Douglas v. Deutsche Bank Nat’l Tr. Co.

[*21] In Douglas v. Deutsche Bank National Trust Company, the District of Columbia Court of Appeals addressed the consequences of submitting fabricated legal citations obtained through artificial intelligence (AI). The case arose from a foreclosure dispute between Barry Douglas and Deutsche Bank National Trust Company. After the lower court ruled in favor of Deutsche Bank, Douglas appealed the decision.¹ During the appeal, the court discovered that Deutsche Bank's brief contained citations to several cases that could not be verified.²

The court ordered Deutsche Bank to explain why its brief should not be stricken. Attorney Loishirl Hall acknowledged that four of the cited cases did not exist and admitted that she had used Google's generative AI search tool to locate legal authorities without independently verifying them.³ Although Hall apologized and accepted responsibility, the law firm representing Deutsche Bank stated that it had been unaware of her actions. The court questioned the firm's lack of oversight, [*22] particularly because multiple attorneys were listed on the brief.⁴

The D.C. Court of Appeals struck Deutsche Bank’s brief. Rule 28 of the D.C. Court of Appeals Rules requires arguments supported by real authority.⁵ Attorneys must verify what they file, regardless of whether or not AI was involved.⁶ Fake citations waste the court’s time and leave the client without a credible advocate.⁷ Deutsche Bank’s law firm proposed deleting the bad citations and leaving the rest of the brief in place; however, the court refused, reasoning that this would be effectively no sanction at all and referred the matter to the Office of Disciplinary Counsel.⁸

Senior Judge Glickman wrote a concurring opinion agreeing with the court's decision to strike Deutsche Bank's brief but raising concerns about attorney accountability and the court's ability to impose sanctions. He explained that the D.C. Court of Appeals has less authority to directly sanction attorneys than federal appellate courts. Glickman agreed the attorneys may have been negligent, but he did not see enough evidence of bad faith, which courts generally need to find before they sanction someone on their own authority.⁹ He disagreed with the idea that every attorney whose name appears on a brief must personally check every citation, calling it impractical and burdensome.¹⁰ Instead, he suggested that law firms [*23] should have a reliable process for reviewing their work.¹¹ Glickman also explained that AI can cause problems beyond making up cases. For example, it can give incorrect information about real court decisions or misrepresent what a case actually says. Because of this, he stressed that attorneys need to use their own judgment and check their research rather than simply trusting AI.¹²

This case sheds light on the growth of AI and how, while it can definitely be useful, it can also be misleading. For the future attorneys of this and later generations who have grown up using AI throughout their education and life, understanding how to use it correctly and knowing the risks that come with it is essential. Of course, there may be times when lawyers feel pressure from their workloads, but this case is an example of why maintaining integrity and verifying any and all information AI produces is so important. This decision should serve as an example for future lawyers who have become accustomed to relying on AI. Being familiar with both the benefits and dangers of this incredible technology can help attorneys avoid damaging not only their careers but also the interests of their clients.

1. Douglas v. Deutsche Bank Nat'l Tr. Co., No. 24-CV-1099, slip op. at 1 (D.C. Sept. 3, 2026) (per curiam). ↩︎
2. Id. ↩︎
3. Id. at 1. ↩︎
4. Id. at 1-2, 5. ↩︎
5. Id. at 6-7. ↩︎
6. Id. at 4-5. ↩︎
7. Id. at 5. ↩︎
8. Id. at 7 & n.5. ↩︎
9. Id. at 10-13 (Glickman, J., concurring). ↩︎
10. Id. at 14-15 (Glickman, J., concurring). ↩︎
11. Id. at 15-16 (Glickman, J., concurring). ↩︎
12. Id. at 16-17 (Glickman, J., concurring). ↩︎