New Mexico Attorney Sanctioned After Artificial Intelligence Fabricates Witnesses in Murder Appeal
Santa Fe, Wednesday, 30 September 2026.
Attorney Stephen Aarons received a $5,000 fine and court sanctions after submitting a murder appeal brief containing fake witness testimony generated by ChatGPT without performing manual verification.
New Mexico Attorney Sanctioned After Artificial Intelligence Fabricates Witnesses in Murder Appeal
Attorney Stephen Aarons received a $5,000 fine and court sanctions after submitting a murder appeal brief containing fake witness testimony generated by ChatGPT without performing manual verification [1][3]. The incident, finalized in September 2026, involved the submission of wholly fabricated witnesses, including nonexistent individuals named Danny Stanton and Linda Stanton, within a legal brief for a client convicted of killing his wife earlier in the year [1][2]. The New Mexico Supreme Court issued an order of direct contempt against Aarons, emphasizing that attorneys retain full liability for the veracity of all documents they sign and submit to the court [3]. This case marks a significant escalation in judicial responses to artificial intelligence misuse, moving beyond citation errors to the invention of factual evidence [1].
Judicial Scrutiny Intensifies
During a hearing held in August 2026, appellate judges questioned the attorney’s professional due diligence and reliance on unverified generative artificial intelligence outputs [1]. One judge explicitly asked, Do you watch the news? Do you listen to the radio? Do you read anything about what’s going on in the world? highlighting that the problem with lawyers relying on AI hallucinations is an above-the-fold story every single day [1]. The court noted that Aarons assumed ChatGPT would generate a bulletproof summary of proceedings, but the technology instead produced false testimony regarding the shooter’s clothing and appearance [2]. As part of the disciplinary sanctions, Aarons is barred from appearing before the court pending an investigation by the Disciplinary Board [1].
Communication Breakdown with Client
On 2026-09-11, reports indicated that Aarons claimed he did not inform his incarcerated client about the AI error [1]. Although the client was serving a life sentence, the attorney admitted to telling family members only that there was a problem with the brief, without specifying the use of ChatGPT [1]. The presiding judge criticized this lack of transparency, stating, So you didn’t say, I relied on ChatGPT, and it submitted lies to the court, and the court is mad at me, and now there’s a hearing [1]. This failure to communicate critical developments to the client underscores the ethical risks associated with unchecked automation in legal representation [3].
Implications for Professional Liability
This incident follows a pattern of AI-related legal errors, including a May 2026 case where judges in the Supreme Court of the State of New York Appellate Division reprimanded several lawyers for similar failures [1]. The legal consensus established by this case is that generative AI is strictly an auxiliary tool and cannot be used as a defense for professional negligence [3]. Corporate leadership and professional service firms integrating large language models must now implement robust governance protocols to avoid rising operational, compliance, and legal liabilities [1]. When AI begins to fabricate facts, any direct reliance on it without human review is a fatal injury to substantive justice [3].