Lawyer fined for using AI to fabricate evidence in legal documents

A New Mexico lawyer, Stephen Arons, was fined $5,000 after submitting fabricated police testimony and fictional witnesses in a murder appeal document, which was generated by OpenAI’s ChatGPT. The New Mexico Supreme Court ruled that Arons violated court rules by failing to verify the accuracy of the submitted documents. Arons admitted to using AI to draft the legal filing but claimed it was an unintentional mistake. The court also referred him to the state bar for further investigation. Judge C. Shannon Bacon criticized Arons for not being aware of the risks of relying on AI-generated content, stating that the issue of AI hallucinations is a daily headline. This case highlights the growing concern over lawyers using AI without proper verification, as multiple legal professionals have faced penalties for similar errors. Arons explained that he had used ChatGPT to summarize court transcripts but did not realize the AI would fabricate significant details. The court found that the false testimony was from entirely made-up witnesses, which is more severe than previous cases involving misquoted legal references. The case has sparked discussions about the ethical responsibilities of legal professionals when using AI tools. Arons expressed regret but asked the bar to consider it an accident, urging other professionals to learn from the incident.