A US federal judge has dismissed a self-represented litigant’s 67-page employment complaint and prohibited her from using AI to prepare its replacement.
Judge Krissa Lanham of the District of Arizona said the pleading appeared AI-generated because it resembled other such complaints encountered by the court. It contained repetitive factual allegations, unclear and conflicting claims, and facts insufficiently connected to particular causes of action. The plaintiff later told Reuters that AI had been used in part. The court permitted an amended complaint but required the plaintiff to draft it personally and limited it to 25 pages.
The order does not establish that using AI caused the defects, nor does it prohibit AI throughout the proceeding. For later filings, the judge instead requires disclosure of any generative-AI use, identification of the affected passages, and certification that the filer independently reviewed and verified the material.
The development matters because the court’s concern was not fabricated authority. It was the failure to convert a mass of generated content into a coherent legal case. That broadens the emerging judicial response from checking citations to examining whether the filer has genuinely selected the facts, identified the claims and exercised independent judgment.
For Australian lawyers, the practical lesson is that verification cannot be confined to factual accuracy. AI-assisted work must also be relevant, proportionate and intellectually adopted by the person responsible for filing it.
Sources: court order; Reuters.
Leave a Reply