Case Information
United States District Court O Southern District of Texas ENTERED July 23, 2025 UNITED STATES DISTRICT COURT Nathan Ochsner, Clerk SOUTHERN DISTRICT OF TEXAS LAREDO DIVISION
KARINA ELIZONDO §
§ VS. § CIVIL ACTION NO. 5:25-cv-50
§
CITY OF LAREDO §
ORDER
Before the Court is the question of whether to impose sanctions on Plaintiff’s counsel, Edward L. Piña, for filing a response to Defendant’s motion to dismiss that contained multiple fictitious or materially inaccurate case citations (Dkt. No. 18). The Court ordered Mr. Piña to show cause why he should not be sanctioned, and Mr. Piña filed a written response (Dkt. Nos. 23, 24). [1] After reviewing the response, the Court concludes that Mr. Piña violated Rule 11 of the Federal Rules of Civil Procedure by submitting a filing with fabricated citations and misstatements of law that were not properly reviewed for accuracy. As such, the Court will impose sanctions.
I. BACKGROUND On May 16, 2025, Defendant filed a motion to dismiss Plaintiff’s discriminatory termination claims, arguing that Plaintiff lacked a right-to-sue letter from the Equal Employment Opportunity Commission as to those claims (Dkt. Nos. 14, 15). Plaintiff filed a response on May 22, 2025 (Dkt. No. 18). Upon review, the Court identified multiple citations in Plaintiff’s response that appeared fictitious or materially inaccurate. The Court described the full nature of the misleading citations in its Order to Show Cause issued on June 19, 2025, which required Mr. Piña to explain why he should not be sanctioned under Rule 11, the Court’s inherent authority, and the local rules (Dkt. No. 23). Without rehashing the details in depth here, it will suffice to say that Plaintiff’s response cited four cases with wildly inaccurate citations—misstating case numbers, dates of decisions, reporters, Westlaw citations, page numbers, and confusing the federal and state court systems ( see Dkt. Nos. 18 at 4–6; 23 at 2–3). More importantly, Plaintiff also misrepresented the legal holdings of the cited cases ( see Dkt. Nos. 18 at 4–6; 23 at 2–3). The only accurate citation in the brief was one originally provided by Defendant ( see Dkt. Nos. 15 at 4; 18 at 4; 23 at 3–4). These issues led the Court to conclude that generative artificial intelligence may have been used when the brief was prepared, without adequate verification for accuracy (Dkt. No. 23 at 1–2, 4).
In response, Mr. Piña stated he was shocked at the Court’s mention of generative artificial intelligence, as his personal technological prowess “is, at best, limited” (Dkt. No. 24 at 4). However, he discovered that his law clerk used generative artificial intelligence tools to produce the case citations (Dkt. No. 24 at 4). Mr. Piña acknowledged that he failed to verify their accuracy, and that the ultimate responsibility falls on him (Dkt. No. 24 at 1–2, 4–5). Mr. Piña offered his sincere apologies and regrets to the Court, the parties, and opposing counsel (Dkt. No. 24 at 1–2). He claims that he has since implemented new internal policies prohibiting the use of generative AI tools in drafting legal filings and requiring heightened citation review (Dkt. No. 24 at 6).
II. LEGAL STANDARDS Federal Rule of Civil Procedure 11(b)(2) requires attorneys presenting legal filings to the Court to certify that “the claims, defenses, and other legal contentions are warranted by existing law or by a nonfrivolous argument for extending, modifying, or reversing existing law or for establishing new law.” The rule requires the signing attorney “to satisfy himself that the filed paper is factually and legally responsible,” and by signing, he indicates “that he personally has applied his own judgment.” Pavelic & LeFlore v. Marvel Ent. Grp. , 493 U.S. 120, 125 (1989). This is a “nondelegable responsibility.” Id. at 126. An attorney’s subjective good faith is no shield against Rule 11 sanctions. Jenkins v. Methodist Hosps. of Dall., Inc. , 478 F.3d 255, 264 (5th Cir. 2007) (citation omitted). The same is true for local rule violations. In re Goode , 821 F.3d 553, 559 (5th Cir. 2016) (Prado, J.).
After finding a Rule 11 violation, district courts retain “considerable discretion in determining the ‘appropriate’ sanction to impose upon the violating party.” Thomas v. Cap. Sec. Servs. , 836 F.2d 866, 877 (5th Cir. 1988) (quoting Fed. R. Civ. P. 11(c)). However, “the sanction imposed should be the least severe sanction adequate to the purpose of Rule 11,” which is “deterrence, punishment, and compensation.” Id. at 878– 79; Fed. R. Civ. P. 11(c)(4) (“A sanction imposed under this rule must be limited to what suffices to deter repetition of the conduct or comparable conduct by others similarly situated.”).
In the Southern District of Texas, Chief Judge Crane issued General Order 2025- 04 on May 7, 2025, addressing generative AI use in court filings. U.S. Southern District of Texas General Order No. 2025-04 (May 7, 2025), https://www.txs.uscourts.gov/ file/8731/download?token=LzfiVT-t. General Order 2025-04 cautions attorneys against submitting filings drafted using generative AI “without checking the submission for accuracy as certain technologies may produce factually or legally inaccurate content and should never replace the lawyer’s independent legal judgment.” It notes that attorneys will be held responsible for the contents of their filings under Rule 11, “regardless of whether generative artificial intelligence drafted any portion of that filing.” Id.
Mere days before the filing at issue here, the Undersigned also updated her local rules to similarly caution attorneys and self-represented litigants who choose to use generative AI when preparing legal filings. J. Garcia Marmolejo Civ. Ct. P. 8. The Undersigned’s new rule explicitly notes that “[t]he Court will not accept the excuse that such content was prepared by AI, staff, or others when assessing potential violations of Rule 11 or applicable ethical obligations.” Id. Importantly, the Undersigned’s rules do not discourage or prohibit the use of generative AI. That would be an imprudent limitation on a rapidly evolving technology, which may have many beneficial uses when responsibly employed. Rather, the Undersigned’s local rule warns attorneys that because generative AI tools “are capable of producing content that may be factually incorrect or legally unsound,” they “must ensure that any filing prepared with the assistance of generative artificial intelligence . . . is thoroughly reviewed for factual and legal accuracy prior to submission.” Id.
III. DISCUSSION The Court finds that Mr. Piña’s filing of a brief containing fabricated or materially inaccurate case citations, which he acknowledges stemmed from the use of generative AI, constitutes a Rule 11 violation. “The Response presents a false statement of law to the court, and it appears that [Mr. Piña] failed to make a reasonable inquiry required by Rule 11 . . . into the validity of the arguments he presented.” Gauthier v. Goodyear Tire & Rubber Co. , No. 1:23-CV-281, 2024 WL 4882651, at *3 (E.D. Tex. Nov. 25, 2024); Wadsworth v. Walmart Inc. , 348 F.R.D. 489, 495 (D. Wyo. 2025) (“Because there is no dispute that Respondents cited fake cases in a signed motion, Respondents’ conduct violates Rule 11(b)(2).”). While Mr. Piña attributes the errors to his law clerk’s unsupervised AI use, this does not excuse his failure to verify the cited authorities (Dkt. No. 24 at 4–5). As stated above, an attorney’s responsibility to review a legal filing for factual and legal accuracy is “nondelegable.” Pavelic , 493 U.S. at 126. By signing and filing the response, Mr. Piña represented that he had conducted a reasonable inquiry into the accuracy of its contents, which he did not ( see Dkt. No. 24 at 4 (conceding that he “read [his law clerk’s] draft but did not thoroughly examine the cases cited therein”)). See Park v. Kim , 91 F.4th 610, 615 (2d Cir. 2024) (“At the very least, the duties imposed by Rule 11 require that attorneys read, and thereby confirm the existence and validity of, the legal authorities on which they rely.”)
The Court recognizes Mr. Piña’s candid acknowledgment of the problem and contrition. His conduct subsequent to the show-cause order weighs in favor of a measured sanction. See Benjamin v. Costco Wholesale Corp. , No. 2:24-CV-7399 (LGD), 2025 WL 1195925, at *9 (E.D.N.Y. Apr. 24, 2025) (“The Court’s sanction is lower than in many other similar cases in part because of [the attorney’s] candor and sincere regret.”). Nevertheless, “this entire matter has been a waste of the Court’s resources,” as the Court was “left with no choice but to survey the case law regarding attorney misconduct relating to the use of AI.” Benjamin , 2025 WL 1195925, at *7; see also Mata v. Avianca, Inc. , 678 F. Supp. 3d 443, 448 (S.D.N.Y. 2023) (“Many harms flow from the submission of fake opinions.”); Ferris v. Amazon.com Servs. , No. 3:24-CV-304-MPM-JMV, 2025 WL 1122235, at *1 (N.D. Miss. Apr. 16, 2025) (“AI is a powerful tool, that when used prudently, provides immense benefits. When used carelessly, it produces frustratingly realistic legal fiction that takes inordinately longer to respond to than to create.”). Given that the Laredo Division is one of the busiest court dockets in the nation, there are scant resources to spare ferreting out erroneous AI citations in the first place, let alone surveying the burgeoning caselaw on this subject.
The seriousness of filing fictitious case citations—and the corresponding waste of judicial and party resources—warrants a penalty sufficient to deter future violations like Mr. Piña’s, which courts have unfortunately “seen a rash of” in recent years. Willis v. U.S. Bank Nat’l Ass’n as Tr., Igloo Series Tr. , No. 3:25-CV-516-BN, 2025 WL 1224273, at *2 (N.D. Tex. Apr. 28, 2025) (citations omitted); Benjamin , 2025 WL 1195925, at *1 (collecting cases).
The Court also observes that Defendant, the City of Laredo, could have flagged these fictitious citations in a reply brief or supplemental filing. See Gauthier , 2024 WL 4882651, at *1 (opposing party flagged AI misuse in its reply brief); Ferris , 2025 WL 1122235, at *1 (opposing party alerted the court to hallucinated cases). Although this oversight does not rise to the level of sanctionable conduct, the Court expects all parties to assist in maintaining the integrity of the judicial process by alerting the Court to such errors. The Court encourages greater diligence in flagging citation errors in the future— otherwise, the risk is too great that such errors will persist undetected, potentially leading to an outcome unsupported by law.
Lastly, the Court appreciates Mr. Piña’s efforts to take corrective measures at his firm by prohibiting the use of generative AI in response to this incident (Dkt. No. 24 at 6). But it bears emphasizing that the Court’s rules do not prohibit the use of such tools.
Rather, as explained in General Order 2025-04 and the Undersigned’s local rules, attorneys must exercise independent legal judgment and thoroughly review any filing prepared with the assistance of generative AI before submission.
IV. CONCLUSION For the foregoing reasons, the Court SANCTIONS Edward L. Piña as follows: 1. A monetary sanction of $2,500, to be paid into the Court registry within fourteen days of this Order . 2. Completion of a minimum of three hours of Continuing Legal Education in ethics or legal technology, including at least one hour on the use of generative AI in the legal context , within ninety days of this Order , with proof of compliance filed on the docket.
3. Service of a copy of this Order on Plaintiff Karina Elizondo and filing of a certificate of service within fourteen days of this Order .
It is so ORDERED .
SIGNED July 23, 2025.
______________________________ Marina Garcia Marmolejo United States District Judge
[1] As the Court afforded Mr. Piña an opportunity to respond in writing, a sanctions hearing was not required. Merriman v. Sec. Ins. Co. of Hartford , 100 F.3d 1187, 1192 (5th Cir. 1996) (no hearing required if sanctioned party is afforded an opportunity to respond in writing); Collins v. Lawrence , 273 F.3d 1104 (5th Cir. 2001) (citing Merriman , 100 F.3d at 1191) (same).