Lawyer says he didn’t realize ChatGPT could ‘hallucinate’ after getting caught submitting fake testimony in murder appeal

As technology grows more advanced, Artificial Intelligence (AI) has become a go-to resource for people seeking knowledge-based assistance. However, it isn’t always guaranteed that the information given is accurate. In the case of one lawyer, his utilization of OpenAI’s ChatGPT led him to receive a court fine after he was caught submitting testimony that didn’t exist in a legal brief during a murder appeal.

The Guardian reported that Santa Fe-based Stephen Aarons was representing Oscar Renee Sandoval as his defense lawyer after the latter was accused of killing his partner and the mother of his children, Shiereen Al-Jibury, in 2025. Sandoval was later found guilty and sentenced to life in prison after pleading not guilty in February of the same year.

After Sandoval was found guilty, Aarons, who has over 40 years of legal experience under his belt, reportedly submitted a legal brief that “contained false testimony from wholly fabricated witnesses” in August 2025. A few weeks after, a motion was filed by the state of New Mexico to remove the fabricated portions of the brief.

Lawyer wanted a ‘bulletproof’ brief in appeal using ChatGPT

The New Mexico Supreme Court reportedly requested that Aarons explain how the fake testimonies were included in the appeal, which was said to have contained unsubstantiated statements that the “shooter ​was wearing dark pants and a white shirt.” The lawyer had used ChatGPT by feeding it transcripts from Sandoval’s trial and case materials, according to the court.

Aarons claimed that the AI app would give out a “bulletproof summary” for the case during an Aug. 21 hearing. But what he got instead were fake witnesses named Officer Sanchez, Officer Michelle Amarillo, Teresa Marquez, and Manal Al-Jibury.

On Sept. 11, Aarons stood before the court, where he argued that he wasn’t aware that ChatGPT could “hallucinate” facts to that degree, yet acknowledged his error in judgment. He claimed to be remorseful, but asked for the disciplinary board to consider that it was an “honest mistake” on his part. He added, “It is a lesson learned for all professionals who rely upon this powerful but sometimes unstable technology.”

Supreme Court Justice Shannon Bacon, however, was more critical of the alleged mistake. “Counsel, do you watch the news? Do you listen to the radio? Do you read anything about what’s going on in the world?” Bacon said. “Because the problem with lawyers relying on AI hallucinations is an ‘above-the-fold’ story every single day.”

Aarons was found to have “demonstrated a lack of remorse and a lack of concern for his client” by a panel during the hearing. He was later fined $5,000 and referred to an attorney disciplinary board to undergo an investigation after being held in contempt. Sandoval’s appeal is still pending, per the outlet, but was reassigned to public defender Kim Chavez Cook.

Have a tip we should know? [email protected]