2nd Circuit: Lawyer Admits AI Brief Errors, Must Revise

• Attorney Elliot Shields admitted to the 2nd U.S. Circuit Court of Appeals that he used artificial intelligence to draft his client's brief.
• The AI use led to errors, specifically citing paraphrased content as direct quotes, which Shields attributed to rushing and transposing notes.
• Shields represents Luisa Janssen Harger Da Silva, who was awarded approximately $82 million by a district court after losing limbs in a subway accident.
• The Metropolitan Transportation Authority (MTA) appealed the $82 million judgment, leading to the current proceedings.
• The judicial panel granted Shields one week to file a corrected brief, emphasizing the need for accuracy in legal submissions.

This episode involving Elliot Shields' AI brief errors serves as a stark reminder of the critical professional responsibility lawyers bear when incorporating artificial intelligence into their legal writing.

During recent oral arguments before the 2nd U.S. Circuit Court of Appeals in New York, attorney Elliot Shields, representing a client in a significant appeal, acknowledged using artificial intelligence to draft portions of his brief. Shields, who practices with Roth & Roth, admitted that this reliance led to errors in the filing, specifically the misattribution of paraphrased content as direct quotes. This admission came during a hearing on Friday, September 4, 2026, where Shields addressed a panel of judges.

Prior to the hearing, Shields had already communicated with the 2nd Circuit, sending letters to acknowledge the inaccuracies. At the oral arguments, he clarified that the issue was not the fabrication of legal cases, but rather the incorrect presentation of summarized information as verbatim quotations. He candidly told the court, "We were rushing, and we were sloppy," further explaining, "I think that I transposed notes and put them in quotations."

The judicial panel, comprising 2nd Circuit Judge Maria Araújo Kahn, Senior 2nd Circuit Judge Reena Raggi, and U.S. District Judge Rachel Kovner from the Eastern District of New York, responded by granting Shields a one-week period to submit a revised brief. This incident highlights the growing scrutiny surrounding the use of AI in legal drafting, particularly concerning the accuracy of citations and direct textual representations.

The brief in question pertains to an appeal filed by the Metropolitan Transportation Authority (MTA) against a substantial judgment. Shields represents Luisa Janssen Harger Da Silva, a student from Brazil whose life was irrevocably altered after a tragic incident in Brooklyn, New York City. Da Silva fainted and fell onto a subway track, subsequently being struck by a train, which resulted in the loss of an arm and a leg.

In 2025, the U.S. District Court for the Eastern District of New York had ruled in Da Silva's favor, awarding her approximately $82 million. The district court determined that the MTA bore responsibility, finding that the authority was aware of the potential for individuals to fall onto the tracks but failed to implement adequate safety measures. The MTA subsequently initiated the appeal to the 2nd U.S. Circuit Court of Appeals, setting the stage for the recent oral arguments where the AI-generated brief errors came to light.

This episode involving Elliot Shields' AI brief errors serves as a stark reminder of the critical professional responsibility lawyers bear when incorporating artificial intelligence into their legal writing. The misrepresentation of paraphrasing as direct quotes, even if unintentional and attributed to haste, underscores the potential for ethical breaches and damage to credibility with the court. The incident highlights the imperative for attorneys, including those at firms like Roth & Roth, to meticulously verify all AI-generated content, especially when it involves citations and direct textual representations.

The case illustrates a significant challenge in legal ethics: maintaining rigorous human oversight when leveraging AI in legal writing. Over-reliance on artificial intelligence without thorough human review can lead to mistakes that compromise the integrity of legal filings and the trust placed in legal professionals. This cautionary tale emphasizes that while AI tools can assist in drafting, the ultimate responsibility for accuracy and adherence to court rules remains squarely with the attorney of record. The 2nd Circuit's directive for a corrected brief reinforces the expectation that all submissions meet the highest standards of precision and honesty.

This incident underscores the critical professional responsibility for lawyers to meticulously verify all AI-generated content, particularly citations and direct quotes, to prevent ethical breaches and maintain credibility with the court. It serves as a cautionary tale regarding the risks of over-reliance on AI without thorough human oversight in legal drafting.

Source: Original reporting via ABA Journal and news services.