The New Mexico Supreme Court has held attorney Stephen Aarons in direct contempt of court after he submitted a legal brief containing false testimony from fabricated witnesses generated by ChatGPT. The court referred Aarons to a disciplinary board, noting his failure to verify AI-generated content before filing.
What Happened
Aarons, a criminal defense lawyer with over 40 years of experience in New Mexico, filed an appeal for his client, Oscar Renee Sandoval, who was sentenced to life in prison in February 2025 for the murder of his partner, Shiereen Al-Jibury. The brief, which Aarons admitted to generating with AI, included fake police testimony and other legal misrepresentations. According to the state's top court, Aarons did not verify the factual claims or legal authority in the document before signing and filing it, nor did he inform his client of these errors. The court order stated that Aarons 'demonstrated a lack of remorse and a lack of concern for his client.'
Why It Matters
This case highlights the ongoing risks of integrating large language models into professional legal workflows without rigorous human verification. While AI tools can accelerate drafting, they remain prone to hallucination, particularly when generating specific factual details or legal citations. The New Mexico Supreme Court's decision reinforces the responsibility of attorneys to act as final editors and validators of any AI-assisted work product, especially in high-stakes criminal appeals where accuracy is critical to a client's freedom.
The Bottom Line
Aarons has been referred for further disciplinary proceedings, underscoring that reliance on AI in legal practice requires strict oversight to prevent fabricated evidence from entering court records.