Since Steven A. Shwartz, an attorney at Levidow, Levidow & Oberman, infamously filed a legal brief that was drafted by ChatGPT, which fabricated cases, two federal judges have issued certifications requiring explicit disclosures for attorneys using generative artificial intelligence tools in legal work before their court.

On Tuesday, U.S. District Judge Brently Starr in the Northern District of Texas issued a standing order requiring attorneys to file mandatory certifications attesting that they did not use generative AI, such as OpenAI’s ChatGPT or Google’s Bard, in drafting their court filings. Or, if they did use it, attesting that its output ”was checked for accuracy, using print reporters or traditional legal databases, by a human being.”