Dutch lawyers receive warnings for using ChatGPT to cite nonexistent cases in court
Three lawyers in the Netherlands were warned for misusing artificial intelligence in court, with two ordered to take an AI training course, NOS reported.
The lawyers relied on AI programs, including ChatGPT, to support their arguments, citing previous court rulings that either did not exist or were unrelated to the cases.
Wouter Timmermans, the legal regulator in Gelderland, told NOS that such unverified AI-generated information poses a risk in the courtroom. “The most serious problem I’ve seen is lawyers blindly submitting AI-generated content in court without checking it,” Timmermans said. “AI often produces text that reads logically, even if it is nonsense. You cannot trust it 100 percent.”
Judges in Arnhem, Rotterdam, and Groningen recently flagged suspected AI misuse after lawyers referenced nonexistent or irrelevant cases. Timmermans noted that while more than three investigations into AI misuse have been conducted, not all lawyers were required to attend warning meetings. Several inquiries are still ongoing.
Although warning meetings are the mildest disciplinary action, AI is now a key focus for regulators, who anticipate increased use of AI in legal practice.
Lodewijk Smeehuijzen, a law professor at the Vrije Universiteit Amsterdam, said lawyers should be held accountable for relying on AI without verification. “You cannot cite sources that don’t exist. These AI programs do not always produce correct information. Failing to check is serious and harms your reputation,” he told NOS.
AI is also entering courtrooms without lawyers. In December, a man in Den Bosch attempted to resolve a dispute over an apartment sale in Andorra using ChatGPT instead of legal counsel. Although a previous ruling confirmed a valid contract, the seller refused to transfer the property. The man, guided by AI, submitted legal arguments citing irrelevant provisions.
