New Mexico's Supreme Court has fined attorney Stephen Aarons $5,000 and held him in contempt for submitting an appeal brief in a murder case that included fabricated witnesses and fake police testimony generated by artificial intelligence. The court found that Aarons failed to verify the factual claims and legal authority in his AI-generated brief, according to a filing reported by Reuters and The Verge.
The brief, filed on behalf of a client convicted of murder, contained what the court described as "false testimony from wholly fabricated witnesses," as well as false testimony about the shooter's clothing and appearance. The sanctions order, issued Wednesday, represents one of the most severe judicial rebukes yet in the rapidly growing number of cases where lawyers have relied on AI chatbots that produce plausible but nonexistent legal citations, quotes, and evidence.
How the AI hallucinations unfolded
According to The Verge, the New Mexico Supreme Court's filing says Aarons's brief included fabricated witnesses and fake police testimony. The court held him in contempt for failing to "verify the factual claims and legal authority in his AI-generated brief." At an August hearing, Justice C. Shannon Bacon questioned how Aarons could have been unaware of the risks posed by AI. Her skepticism underscored the court's view that lawyers cannot delegate their duty of candor to a machine.
Aarons, according to reports from MSN and The Detroit News, blamed ChatGPT for the fake witnesses. The headlines from those outlets frame the case as a lawyer pointing the finger at the popular AI chatbot after it produced false information. But the court's order focused on the lawyer's professional responsibility. Regardless of which tool generated the text, the filing said, Aarons signed and submitted it without checking its contents.
The brief contained false testimony from wholly fabricated witnesses and false testimony about the shooter's clothing and appearance.
The case is a stark illustration of the "hallucination" problem that has plagued large language models. These systems are designed to predict likely sequences of words, not to retrieve verified facts. When asked for legal authority or witness statements, they can invent names, quotes, and details that sound authentic. In a criminal appeal, where liberty is at stake, such errors are especially dangerous.
A broader warning from Illinois
The New Mexico sanction arrives alongside a separate warning from Illinois. According to Law.com, the Illinois Appeals Court said it hopes a higher fine will deter AI hallucinations in legal filings. The court's message is clear: as AI use becomes more common in law offices, judges are willing to impose escalating penalties to protect the integrity of the judicial process.
That perspective differs from the more narrow, case-specific coverage in The Verge, which focuses on the New Mexico court's contempt finding and Justice Bacon's pointed questions. MSN and The Detroit News emphasize the lawyer's attempt to blame ChatGPT. Law.com, by contrast, treats the issue as a systemic deterrence problem. Together, the reports show a legal system struggling to adapt to tools that can generate persuasive nonsense at scale.
Why it matters
The American legal system depends on verifiable evidence and accurate citations. When a lawyer submits a brief, the court assumes the lawyer has checked the facts. AI-generated hallucinations break that assumption. In recent years, courts across the country have sanctioned lawyers for citing nonexistent cases, misquoting opinions, and inventing legal doctrines. The New Mexico case goes further by involving fabricated witnesses and fake police testimony in a murder appeal.
Legal ethicists say the solution is not to ban AI outright but to require strict verification. Lawyers can use AI for brainstorming, organizing arguments, or drafting routine language, but they must independently confirm every factual claim and citation. Bar associations in several states have issued guidance urging lawyers to understand AI's limitations and to supervise its output. The New Mexico court's $5,000 fine and contempt holding are a reminder that those duties are enforceable.
- Verify everything: AI-generated citations, quotes, and witness statements must be checked against primary sources.
- Understand the tool: Lawyers should know that chatbots can hallucinate and should not treat them as legal databases.
- Expect sanctions: Courts are increasingly willing to fine lawyers for AI-related errors, with Illinois signaling higher penalties.
For Aarons, the consequences are immediate: a financial penalty, a contempt finding, and a damaged professional reputation. For the broader legal profession, the case is a cautionary tale about the limits of automation. The Illinois appeals court's hope for a higher fine suggests that judges may not stop at $5,000. As AI becomes more embedded in legal work, the line between assistance and abdication will be tested in courtrooms nationwide.
The New Mexico Supreme Court's order does not mince words. It holds Aarons responsible for failing to verify his AI-generated brief. In doing so, it joins a growing body of precedent: lawyers may use technology, but they cannot outsource their ethical obligations. The murder appeal may have been the setting, but the real trial was about whether AI can be trusted in matters of life and liberty. The court's answer, at least for now, is a resounding no.
