Fake Cases Made In AI Make Lawyers Have Stones

AI is very dangerous if used in court (photo: pexels)

JAKARTA - The US personal injury law firm, Morgan & Morgan, sent urgent emails this month to its more than 1,000 lawyers: Artificial intelligence (AI) can create fake legal cases, and use artificial information in court documents that can get you fired.

A federal judge in Wyoming recently threatened to impose sanctions on the two lawyers at the firm who included a fictitious case excerpt in a lawsuit against Walmart. One of the lawyers admitted in court documents last week that he used an AI program that was "alucinating" in creating the cases and apologizing for what he called an unintentional mistake.

AI's tendency to produce fiction in court documents has led courts across the country to question or sanction lawyers in at least seven cases over the past two years, creating "new headaches" with high-tech parties and judges.

The Walmart case stands out because it involves well-known law firms and large corporate defendants. However, similar examples have emerged in various types of lawsuits since chatbots such as ChatGPT started the AI era, highlighting the risks of new litigation.

Morgan & Morgan's spokesman did not respond to a request for comment. Walmart declined to comment. The judge has not yet decided whether to impose sanctions on lawyers in the Walmart case, which involves a allegedly flawed toy hostboard.

Advances in a generative AI help reduce the time it takes lawyers to research and compile legal documents, prompting many law firms to cooperate with AI vendors or build their own AI tools. Turny-three percent of lawyers surveyed by Thomson Reuters, the parent company Reuters, last year said they had used AI for work, and 12% said they were using it regularly.

However, generative AI is known to be confident in creating facts, and lawyers who use them must be careful, law experts say. AI sometimes generates false information, known as "halusination" in the industry, as this model generates responses based on the statistical patterns learned from large datasets, not by verifying the facts in the dataset.

Lawyers' ethical rules require lawyers to examine and be responsible for their court documents or risk sanctions. The American Bar Association last year notified its 400,000 members that the obligation extends to even "unintentional false statements" generated through AI.

"The consequences have not changed just because legal research tools have developed," said Andrew Perlman, the dean of Suffolk University's legal school and supporters of AI's use to improve legal work.

"When a lawyer is caught using a ChatGPT or any generative AI tool to make a quote without examining it, it's an inability, that's as simple as it is," Perlman said.

In one of the earliest court warnings regarding the use of AI by lawyers, a federal judge in Manhattan in June 2023 fined two New York lawyers $ 5,000 for citing a case created by AI in a personal injury case against airlines.

Another New York federal judge last year considered imposing sanctions in a case involving Michael Cohen, former lawyer and "fixer" Donald Trump, who said he had accidentally given a false case quote to his lawyer, who was later filed in Cohen's criminal campaign tax and finance case.

Cohen, who used Google's AI Bard chatbot, and his lawyer was not sanctioned, but the judge called the incident "embarrassing."

In November 2024, a Texas federal judge ordered a lawyer citing cases and excerpts that were not included in the unlawful termination of employment lawsuit to pay a fine of 2,000 US dollars (Rp32 million) and take courses on in-law-generative AI.

A federal judge in Minnesota last month said that a misinformation expert had destroyed his credibility in court after he admitted accidentally citing false citations produced by AI in a case involving Vice President Kamala Harris' "deepfake" parody.

Harry Surden, professor of law at a Colorado University law school studying AI and law, said he recommended lawyers take time to study "the strengths and weaknesses of those tools." He said these increasing examples show "lack of AI literacy" in the profession, but the technology itself is not the problem.

"Lawyers always make mistakes in their documents before there is an AI," he said. "This is nothing new."