Skip to Main Content

ChatGPT and Generative AI Legal Research Guide

Judges' Policies About Use of Generative AI

AI Court Order Tracker

AI Court Order Tracker by Ropes & Gray. This tool is designed to assist clients in understanding the varying approaches courts take regarding the use of AI in court filings. The tracker categorizes court orders into several types: some aim to educate on AI use, others outright prohibit it, and most require disclosure or verification of AI-generated content's accuracy. Shannon Capone Kirk, the firm's managing principal and global head of advanced e-discovery and AI strategy, highlighted the growing trend of such orders from individual judges and courts. He emphasized the legal profession's need to stay informed as courts address issues of transparency and accuracy in AI use. Amy Jane Longo, a litigation & enforcement partner, cautioned about the potential adverse outcomes of over-relying on generative AI. She advised attorneys to be aware of jurisdiction-specific requirements for notifying courts about AI use and to prepare for future regulations and professional conduct rules related to AI.

U.S. Court of International Trade

Another Judge Issues AI Order, Citing Confidentiality Risks, Sarah Martinson, Law360, June 30, 2023 

U.S. Court of International Trade Judge Stephen A. Vaden has issued an order requiring attorneys to reveal the generative AI tools used in the drafting of their briefs. They must also specify where in the briefs these tools were used and confirm that the use of such tools hasn't led to the unauthorized exposure of confidential or proprietary business data. Judge Vaden highlighted the challenge generative AI tools pose in protecting sensitive information, despite the court's specific rules designed to safeguard such data. 

U.S. District Court for the District of Montana

Federal Judge Forbids Use of ChatGPT by Out-of-State Lawyers, Eugene Volokh, The Volokh Conspiracy, May 29, 2023 

U.S. District Court for the District of Montana Judge has added a provision to the standard pro hac vice admission order, stating that the use of AI automated drafting programs, including Chat GPT, is prohibited. This restriction applies only to lawyers seeking pro hac vice admission, indicating the judge's disapproval of AI programs specifically in this context. 

U.S. District Court for the District of Texas

No ChatGPT in my Court: Judge Orders all AI-generated Content Must be Declared and Checked, Devin Coldewey, Techcruch, May 30, 2023 

U.S. District Court for the District of Texas Judge Brantley Starr, has introduced a requirement for attorneys in his court to confirm that none of their legal arguments were drafted by generative artificial intelligence, or if they were, they were reviewed by a human. Judge Starr's mandate, called the "Mandatory Certification Regarding Generative Artificial Intelligence," is enforceable in his courtroom and requires attorneys to certify the sources of their legal arguments, whether drafted by AI or not. This development underlines the growing concerns and challenges posed by AI in the legal profession. This decision follows an incident where attorney Steven Schwartz utilized an AI, ChatGPT, to assist with his legal research for a federal filing, only to find that all six cases and relevant precedent provided by the AI were entirely fabricated.  

5th Circuit Court of Appeals

In First for A U.S. Appeals Court, 5th U.S. Circuit Court Considers Rule Requiring Lawyers to Certify they Did Not Rely on AI to Create Filings, Bob Ambrogi, LawSites, November 29, 2023 

The 5th U.S. Circuit Court of Appeals is considering a historic rule change requiring lawyers and litigants to certify whether they used AI in drafting court filings. This appears to be the first federal appeals court to propose such a rule. Many courts began exploring AI rules after lawyers were sanctioned for filing a ChatGPT-generated brief filled with bogus cases. The 5th Circuit published the proposal and is seeking public comments until January 4th, 2024. False AI certifications could lead to rejection of filings and sanctions.