The usage of AI by law firms has been increasing each year. It was estimated that 30% of law firms in the US use AI tools, which is higher than 11% in 2023. Generally, these law firms use AI tools to deal with menial tasks like document processing and legal research. This resolves the difficulty of devising a proper legal strategy in court because they were bombarded with menial tasks.
However, what happens when a lawyer goes beyond utilizing AI tools for menial, repetitive tasks? What happens when AI becomes an integral part in a lawyer’s brief? Introducing Mata v. Avianca, a case of AI hallucination and a lawyer who believed in it.
It Started as a Dismissed Case
When plaintiff Roberto Mata sued Avianca after an employee hit him with a serving cart, it was initially dismissed by the Court. The dismissal was caused by the expiry of the statute of limitations. However, Mata stood firm on his case. His lawyers submitted a brief, stating that the lawsuit should continue and referencing other cases.
However, upon reviewing the case, Avianca’s lawyers cannot find the cases cited in the brief. Even the cited court opinion was not found. The judge then ordered Mata’s attorneys to provide copies of the referred cases and opinion in the brief, and they did. They even listed the court and judges who issued them, as well as the docket and case numbers. Still, Avianca’s lawyers were unable to locate the cases. At that point, they had an inkling that AI might have been involved. In the court file, they claim that, as lawyers who specialize in Aviation Law, they can tell if a case is not real.
A “Bogus” Opinion
Judge Castel then conducted his own inquiry about the brief and called for a hearing. During the hearing, it was revealed that ChatGPT was used in the legal brief and the cases were fabricated.
When Schwartz, Mata’s lawyer, was testifying, he stated, “My reaction was, ChatGPT is finding that case somewhere. Maybe it’s unpublished. Maybe it was appealed. Maybe access is difficult to get. I just never thought it could be made up.” Schwartz added that although he didn’t find the case as well, he still never thought that the case could be fabricated.
As a result, the case was dismissed, and the lawyers using ChatGPT were fined. In his decision, the judge stated:
“Many harms flow from the submission of fake opinions. [xxx] There is potential harm to the reputation of judges and courts whose names are falsely invoked as authors of the bogus opinions and to the reputation of a party attributed with fictional conduct. It promotes cynicism about the legal profession and the American judicial system. And a future litigant may be tempted to defy a judicial ruling by disingenuously claiming doubt about its authenticity.”
What is AI Hallucination
What happened in this case is a phenomenon called “AI hallucination”. It is when an AI hallucinates or adds misleading information. Although AI is admittedly becoming more powerful, it will always hallucinate, according to a New York Times article. This is because AI can only process language, but cannot discern what’s true and what’s not. Unlike humans, AI cannot fully grasp context, intent, or credibility. Despite the convenience it markets itself to have, AI still require human oversight to make sure it works properly. This impracticality and risk of misinformation defeats the purpose of AI in law firms, where accurate information is the standard.
Who Should Lawyers Trust
At the end of the day, humans do the fact-checking for AI. While admittedly, AI can be more efficient in processing documents and other menial tasks, there is no due diligence behind it. AI does not fact-check, nor can we hold AI accountable for false information.
Which is why law firms, where the truth is held to the highest standard, should be handled by humans. Whether it’s menial tasks like answering emails or processing exhibits, only humans can do tasks with prudence and due diligence. Law firms need legal professionals who are trained by lawyers for lawyers, not an AI chatbot that could put their legal careers at risk.