Lawyer fined for AI-generated false witnesses in murder appeal

lawyer fined for AI-generated false witnesses in murder appeal

New Mexico’s Supreme Court sanctioned an attorney after he filed a brief that contained made-up witnesses and false police testimony churned out by an AI tool. The court’s Wednesday filing says Stephen Aarons got a $5,000 fine and was held in contempt for skipping the step of checking the facts and legal citations in his AI-generated brief.

According to that filing, the brief quoted witnesses who don’t exist and gave wrong descriptions of the shooter’s clothes and looks. Those fabricated bits were trotted out to back an appeal trying to vacate a murder conviction. The court ruled that leaning on the AI output without any independent check was a serious breach of professional duty.

august hearing justice c comments on ai-generated false testimony

At an August hearing, Justice C. Shannon Bacon pressed Aarons on whether he grasped the risks of using generative AI in law work. Aarons confessed he’d turned to ChatGPT hoping for a “bulletproof summary” of the trial record. The justice shot back asking if he follows news, listens to the radio, or reads anything about the rising worries over AI hallucinations, noting that stories of lawyers leaning on made-up AI content have become everyday headlines.

In a statement to Reuters, Aarons expressed regret, saying he hopes the disciplinary board sees the slip as unintentional. He stressed he never meant to mislead the court but admitted his reliance on the tech wasn’t cautious enough.

The episode points to a wider pattern in the legal field. As more lawyers bring AI into research and drafting, courts are seeing more filings peppered with false citations, invented quotes, or misleading info spun by these systems. Last year a judge slammed two law firms for a brief riddled with numerous inaccurate and misleading legal references. In another case, lawyers for MyPillow founder Mike Lindell got hit with penalties for slipping AI-crafted misquotes and fabricated citations into a court document.

warning sign for professionals using ai in legal settings

Those developments flash a warning sign for anyone building or using AI in professional settings. The tech can speed up tasks like summarizing huge chunks of text or drafting first-draft arguments, but it doesn’t promise factual accuracy. Users ought to treat AI output as a rough draft that needs thorough checking against trustworthy sources.

For developers, the incident highlights the need to design systems that plainly signal their limits. Features that flag uncertain outputs, give confidence scores, or prompt a user review can help keep hallucinated information from being passed off as fact.

Legal practitioners should adopt strict verification routines when they rely on AI help. That means cross-checking any generated statement against primary case law, statutes, or credible secondary sources before it lands in a filing. Training programs that teach lawyers about how generative models work and where they tend to stumble are also wise.

Ultimately, the New Mexico ruling reminds the AI-enabled workforce that innovation must be paired with diligence. Tools like ChatGPT can boost productivity, but they can’t erase the professional duty to make sure every assertion filed with a court is true and backed up. By marrying AI’s strengths with rigorous human oversight, users can reap the technology’s benefits without sacrificing integrity or inviting sanctions.

Frequently asked questions

Why was lawyer Stephen Aarons fined by New Mexico Supreme Court?

Aarons was fined $5,000 and held in contempt for filing a brief that contained made-up witnesses and false police testimony generated by an AI tool without checking the facts or legal citations, which the court called a serious breach of professional duty.

What false information did the AI-generated brief contain according to the court filing?

The brief quoted witnesses who do not exist and gave incorrect descriptions of the shooter’s clothing and appearance, which were used to support an appeal seeking to vacate a murder conviction.

What did Justice C. Shannon Bacon ask Aarons during the August hearing about his use of generative AI?

Justice Bacon pressed Aarons on whether he understood the risks of using generative AI in legal work, asking if he follows news, listens to the radio, or reads anything about the growing concerns over AI hallucinations.

What broader pattern in the legal field does the Aarons case illustrate, and what advice does the article give lawyers using AI?

The case shows a rising trend of lawyers submitting filings with false citations, invented quotes, or misleading AI-generated content; lawyers should verify every AI-produced statement against primary law, statutes, or credible sources before filing.

Source: The Verge