Counsel Took Full Responsibility for His Mistakes, Was not Evasive, and Prioritized Candor to the Court and Opposing Counsel

From Judge Jennifer Choe-Groves (Ct. of Int’l Trade) Thursday in Disruptive Resources, LLC v. Ballistic Barrier Prods. Inc. (D. Del.):

Before the Court is an action involving claims of patent infringement between Plaintiff Disruptive Resources … and Defendants Ballistic Barrier Products Inc. and Mayday Security Solutions LLC …. After Plaintiff’s counsel disclosed to the Court that the previously filed Joint Claim Construction Brief … featured multiple errors due to his use of generative artificial intelligence (“AI”), the Court ordered counsel to show cause in writing as to why the Court should not impose sanctions for counsel’s conduct pursuant to Federal Rule of Civil Procedure 11 ….

In counsel’s letter to the Court, counsel described the events that led to his disclosure of inaccuracies attributable to generative AI. Portions of counsel’s reply brief in support of Plaintiff’s claim constructions “included an erroneously proposed construction that deviated from the construction of the terms Plaintiff had advanced through the claim construction process[.]”

Counsel stated that he used the AI system “Strongsuit” to generate an initial outline of Plaintiff’s reply brief and to pull statements from a deposition that supported Plaintiff’s position on claim construction. Counsel used directed prompts to complete an initial draft of the reply brief that integrated extrinsic evidence mentioned in the deposition.

When editing the draft, counsel “jumped from section to section” and mistakenly “believed [he] had edited the entire brief, when in reality, [he] had missed the 1-page section” that proposed an incorrect construction. Counsel admits that he failed to conduct a final review of the entire brief and “overlooked some of the arguments being made.”

Upon discovering the erroneous statements in the brief, counsel notified opposing counsel and proposed amending the Joint Brief to implement footnotes to address the erroneous proposed constructions. The Parties later filed a [Joint Letter] disclosing these issues to the Court. Counsel argues that his actions did not violate Rule 11 or any rules of professional conduct because the reply brief was not presented for an improper purpose, the claims and legal contentions were supported by authentic existing law, and all factual contentions were supported by evidenceCounsel suggests that his errors were “more akin to taking litigation positions that are ultimately borne out to not be adopted by the Court.”

The Court observes that the Joint Letter identified that two sections of the Joint Brief contained a concluding statement that proposed an incorrect claim construction, and a third section was withdrawn entirely because it “contained more pervasive errors.” Plaintiff sought to strike the problematic section of its reply brief, causing Defendants to also strike their subsequent briefing that responded to the inaccurate AI-generated arguments put forth by Plaintiff. This demonstrates that counsel’s erroneous constructions not only impacted the accuracy and strength of Plaintiff’s legal arguments, but caused opposing counsel to spend time responding to erroneous arguments that have now been struck.

The incorporation of the use of generative AI in the legal field is growing exponentially, and this is exemplified by counsel’s admitted use of large language models in his legal practice to ease the workload on his smaller law firm. Nevertheless, “when using AI, litigants must still strictly adhere to all rules of professional conduct.” In both his letter to the Court and statements made during the hearing, counsel expressed regret over the unfortunate outcome of his use of AI without thorough review and recognized the importance of maintaining a higher standard of diligent lawyering than what led to the erroneous briefing.

The circumstances of this case are different compared to other instances when the use of AI led to legal filings with hallucinated case citations or fictitious judges. See Mata v. Avianca, Inc. (S.D.N.Y. 2023) (an attorney “submitted non-existent judicial opinions with fake quotes and citations created by the [AI] tool ChatGPT, then continued to stand by the fake opinions after judicial orders called their existence into question.”). In this case, counsel took full responsibility for his mistakes, was not evasive, and prioritized candor to the Court and opposing counsel by disclosing the errors quickly. Contra Mata (finding that the attorney “knew before making a submission to the Court that the full text of [a fake case] could not be found but kept silent.”).

The Court expects that these events and this Opinion shall serve as a cautionary warning and learning experience for counsel moving forward. The Court warns counsel that any future incidents involving AI mistakes in this case may result in sanctions.

Upon consideration of the circumstances as described by counsel, and his appreciation for the gravity of filing erroneous legal arguments drafted by generative AI without proper review, the Court declines to sanction counsel for his conduct….