A lawyer from New Mexico received disciplinary action for using ChatGPT

A New Mexico state court has issued a disciplinary ruling against a lawyer who used ChatGPT to prepare court documents, including a motion that cited fake testimony from made-up witnesses, which distorted the evidence in a murder case. According to official documents, the lawyer failed to verify the accuracy of the information provided, relying on the model's responses, and it was discovered during the proceedings that the references to witness testimony presented in court were unfounded, as the individuals had never made such comments in the case of Oscar Sandoval. The incident was part of a broader process in a murder case in Mesilla, where the defendant was sentenced to life plus 4.5 years, and the use of unreliable data threatened the procedural integrity of the case and led to sanctions against the defense. The lawyer acknowledged that he was not aware of the ability of language models to "hallucinate" - generate facts that appear convincing but are not true, and the case highlights the risks of automating legal work without proper human oversight. The judicial system continues to face challenges related to the implementation of generative technologies, and the decision in this case sets a precedent for accountability for verifying AI-generated content in the context of professional legal activity.