The Volokh Conspiracy
Mostly law professors | Sometimes contrarian | Often libertarian | Always independent
Apparent AI Hallucinations in Filing from Two >500-Lawyer Firms
Don't assume your firm is safe.
From Monday's decision by Judge Michael Simon (D. Ore.) in Green Building Initiative, Inc. v. Peacock:
GBI cites Stell v. Cardenas, No. 3:21-cv-413-HZ, 2022 WL 1696093, at *6 (D. Or. May 26, 2022), and Page v. Parsons, 249 F. Supp. 3d 998, 1017 (D. Or. 2017). Stell, however, is totally fake; it is an "hallucinated" case. There is no case at 2022 WL 1696093. Nor does there exist a case with the parties "Stell" and "Cardenas" (in either orientation—in the District of Oregon or elsewhere). Moreover, Case No. 3:21-cv-413 is a real case, styled JS Halberstam Irrevocable Grantor Trust v. Davis, and it was presided over by the undersigned ("SI"), not U.S. District Judge Hernández ("HZ"). Further, that case was closed by judgment on May 9, 2022, and no decision was issued on May 26, 2022.
Page v. Parsons is almost real. There is a case titled Page v. Parsons that discusses the Oregon anti-SLAPP statute, but it is not a federal case from the District of Oregon. See Page v. Parsons, 249 Or. App. 445 (2012). Page, however, does not discuss when fees are due or may be advanced under the statute. And as a matter of logic, that Oregon Court of Appeals decision cannot support the proposition that GBI cited it for—that federal courts applying Oregon's anti-SLAPP statute have consistently held that fee awards are appropriate immediately following resolution of the special motion to strike. Indeed, a basic internet search easily reveals both the real citation for Page v. Parsons, and the real case listed at 249 F. Supp. 3d 998 (a death penalty habeas decision from the Eighth Circuit). See Deck v. Steele, 249 F. Supp. 3d 991, 998 (E. D. Mo. Apr. 13, 2017), rev'd and remanded, Deck v. Jennings, 978 F.3d 578 (8th Cir. 2020).
GBI's fake case citations appear to be hallmarks of a generative artificial intelligence ("AI") tool, such as ChatGPT…. GBI is hereby ordered to show cause within two weeks as to why the Court should not issue sanctions on both GBI and its counsel for the use of fake citations in violation of Rule 11(b). GBI is also ordered to disclose what it believes to be an appropriate sanction. GBI's proposed sanctions should reflect the nature and seriousness of its use of fake citations and appropriate steps to ensure that this type of violation will not occur again….
Editor's Note: We invite comments and request that they be civil and on-topic. We do not moderate or assume any responsibility for comments, which are owned by the readers who post them. Comments do not represent the views of Reason.com or Reason Foundation. We reserve the right to delete any comment for any reason at any time. Comments may only be edited within 5 minutes of posting. Report abuses.
Please to post comments
What (if any) are the options for the client if their attorneys are sanction by a court for false filings? Do they have a cause of action for malpractice?
An attorney personally being sanctioned — for any type of wrongdoing, not limited to using ChatGPT — is not in itself a basis for a malpractice claim by his client. An attorney's misconduct being a reason for his client losing, of course, would be; you need that causal connection.
EDIT: To be clear, if the client itself is sanctioned, the client would have contract and/or tort remedies against its counsel.
HA!
David Bernstein is an attorney for Green Building Initiative (the company getting their hand slapped).
(OK, not our DB.)
The solution is simple: Lifetime disbarment for any attorney caught doing this -- that would end this tomorrow.
Seems appropriate.
I'm pretty aggressive on this matter, but I think 6 month's suspension from practicing law would wake the rest of them up.
Everything's simple when you're a simpleton.
So, I suppose my question is this...
Are there any unifying characteristics between the lawyers that are doing this? Are they new associates who are writing the cases and are just "used to" using AI to write summaries? Are they 50 year old partners who are trying something new, and don't know any better? Are they paralegals who are "writing" the filing and the lawyer actually on the case just takes the paralegal's word for it, says it looks good, and doesn't check any further?
Just curious.
The last case I saw here it was supposedly an intern from a law school.