Blab AI
New Mexico Supreme Court Fines Defense Lawyer $5,000 for AI-Generated Fake Testimony
On Wednesday, the New Mexico Supreme Court imposed a $5,000 fine on Santa Fe defense attorney Stephen Aarons after he submitted a criminal appeal brief laced with fabricated police testimony and invented witness statements. The brief had been produced with the assistance of OpenAI’s ChatGPT, a large‑language model capable of generating convincing text yet prone to hallucinations.
Aarons was defending Oscar Renee Sandoval, who is appealing a first‑degree murder conviction and a life sentence. Sandoval denied responsibility for the killing of his children’s mother and received his sentence last year. The appeal remains pending and has now been assigned to public defender Kim Chavez Cook.
At a hearing on August 21, Aarons told the court that he had fed a computer‑generated transcript and other case documents into ChatGPT, expecting the chatbot to produce a “bulletproof summary.” He admitted he did not grasp how easily the model could fabricate facts. In a statement, Aarons expressed remorse and asked the disciplinary board to consider the incident an honest mistake, noting that it had served as a lesson for all professionals who depend on powerful yet sometimes unstable technology.
The court ruled that the brief contained false testimony from wholly fabricated witnesses and noted Aarons’ lack of remorse and concern for his client. Justice C. Shannon Bacon questioned the lawyer’s awareness of AI risks, asking, “Counsel, do you watch the news? Do you listen to the radio? Do you read anything about what’s going on in the world?” and adding, “Because the problem with lawyers relying on AI hallucinations is an above‑the‑fold story every single day.”
Aarons was found in direct contempt of court and fined $5,000. The panel also referred him to the attorney disciplinary board for further investigation.
The incident fits into a wider pattern across the United States, where courts have disciplined lawyers for submitting briefs that contain AI‑generated errors—including fabricated citations and incorrect legal references. Some judges have also faced scrutiny over their own use of AI tools.
OpenAI’s ChatGPT, launched in November 2022, is widely employed for drafting documents, answering questions, and summarizing information. Nevertheless, the model is prone to hallucinations, generating statements that sound credible but are not grounded in fact. Legal professionals are advised to verify any output before submitting it to a court.
The New Mexico Supreme Court, the state’s highest appellate court, convenes in the New Mexico Supreme Court Building in Santa Fe. Its ruling underscores the need for careful oversight when employing generative AI in legal practice.
The result of Aarons’ disciplinary proceedings is still pending. Sandoval’s appeal continues under the supervision of a public defender, and the court has not yet ruled on the merits of the appeal.
The incident underscores the growing need for clear guidelines governing the use of AI in legal contexts and the potential consequences that arise when those guidelines are ignored.
Aarons was defending Oscar Renee Sandoval, who is appealing a first‑degree murder conviction and a life sentence. Sandoval denied responsibility for the killing of his children’s mother and received his sentence last year. The appeal remains pending and has now been assigned to public defender Kim Chavez Cook.
At a hearing on August 21, Aarons told the court that he had fed a computer‑generated transcript and other case documents into ChatGPT, expecting the chatbot to produce a “bulletproof summary.” He admitted he did not grasp how easily the model could fabricate facts. In a statement, Aarons expressed remorse and asked the disciplinary board to consider the incident an honest mistake, noting that it had served as a lesson for all professionals who depend on powerful yet sometimes unstable technology.
The court ruled that the brief contained false testimony from wholly fabricated witnesses and noted Aarons’ lack of remorse and concern for his client. Justice C. Shannon Bacon questioned the lawyer’s awareness of AI risks, asking, “Counsel, do you watch the news? Do you listen to the radio? Do you read anything about what’s going on in the world?” and adding, “Because the problem with lawyers relying on AI hallucinations is an above‑the‑fold story every single day.”
Aarons was found in direct contempt of court and fined $5,000. The panel also referred him to the attorney disciplinary board for further investigation.
The incident fits into a wider pattern across the United States, where courts have disciplined lawyers for submitting briefs that contain AI‑generated errors—including fabricated citations and incorrect legal references. Some judges have also faced scrutiny over their own use of AI tools.
OpenAI’s ChatGPT, launched in November 2022, is widely employed for drafting documents, answering questions, and summarizing information. Nevertheless, the model is prone to hallucinations, generating statements that sound credible but are not grounded in fact. Legal professionals are advised to verify any output before submitting it to a court.
The New Mexico Supreme Court, the state’s highest appellate court, convenes in the New Mexico Supreme Court Building in Santa Fe. Its ruling underscores the need for careful oversight when employing generative AI in legal practice.
The result of Aarons’ disciplinary proceedings is still pending. Sandoval’s appeal continues under the supervision of a public defender, and the court has not yet ruled on the merits of the appeal.
The incident underscores the growing need for clear guidelines governing the use of AI in legal contexts and the potential consequences that arise when those guidelines are ignored.
