New Mexico Lawyer Sanctioned for Citing Fake ChatGPT-Generated Testimony
A defense attorney in New Mexico faced court sanctions after submitting fictitious witness testimony generated by ChatGPT in a legal filing. The lawyer admitted in court that he did not know artificial intelligence could hallucinate fabricated facts. This incident highlights the ongoing risks of professionals relying on generative AI without verifying outputs in high-stakes domain like the legal system. It underscores the critical need for legal AI literacy and strict verification protocols as AI tools become widespread in professional workflows. The lawyer cited non-existent testimony from made-up witnesses provided by ChatGPT, leading directly to judicial disciplinary measures. This case adds to a growing series of court rulings establishing that legal professionals are personally responsible for verifying any AI-assisted work submitted to the judiciary.
## BACKGROUND
Large Language Models (LLMs) like ChatGPT are prone to a behavior known as "hallucination," where they generate plausible-sounding but entirely fabricated facts, quotes, or references. Courts across the United States have increasingly penalized attorneys who present AI hallucinations as legal evidence, establishing that professional standards require manual checking of all AI-generated content.