook, was appointed to the case in early September.

Court questions fabricated testimony

Last month, the New Mexico Supreme Court ordered Aarons to explain how fabricated statements had appeared in his appeal.

One passage attributed a statement to a supposed witness that “the perpetrator was wearing dark pants and a white shirt”.

Aarons said he had provided ChatGPT with trial transcripts and other material because he believed the system would produce an “accurate summary” of the proceedings.

The justices questioned whether Aarons could reasonably have been unaware of the risks associated with AI-generated legal material.

Justice S. Shannon Bacon pointed to the extensive public reporting about lawyers submitting fabricated AI-generated material in court filings.

The case adds to growing concerns within the US legal system over the use of generative AI and the responsibility of lawyers to verify every factual and legal claim before submitting documents to a court.

Stay updated with the latest news from Greece and around the world on greekcitytimes.com.
Contact our newsroom to share your updates, stories, photos, or videos. Follow GCT on Google News and Apple News.