Second Circuit Cautions Lawyer AI Brief: Errors Lead to Refiling

• Attorney Elliot Shields admitted to the Second Circuit that AI likely caused citation and quotation errors in his brief for an $82 million personal injury case.
• U.S. Circuit Judge Maria Araújo Kahn warned of AI's "perils," citing a Connecticut lawyer's suspension for similar issues, emphasizing the need for careful review.
• The court ordered Shields to file a corrected brief within one week, highlighting judicial intolerance for unverified AI-generated content.
• This incident follows several other cases where lawyers faced sanctions or chastisement for AI-created false citations, hallucinations, or non-existent case references.
• The underlying case involves an $82 million award to a client who lost limbs in a subway accident, with the MTA appealing and arguing resource limitations for safety measures.

The court's admonition underscores a growing judicial intolerance for unverified AI-generated content, signaling that attorneys who fail to meticulously review such output risk severe professional repercussions.

A recent hearing before the Second Circuit saw attorney Elliot Shields of Roth & Roth LLP face scrutiny after admitting that artificial intelligence likely contributed to significant errors in a legal brief. Shields, representing client Luisa Janssen Harger Da Silva in a high-stakes $82 million personal injury case, acknowledged that his submission contained multiple quotation and citation inaccuracies. These errors, which he described in letters to the court as "professionally embarrassing and humiliating," were attributed to the use of AI tools.

U.S. Circuit Judge Maria Araújo Kahn directly questioned Shields about the origin of the mistakes, prompting his admission that they likely "arose from the use of AI." Judge Kahn, a Joe Biden appointee, emphasized the potential pitfalls of such technology, referencing a recent case where a Connecticut lawyer faced suspension for similar AI-related issues. The court has since directed Shields to file a corrected brief within one week, underscoring the immediate need for accuracy in legal filings.

The incident involving Elliot Shields AI citation errors highlights a broader trend of judicial concern regarding the uncritical adoption of AI in legal work. Judge Kahn's warning that AI, while a valuable resource, carries significant "perils," resonates with increasing reports of legal AI hallucinations sanctions across the country. U.S. District Judge Rachel Kovner, who joined the appellate panel, further noted that Shields' brief contained "any number of quotes that are not in the cases they’re attributed to," indicating a fundamental breakdown in verification.

This is not an isolated occurrence. Earlier this summer, a Washington attorney faced sanctions after using AI tools like ChatGPT and Claude to prepare filings that included dozens of false citations. Last year, a Georgia lawyer was chastised for citing AI-created "hallucinations" in a high-profile assault case. Similarly, in 2023, New York lawyer Steven Schwartz found himself in a difficult position after his ChatGPT-crafted brief cited non-existent cases, leading him to lament, "It just never occurred to me that it would be making up cases." Shields, however, clarified that his errors stemmed from accidentally transposing notes and presenting paraphrases as direct quotes, rather than fabricating cases, attributing it to being "rushing and sloppy."

At the heart of the proceedings is the appeal of an $82 million jury award to Luisa Janssen Harger Da Silva, a former student tourist from Brazil. In 2016, Da Silva suffered devastating injuries, losing her left arm and left leg, after fainting onto a Brooklyn subway track and being struck by a train. Her legal team contends that the Metropolitan Transportation Authority (MTA) is liable because it failed to implement crucial safety measures, such as sliding screen doors, despite possessing over a decade of data indicating the dangers posed by exposed tracks. They argue that such safety installations are common in other major transit systems worldwide.

MTA attorney Max Crema countered these arguments, characterizing the absence of platform screen doors in New York as a "trade-off" necessitated by the agency's "limited public resources." Crema stated that the MTA had extensively studied the issue, including convening a special task force, and ultimately opted for an "incremental course, further consideration and pilot project planning." He asserted that this was a deliberate, study-based decision, not a negligent oversight, thereby entitling the MTA to qualified immunity. Judge Kovner challenged this stance, questioning whether such immunity would still apply if the underlying studies were "plainly inadequate," to which Crema responded that the immunity is intended to be "broad immunity for planners."

The Second Circuit cautions lawyer AI brief use, underscoring a critical juncture for lawyer professional ethics AI. The court's admonition underscores a growing judicial intolerance for unverified AI-generated content, signaling that attorneys who fail to meticulously review such output risk severe professional repercussions. The incident involving Elliot Shields and Roth & Roth LLP serves as a stark reminder that while AI tools offer efficiency, they do not absolve legal professionals of their fundamental duty to ensure the accuracy and integrity of all court submissions.

Legal practitioners and compliance officers must implement stringent verification protocols for all AI-generated legal content, particularly citations and quotations. The increasing instances of sanctions and public chastisement demonstrate that courts are actively scrutinizing and penalizing unverified AI use, making diligent human oversight an indispensable component of modern legal practice to avoid ethical breaches and potential suspension.

Lawyers and compliance officers must implement stringent verification protocols for all AI-generated legal content, especially citations and quotations, to avoid professional sanctions, ethical breaches, and potential suspension, as courts are actively scrutinizing and penalizing unverified AI use.

Source: Original reporting via Courthouse News Service