Garner v. Transit Employees Federal Credit Union
U.S. District Court, District of Columbia · D.D.C. · District of Columbia bar guidance
Conduct
Here the court found plaintiff's opposition brief quoted three D.C. Circuit decisions for language or propositions that do not appear in them.
Consequence
Order to show cause why counsel should not be sanctioned, with a citation-by-citation accounting and an explanation of any generative AI use.
Lesson
Here the court required counsel to audit every citation in the brief, not just the three it identified, as part of the show-cause response.
Verified September 30, 2026
- Citation
- Garner v. Transit Employees Federal Credit Union, No. 25-4024 (LLA), Memorandum Opinion and Order (D.D.C. Sept. 22, 2026) (AliKhan, J.), ECF No. 12
- Filing date
- September 22, 2026
Summary
Wanda Garner sued her former employer, the Transit Employees Federal Credit Union (TEFCU), under Title VII, the D.C. Human Rights Act, and the D.C. Family and Medical Leave Act. Resolving TEFCU's motion for summary judgment or dismissal, the court turned last to Ms. Garner's opposition brief, which it said "is littered with incorrect and misleading citations and quotations." It gave three examples. Counsel quoted Aka v. Washington Hospital Center, 156 F.3d 1284 (D.C. Cir. 1998), "for propositions that do not appear in the case," including that comparators need not be "mirror images" and that discriminatory intent "often must be inferred from circumstantial evidence found in affidavits and depositions." Counsel quoted Waterhouse v. District of Columbia, 298 F.3d 989 (D.C. Cir. 2002), for a statement the court said the D.C. Circuit "did not say," and quoted Brown v. Brody, 199 F.3d 446 (D.C. Cir. 1999), for language on similarly situated employees that "is not in Brown." The court added that it "could go on." The order does not name counsel; it refers only to "Ms. Garner's counsel," "counsel for Ms. Garner," and "Plaintiff's counsel."
- AI tool:
- Unidentified (the court orders counsel to explain "whether he used generative artificial intelligence when writing his brief"; no finding or admission of AI use)
- Amount or terms:
- None imposed as of the September 22, 2026 order; show-cause response due October 6, 2026
What is the current procedural posture?
District Judge Loren L. AliKhan granted the motion only as to the D.C. Family and Medical Leave Act claim, which was dismissed with prejudice, and otherwise denied it, holding summary judgment premature under Rule 56(d). She then ordered plaintiff's counsel to "SHOW CAUSE on or before October 6, 2026 why he should not be sanctioned for the incorrect and misleading citations and quotations in his opposition brief." The submission must include (1) a list of every authority cited in the opposition with an explanation of whether each citation or quotation is substantiated, (2) a description of what caused the errors, "including whether counsel used generative artificial intelligence when drafting," and (3) a statement of what sanction, if any, would be appropriate. The order identifies the potential bases as counsel's duty of professional judgment, D.C. Rule of Professional Conduct 3.3(a)(1), and Fed. R. Civ. P. 11(b)(2). No sanction had been imposed as of the order.
Why does Garner v. Transit Employees Federal Credit Union matter for law firms using AI?
The show-cause order in Garner arrives at the end of a ruling that went largely in the plaintiff’s favor. The court denied summary judgment as premature, rejected the defendant’s forfeited Rule 12(b)(6) argument and its “phantom request to strike,” and dismissed only the D.C. Family and Medical Leave Act count. The court then turned to the plaintiff’s own brief, stating that “[i]t is unacceptable to file briefs ‘containing false, misleading, or nonexistent quotations or authorities,’” and citing the D.C. Rules of Professional Conduct and Rule 11(b)(2) as the duties potentially implicated.
The three examples the court gave are not invented cases. Aka, Waterhouse, and Brown are real D.C. Circuit decisions; the problem the court identified is that the brief put quotation marks around language those opinions do not contain. The court did not find that generative AI was involved. It asked counsel to explain “whether he used generative artificial intelligence when writing his brief,” alongside a citation-by-citation accounting and counsel’s own view of an appropriate sanction.
The matter is pending. Firms documenting compliance may wish to consider that the response this court required (a full list of authorities, an explanation of each, and a description of how the errors arose) is the kind of record a firm’s own citation-verification process could already have produced before the brief was filed.
Implications for your firm
Operational steps a firm reading this case may wish to consider documenting. Strategic and rule-application calls belong to your firm's attorneys.
- Verify every quotation against the cited opinion before filing. Here the three examples were real D.C. Circuit cases quoted for words the court found are not in them.
- Document how each brief was drafted, including any AI tool used. Here the court required counsel to describe what caused the errors, including whether generative AI was used.
- Review the full citation list when one error surfaces. Here the court ordered a list of all authorities cited in the opposition with an explanation for each.
- Consider that a court may raise citation problems in the same ruling that goes largely for the filer on the merits: here the court largely denied the defendant's motion and in the same order directed counsel to show cause.
Sources
Primary sources
- AI attribution is a Charlotin tracker inference ('Implied'): the order asks counsel to explain whether he used generative AI but makes no finding that he did.
- Counsel's identity: the court's own copy does not name plaintiff's counsel; a Westlaw copy lists counsel of record, but that is not a primary source and the entry does not rely on it.
- The outcome of the show-cause order (response due October 6, 2026) had not been located as of 2026-09-30.