A New Mexico court fined attorney Stephen Aarons $5,000 (approximately 1.6 million forints) after an appeal he filed in a murder case was found to contain false material generated by artificial intelligence. Reuters reported that the filing included invented witnesses and police statements that the court identified as fabricated.
What happened
The appeal contained multiple serious inaccuracies: it included purported witness “testimonies” from people who do not exist and made false claims about the shooter's clothing and appearance. Judge C. Shannon Bacon questioned Aarons at an August hearing about how he could have been unaware of the risks posed by AI and why he did not verify the assertions in the submission.
Aarons acknowledged using ChatGPT to prepare the filing, explaining he expected the tool to produce an "unassailable summary" of the trial. The court criticized him for failing to check the AI-generated factual statements and legal citations.
Why this matters
The case highlights that chatbots and generative AI systems can "hallucinate"—produce plausible but false information. This poses a particular danger in legal contexts, where unverified AI-generated content can undermine proceedings and lead to misinformation in official documents.
Consequences and next steps
Beyond the $5,000 fine, the court reprimanded Aarons for not exercising adequate verification. Aarons expressed regret and said he hopes any disciplinary review will consider the mistake a "good-faith error." The Reuters report provided the basis for the court's action and the judge's comments.
The incident serves as a caution to legal professionals: when using AI tools, rigorous fact-checking of both factual claims and cited authorities is essential, especially in cases that affect individuals' liberty.
