Federal judge in New York warns: The use of artificial intelligence could harm clients and lawyers’ training

A federal judge in Manhattan this week issued a strong warning about the growing use of artificial intelligence by law firms, stressing that reliance on AI tools could create risks for clients and harm the training of young lawyers — while deciding not to impose sanctions on lawyers for errors caused by AI in a document filed in court, as Reuters reports.

District Judge Arun Subramanian called the episode a “wake-up call” for the legal profession, warning that legal work generated by AI must be carefully checked for accuracy. “Even though clients may ask for AI to be used to save money, there is a risk of negative consequences when things go wrong. For this reason, it might be wise for clients to think carefully before asking lawyers to use AI for tasks such as drafting documents,” he said, according to Reuters.

The case stems from a copyright infringement lawsuit, in which last month the judge had questioned whether the “unchecked” use of AI had led to citation errors in the filed documents. Lawyer Cynthia Arato, who represents the defendant Foundation Media, apologised and explained that the errors had come from another lawyer at her firm, who had prepared the initial draft and finalised it; according to her, he had acted “contrary to the firm’s policies and training”, while the failure to verify the AI-generated citations was “an isolated mistake and not any malice or intent to deceive the court”, Reuters reports.

Her firm, Shapiro Arato Bach, is implementing additional safeguards against the misuse of AI. In a filing on Thursday, Arato revealed that the firm had used artificial intelligence tools created by the technology company Harvey to help prepare the insufficiently verified document, as Reuters reports. Harvey and Foundation Media did not immediately respond to requests for comment.

Judge Subramanian stressed that the legal profession must “think about ways to ensure that young lawyers receive the necessary training, even as reliance on these technological tools grows”. He suggested that firms could require that initial drafts of legal documents be prepared without the help of artificial intelligence. “The use of AI may be inevitable, but mentoring young lawyers is a shared responsibility that we all share,” the judge said.

This case is only one of many others in which state and federal judges are assessing how to sanction lawyers for the misuse of artificial intelligence, Reuters notes: recent similar cases include a document called “AI slop” filed by a lawyer in Florida, a State Farm lawyer fined in Los Angeles, and a judge in Oklahoma whose AI-assisted decision contained false citations.