Judge Stephen Alexander Vaden of the Court of International Trade (“CIT”) issued an order on June 8, 2023 to address increasing concerns related to the use of generative artificial intelligence platforms in drafting documents that are publicly filed in the litigation process.  The order reflects a growing awareness of the novel risks associated with using artificial intelligence to draft court documents, including, as Judge Vaden notes, with respect to the inadvertent disclosure of business proprietary information.

Judge Vaden’s reasoning is that should an attorney utilize ChatGPT (or a similar platform) to assist in drafting documents submitted to the court, the search terms used could contain business proprietary terms that were otherwise not released to the public as part of the litigation.  These platforms retain information in the search terms indefinitely as part of its “learning” process, and therefore any potentially business proprietary information could become public (whether intended or not).  Consequently, the information that was originally treated as business proprietary in litigation, and during the underlying administrative proceeding, would no longer be protected.

As a result, the order instructs practitioners in cases assigned to Judge Vaden who utilize artificial intelligence resources to assist in preparing briefs, or any other documents submitted to the Court, to provide the following:

• A disclosure notice that identifies the program used and the specific portions of text drafted with the assistance of that program; and
• A certification that the use of the program has not resulted in the disclosure of any confidential or business proprietary information to any unauthorized party.

The order also states that following such a disclosure, any party may file any motion provided for under the CIT’s rules to seek relief that it believes the facts disclosed warrant.

Judge Vaden’s order follows a recent order by U.S. District Judge Brantley Starr of the Northern District of Texas, which requires litigants before the Court to certify either than no portion of any filing will be drafted by generative artificial intelligence, or that any language so drafted be checked for accuracy “by a human being.”  

[View source.]