Lawyer Uses ChatGPT to Generate Fake Witnesses and Testimony in Murder Appeal

ChatGPT Generates Fabricated Legal Brief

New Mexico attorney Stephen D. Aarons used ChatGPT to create a legal brief for a murder appeal. The AI tool produced false testimony from fabricated witnesses, including individuals named Danny and Linda Stanton, who did not exist. The brief also contained other AI-generated statements regarding the shooter's clothing and appearance.

Court Discovers AI-Generated Errors

Judges discovered the fabricated content in the brief, leading to a direct contempt order from the New Mexico Supreme Court. Aarons admitted to relying on ChatGPT, stating he "assumed" it would generate a "bulletproof summary of proceedings." He attributed the error to his own "stupidity" and an "honest mistake."

Impact on the Case and Client

Aarons was representing a client found guilty of killing his wife and currently serving a life sentence. The attorney had not informed his client about the use of ChatGPT or the court's discovery of the fabricated content. This incident underscores the potential for AI-generated errors to compromise legal proceedings.

Broader Implications for AI in Law

This case highlights ongoing concerns within the legal profession regarding the use of AI. Judges have noted the frequent occurrence of lawyers submitting AI-generated documents with errors. The incident serves as a cautionary tale about the need for verification when using AI tools in professional contexts, particularly in fields with significant consequences like law.