What happened
The New Mexico Supreme Court sanctioned attorney Stephen Aarons for including AI-generated false testimony in a criminal appeal brief.
The New Mexico Supreme Court issued a sanction on September 9 against Stephen Aarons, a private attorney based in Santa Fe, for his role in the appeal of Oscar Renee Sandoval, who was convicted of murder in 2025. The court found that Aarons submitted a brief containing 'false testimony from wholly fabricated witnesses' and 'fictional statements' regarding the appearance of the shooter, which he admitted were generated by OpenAI's ChatGPT.
Aarons stated in a hearing on August 21 that he used ChatGPT to summarize trial proceedings and case materials, presuming the output would be accurate. He claimed he did not fully understand the extent to which AI could 'hallucinate' facts. The court rejected this defense, with Justice C. Shannon Bacon questioning whether Aarons was aware of the widespread reporting on AI hallucinations, stating that the issue is an 'above-the-fold story every single day.'
The court fined Aarons $5,000 and held him in contempt for failing to verify the accuracy of the filing. The justices also noted his 'lack of remorse and a lack of concern for his client.' The court referred Aarons to an attorney disciplinary board for further investigation. The appeal itself remains pending and has been reassigned to a public defender.
Source details: straitstimes.com ↗
Why it matters
This ruling marks a significant escalation in AI-related legal sanctions, moving beyond fabricated citations to the creation of entirely fictional witness testimony in a criminal case. It highlights the severe risks of relying on generative AI for legal fact-finding without rigorous human verification.
While previous AI-related legal sanctions have primarily involved fabricated case citations or misquoted laws, this incident involves the fabrication of witness testimony in a criminal appeal. This represents a more severe breach of legal ethics and procedural integrity, as it introduces non-existent evidence into the judicial record.
The ruling underscores the professional obligation of attorneys to verify all materials submitted to the court, regardless of their source. It serves as a concrete precedent that using AI as a 'black box' for fact-generation is not a defensible strategy in legal practice, particularly in high-stakes criminal cases.
This case contributes to a growing body of evidence that generative AI tools, while useful for summarization, are unreliable for generating factual content that requires absolute accuracy. It may influence how legal professionals and courts approach the use of AI in future proceedings.
What to watch next
Monitor the outcome of the attorney disciplinary board investigation and whether other jurisdictions adopt similar sanctions for AI-generated factual errors in court filings.
The outcome of the disciplinary board investigation into Aarons' conduct, which could result in additional penalties such as suspension or disbarment.
Whether other state or federal courts cite this New Mexico ruling in their own sanctions against attorneys for AI-related errors, potentially establishing a broader standard for AI accountability in the legal system.
Developments in the Sandoval appeal itself, now handled by a public defender, to see if the fabricated testimony has caused any procedural delays or requires a new hearing.