On September 9, lawyer Steven Aarons was fined $5,000 for submitting appeal documents containing fabricated testimony generated by the ChatGPT chatbot. This was stated in a decision by the New Mexico Supreme Court.
Aarons represented Oscar Rene Sandoval, who was sentenced in February 2025 to life imprisonment for the murder of Shirin Al-Jibouri and an additional four and a half years for assaulting a witness with a weapon.
How the Fabricated Testimony Entered the Documents
According to Ars Technica, the lawyer uploaded court transcripts and other case materials into ChatGPT to obtain a summary. In August 2025, he submitted the document prepared with the chatbot to the court.
A few weeks later, the prosecution demanded the removal of certain information from it.
The appeal documents included four fictitious witnesses, including two police officers. Real participants were also attributed with false testimony about threats from the convicted individual, as well as the clothing and appearance of the shooter.
Additionally, the document distorted the content of two legal precedents.
Aarons admitted that he signed and submitted the text without verifying the facts and legal references. He did not inform his client about the errors or the court’s investigation into his conduct.
In a comment to Reuters, the lawyer explained that he did not understand the extent to which AI could fabricate facts.
Court’s Decision
The court found Aarons guilty of contempt and referred the case to the disciplinary board. He was barred from appearing before the state’s Supreme Court until the investigation is concluded.
The fine must be paid to the state bar’s client protection fund within 30 days.
All previously submitted written positions by the parties were removed from the case. A new attorney from the public defender’s office was assigned to prepare the appeal. The hearing is scheduled to continue in the 2026-2027 court term.
During the hearing, judges emphasized that the responsibility for verifying a document remains with the signing lawyer, regardless of the tool used. Subsequently, Aarons told Ars Technica:
“I regret it, but I hope the disciplinary board considers it a good faith mistake.”
Earlier, in May 2023, lawyer Steven Schwartz submitted fabricated ChatGPT precedents in a case against airline Avianca. He also claimed he was unaware of the chatbot’s ability to generate false information.
