New Mexico attorney Stephen Aarons fined $5,000 over ChatGPT-hallucinated witnesses
The state's Supreme Court holds a criminal defense lawyer in contempt for submitting fake testimony in a murder appeal.
Original cover art, generated for this story. THE VISSION does not republish third-party press imagery.
- The New Mexico Supreme Court has fined defense attorney Stephen Aarons $5,000 and held him in contempt for submitting AI-fabricated witnesses.
- Aarons admitted to using ChatGPT to generate a summary of a murder trial, failing to verify the fictitious testimony and witnesses it created.
- Justice Bacon rebuked Aarons during the hearing, warning that AI hallucination risks are now an above-the-fold story.
The Supreme Court of New Mexico has held defense attorney Stephen Aarons in contempt of court and imposed a $5,000 fine after he submitted an appellate brief containing entirely fabricated witnesses and fictitious police testimony. The ruling marks a severe escalation in judicial penalties for legal professionals who fail to supervise artificial intelligence tools, with the state's highest court warning that claims generated by commercial chatbots must be rigorously verified before they are introduced into judicial records.
Aarons, a prominent criminal defense attorney in Santa Fe, admitted to the court that he used OpenAI's ChatGPT to generate a 'bulletproof summary' of a complex homicide trial for his client's appeal. The chatbot, however, hallucinated several critical witnesses and detailed cross-examination testimonies that never occurred in the actual trial. Aarons copied the AI-generated summaries directly into his formal court submission without cross-checking the references against the official stenographic transcript of the trial, a failure of basic due diligence that was quickly flagged by state prosecutors.
During the tense hearing, New Mexico Supreme Court Justice C. Shannon Bacon sharply rebuked Aarons, asking, 'Counsel, do you watch the news?' Justice Bacon noted that artificial intelligence hallucinations and their legal consequences are now 'above-the-fold stories every single day,' leaving lawyers with no excuse for failing to verify their filings. Aarons accepted the fine and the contempt finding, stating that he was deeply embarrassed by his technological complacency and has implemented a complete ban on generative tools in his practice.
This $5,000 fine is a vivid reminder that the legal profession’s ethical duty of competent representation cannot be outsourced to generative AI. While early instances of 'AI-hallucinated' briefs were met with judicial warnings, courts are now moving to aggressive financial and professional sanctions. This contempt ruling sets a firm precedent that attorneys bear absolute personal and professional liability for the accuracy of AI-generated content, forcing firms to establish strict compliance checks.