Hall v. Academy Charter School, No. 2:24-cv-08630-JMW (E.D.N.Y. 2025)

Case 2:24-cv-08630-JMW Document 32 Filed 08/07/25 Page 1 of 12 PageID #: 123

FILED CLERK

UNITED STATES DISTRICT COURT 8/7/2025 EASTERN DISTRICT OF NEW YORK

--------------------------------------------------------------X

SASHANE HALL,

Plaintiff,

2:24-cv-08630-JMW

-against-

THE ACADEMY CHARTER SCHOOL,

Defendant.

--------------------------------------------------------------X

Suryia Rahman Naresh M. Gehi Gehi and Associates 173-29 Jamaica Ave., Jamaica, NY 11432 Attorneys for the Plaintiff

Adam Granek Guttell Anahi Tapia Jackson Lewis, P.C. 58 South Service Road Suite 250 Melville, NY 11747 Attorneys for Defendant

WICKS, Magistrate Judge:

“[L]awyers are essential to the primary governmental function of administering justice,

and have historically been ‘officers of the courts.’” Goldfarb v. Virginia State Bar, 421 U.S. 773,

792 (1975). As such, an attorney appearing on behalf of a client is obligated to provide

“competent representation,” which means the attorney must possess the “legal knowledge, skill,

thoroughness and preparation reasonably necessary for the representation.” N.Y. Rules of Prof. Conduct 1.1(a). That requires “an advocate to disclose directly adverse and controlling legal authority that is known to the lawyer . . . .”. Id. at cmt. 4 to Rule 3.3. The reason of course is that “[a] tribunal that is fully informed on the applicable law is better able to make a fair and accurate determination of the matter before it.” Id. That is not what occurred here.

The Court is confronted with what seemingly appears to an issue plaguing the modern legal justice system: the use of generative artificial intelligence (“AI”) to assist in the drafting of legal papers that results in the generation of a legal submission containing citations to fictitious or non-existent legal authority – commonly referred to as “hallucinated” or fake case citations. In a brief filed in opposition to Defendant’s pre-motion conference letter in support of a motion for partial dismissal (ECF No. 24), Plaintiff’s counsel relied upon AI-generated content that created fictitious legal citations. That conduct clearly violated Federal Rule of Civil Procedure 11. See Mata v. Avianca, Inc., 678 F. Supp. 3d 443, 462 (S.D.N.Y. 2023) (Castel, J.).

The question presented is whether the circumstances here constitute sanctionable conduct under Rule 11 and the applicable standards for competence and candor before a tribunal.

Plaintiff Sashane Hall (“Plaintiff”) commenced this action against Defendant The Academy Charter School (“Defendant”) on December 18, 2024 asserting claims under Title VII of the Civil Rights Act of 1964, 42 U.S.C. §2000(e) et seq., Title IX of the Education Amendments of 1972, 20 U.S.C. § 1681 et seq., the New York State Human Rights Law §§ 296, et seq., 42 U.S.C. § 1981(a), for discrimination based on sexual orientation, and hostile and abusive working environment. (See generally ECF Nos. 1, 22.) Plaintiff filed her Amended Complaint (ECF No. 22) following a pre-motion conference held on April 4, 2025. (ECF No.

20.) Thereafter, on May 30, 2025, Defendant filed a letter requesting a second pre-motion conference, this time on its anticipated motion to partially dismiss the Amended Complaint under Fed. R. 12(b)(6). (ECF No. 23.) Plaintiff filed her opposition letter in the form of a brief (“Opposition”) on June 6, 2025, arguing, inter alia, that charter schools are not subject to notice of claim requirements because they are not government entities like public schools and school districts. (See ECF No. 24.)1

The cases Plaintiff cites in the Opposition to support this proposition, however, do not exist. (See ECF No. 25 at pp. 1-2.) Instead, Plaintiff includes three hallucinated cases: (1) Laskowski v. Liberty Partners Restaurant Group Inc., 2011 WL 817498, at \(S.D.N.Y. Mar. 9, 2011), (2) Lindner v. Forest Hills Montessori School, 2011 WL 1334869, at *4 (E.D.N.Y. Apr. 6, 2011), and (3) Matter of K.M. v. Bronx Charter School for Better Learning, 2011 NY Slip Op 32728(U) (Sup. Ct. Bronx County). (ECF No. 24 at p. 3; ECF No. 25 at p. 2.) As such, on June 16, 2025, Defendant requested an additional pre-motion conference before moving for an order striking Plaintiff’s opposition entirely and for an award of attorneys’ fees and costs stemming from Defendant’s application. (ECF No. 25 at p. 1.) Defendant argues that “Plaintiff’s citations to fabricated, misleading, and irrelevant legal authority [] cannot be explained away as an innocent ‘typo,’ mistake, or misunderstanding of the law,” and such conduct “deprives both Defendant and the Court the opportunity to properly evaluate Plaintiff’s argument on its merits.” (Id. at p. 3.)

In response, the same day, counsel for Plaintiff2 filed a short letter acknowledging receipt of Defendant’s motion, stated that she understood the “grave implications of misreporting

case law to the Court,” and assured the Court that her firm was internally investigating the issue. (ECF No. 26.) Moreover, in this same letter, counsel for Plaintiff voluntarily discontinued the claims under New York State Human Rights Law—later memorialized in a stipulation filed on June 30, 2025. (ECF Nos. 26, 27.)

On July 2, 2025, the Court issued an Electronic Order “so ordering” the stipulation and further directing that “counsel for Plaintiff show cause, in writing on or before July 18, 2025 why the Court should not impose sanctions upon counsel for the conduct indicated in Defendant’s pre-motion letter at ECF No. 25 for the fabricated legal authority contained in Plaintiff’s opposition likely result[ing] from counsel’s misuse of artificial intelligence. (ECF No. 28) (internal citations omitted) (alteration in original). Counsel for Plaintiff filed her response on July 18, 2025. (ECF No. 29.) In her response, counsel for Plaintiff explained that the Opposition was drafted by a clerk who used Google for research. (Id. at p. 1.) The clerk sent the work product to counsel for Plaintiff who “reviewed the draft Opposition but did not check the citations.” (Id.) Counsel for Plaintiff attributed the “main reason” for her failure to check the citations to the death of her spouse, who recently passed away unexpectedly, stating that “[t]he shock and grief resulting from my husband’s death has had a profound impact on all aspects of my life,” and explaining that her husband’s death “has affected [her] ability to attend to the practice of law with the same focus and attention as before.” (Id. at pp. 1-2.) Counsel for Plaintiff has since taken bereavement leave and is “continuously meeting with medical and mental health professionals.” (Id. at p. 1.)

Rule 11 governs attorneys’ representations to the Court, and therefore, informs situations such as the present one. See generally Fed. R. Civ. P. 11. Rule 11 states:

By presenting to the court a pleading, written motion, or other paper —whether by signing, filing, submitting, or later advocating it—an attorney . . . certifies that to the best of the person's knowledge, information, and belief, formed after an inquiry reasonable under the circumstances . . . 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. Fed. R. Civ. P. 11(b)(2).

“Rule 11 imposes a duty on attorneys to certify that they have conducted a reasonable inquiry and have determined that any papers filed with the court are well grounded in fact, [and] legally tenable.” Benjamin v. Costco Wholesale Corp., No. 2:24-cv-7399 (LGD), 2025 WL 1195925, at \(E.D.N.Y. 2025) (quoting Park v. Kim, 91 F.4th 610, 614 (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.” Park, 91 F.4th at 615. Indeed, a claim is frivolous and warrants sanctions when a “legal position has ‘no chance of success.’” Fishoff v. Coty, Inc., 634 F.3d 647, 654 (2d Cir. 2011) (quoting Morley v. Ciba–Geigy Corp., 66 F.3d 21, 25 (2d Cir. 1995)). When “a court considers whether to impose sanctions sua sponte, it ‘is akin to the court's inherent power of contempt,’ and, ‘like contempt, sua sponte sanctions in those circumstances should issue only upon a finding of subjective bad faith.’” Mata, 678 F. Supp. 3d at 462 (alterations omitted) (quoting Muhammad v. Walmart Stores East, L.P., 732 F.3d 104, 108 (2d Cir. 2013) (per curiam)). “Subjective bad faith is ‘a heightened mens rea standard’ that is intended to permit zealous advocacy while deterring improper submissions.” Mata, 678 F. Supp. 3d at 462 (alterations omitted) (quoting In re Pennie & Edmonds LLP, 323 F.3d 86, 91 (2d Cir. 2003)).

Further, this Court possesses the inherent power to sanction a party “for conduct which abuses the judicial process.” Chambers v. NASCO, Inc., 501 U.S. 32, 44–45 (1991). A court’s inherent power is “governed not by rule or statute but by the control necessarily vested in courts

to manage their own affairs so as to achieve the orderly and expeditious disposition of cases.” Id. at 43 (quoting Link v. Wabash R. Co., 370 U.S. 626, 630–31 (1962)). Notably, a court possesses “‘wide discretion’ to craft an appropriate sanction, and may consider the effects on the parties and the full knowledge of the relevant facts” gained throughout sanctions hearings. Heaston v. City of N.Y., 19-CV-5569 (PKC) (VMS), 2022 WL 182069, at *9 (E.D.N.Y. Jan. 20, 2022) (quoting Oliveri v. Thompson, 803 F.2d 1265, 1280 (2d Cir. 1986)). For instance, courts can exercise their inherent authority by “suspend[ing] or disbar[ing] lawyers,” In re Snyder, 472 U.S. 634, 643 (1985), and in some circumstances impose “particularly severe sanction[s]” such as “outright dismissal” of a case. Chambers, 501 U.S. at 45.

Moreover, “any attorney or other person admitted to conduct cases in any court of the United States or any Territory thereof who so multiplies the proceedings in any case unreasonably and vexatiously may be required by the court to satisfy personally the excess costs, expenses, and attorneys' fees reasonably incurred because of such conduct.” 28 U.S.C. § 1927. For a court to impose sanctions pursuant to § 1927, there must have been bad faith motive on the offending party to delay the litigation. GiftRocket, Inc. v. Buchnik, No. 24 MC 5105 (RPK) (VMS), 2025 WL 888483, at *2 (E.D.N.Y. Mar. 21, 2025). “Unlike Rule 11 sanctions which focus on particular papers, the inquiry under § 1927 is on a course of conduct.” Hernandez v. Money Source Inc., No. 17-CV-6919 (GRB) (AYS), 2022 WL 2702894, at *8 (E.D.N.Y. July 12, 2022) (quoting Bowler v. U.S. Immigration & Naturalization Serv., 901 F. Supp. 597, 605 (S.D.N.Y. 1995); Mahoney v. Yamaha Motor Corp. U.S.A., 290 F.R.D. 363, 367 (E.D.N.Y. 2013)).

As one court observed, “there is nothing inherently wrong with an attorney properly and competently utilizing AI or any of its subsets to practice law or litigate cases.” Versant Funding

LLC v. Teras Breakbulk Ocean Navigation Enter., LLC, No. 17-cv-81140-DIMITROULEAS/MATTHEWMAN, 2025 WL 1440351, at *4 (S.D. Fla. May 20, 2025). Nevertheless, “[a]ttorneys and courts need to be aware of both the benefits and limitations that these AI platforms present.” Id. (quoting Judge Xavier Rodriguez, Artificial Intelligence (AI) and the Practice of Law, 24 SEDONA CONF. J. 783, 791 (2023)). Indeed, “[a] basic prerequisite to the filing of any pleading, motion, response, reply, or paper in court is for the drafting and filing attorney(s) to carefully check every case citation, fact, and argument to make sure that they are correct and proper.” Id. This obligation existed long before AI entered the legal industry landscape.

The appearance of hallucinated citations in briefs generated from AI is no longer in its nascent stage. Regrettably, the number and regularity with which courts have been faced with hallucinations in court filings continues to rise both in this country and abroad. See Damien Charlotin, AI Hallucination Cases, (Aug. 6, 2025) https://www.damiencharlotin.com/hallucinations/ (database tracking legal decisions “in cases where generative AI produced hallucinated content,” evidencing 255 cases to date) (hereinafter “Charlotin Database”). This trend of AI hallucinations has steadily increased over the years. See Cecily Mauran, *120 court cases have been caught with AI hallucinations, according to new database, MASHABLE (May 27, 2025), https://mashable.com/article/over-120-court-cases-caught-ai-hallucinations-new-database.*

Courts addressing filings containing AI-generated hallucinations have fashioned various forms of sanctions imposed upon the offending attorneys after they were found to have used these fabricated citations in “bad faith.” For instance, in Park v. Kim, an attorney who submitted a reply brief with hallucinated cases was referred to the Second Circuit’s Grievance Panel with

consideration of referral to the Committee on Admissions and ordered to provide a copy of the Court’s ruling to their client. 91 F.4th at 616. The Second Circuit found that the act of presenting non-existent cases to the court demonstrated that the attorney made no inquiry into the validity of the arguments she presented and failed to determine her argument was “legally tenable” under Rule 11. Id. at 615. Similarly, in Benjamin, the court held that the offending attorney acted in bad faith by utilizing “AI to produce work she could not (or would not) do on her own,” yet failed to read the case law prior to submitting an affirmation to court. 2025 WL 1195925, at *8. Accordingly, because the facts “suggest[ed] much more than mere carelessness,” the court imposed a $1,000 fine and ordered the offending attorney to serve a copy of the court’s order on their client. Id. at *9.

Likewise, in Mata, one of the earliest cases addressing this issue, the court determined that the offending attorneys acted in bad faith and thus ordered them to, inter alia, provide a copy of the court’s order to their clients and mail a letter and a copy of the court’s order to each of the judges that the offending attorneys falsely attributed hallucinated cases to. 678 F. Supp. 3d at 466. The “bad faith” conduct in Mata involved the offending attorneys citing to non-existent cases generated by ChatGPT in motion papers and for “not reading a single case cited in [the submitted affirmation] and taking no other steps on his own to check whether any aspect of the assertions of law were warranted by existing law.” Id. at 456, 464. Indeed, in Mata, the court noted that taking these affirmative steps, signing and filing the affirmation, and relying on another attorney’s lack of familiarity with the relevant law, a fact known by the offending attorney, the conduct was “an act of subjective bad faith.” Id. at 464. Further, the other offending attorney falsely asserted that his use of ChatGPT was “merely a ‘supplement’ to his research,”

whereas the record reflected that the attorney solely relied on the AI database for his research. See id. at 465. Accordingly, Rule 11 sanctions were appropriate. Id. at 466.

Indeed, the imposition of sanctions resulting from an attorney’s bad faith use of generative AI runs nationwide. See, e.g., Coomer v. Lindell, No. 22-cv-01129-NYW-SBP, 2025 WL 1865282, at *6, *8 (D. Colo. July 25, 2025) (imposing a $3,000 fine on the offending attorneys and their law firm after submitting a brief with hallucinated and inaccurate case outcomes because the attorneys “were not reasonable in certifying that the claims, defenses, and other legal contentions contained [the filing] were warranted by existing law or by a nonfrivolous argument for extending, modifying, or reversing existing law or for establishing new law”) (citing Fed. R. Civ. P. 11(b)(2)); Versant, 2025 WL 1440351, at *7 (ordering the offending attorney to pay the attorney’s fees and costs incurred from the AI generated motion and subsequent replies, attend a CLE course on Artificial Intelligence, and pay a $1000 fine); Wadsworth v. Walmart Inc., 348 F.R.D. 489, 494, 497–98 (D. Wyo. 2025) (revoking one offending attorney’s pro hac vice status because as the drafter he maintained oversight to ensure he filed a meritorious motion on behalf of his client, and imposed monetary fines on certain attorneys for failing to adhere to their Rule 11 obligations, after the attorneys were found to have used the firm’s in-house AI platform to find case law and neglected to check the cases generated); Mid Cent. Operating Eng’r Health and Welfare Fund v. Hoosiervac LLC, No. 2:24-cv-00326-JPH-MJD, 2025 WL 1511211, at *1, *2 (S.D. Ind. May 28, 2025) (imposing a $6,000 fine on the offending attorney for submitting three separate briefs that all contained hallucinated case law in order to “deter repetition of the conduct or comparable conduct by others similarly situated”) (citing Fed. R. Civ. P. 11(c)(4)).

AI-generated caselaw also seems to have unwittingly worked its way into judicial decisions and orders. See Justin Henry, Judge Scraps Opinion after Lawyer Flags Made-Up Quotes (Correct), BLOOMBERG LAW, (July 23, 2025), https://news.bloomberglaw.com/business-and-practice/judge-withdraws-pharma-opinion-after-lawyer-flags-made-up-quotes; Debra Cassens Weiss, After second federal judge withdraws error-riddled ruling, litigants seek explanation, ABA JOURNAL, (July 30, 2025), https://www.abajournal.com/news/article/litigants-seek-explanation-after-second-federal-judge-withdraws-error-riddled ruling#:~:text=Mississippi%20Attorney%20General%20Lynn%20Fitch%20has%20asked%20a %20federal%20judge,ruling%20back%20on%20the%20docket. (noting the court opinion incorrectly “referenced allegations and parties not in the lawsuit, nonexistent declarations by four people, and language not found in the state law being challenged”).

By far, the majority of courts impose sanctions upon the offending lawyer for this sort of conduct and warnings or reprimands have been meted out in cases typically involving pro se litigants. See Charlotin Database, supra. However, there are circumstances where, in the Court’s discretion, monetary sanctions have not been imposed notwithstanding the violation of Rule 11. For example, in U.S. v. Cohen, the Court declined to impose sanctions upon an attorney who submitted three non-existent cases to the tribunal contained in a motion seeking early termination of supervised release. 724 F. Supp. 3d 251, 253, 258 (S.D.N.Y. 2024). There, the attorney who cited fake cases reviewed a draft that was written by his client—another lawyer— and reviewed by another attorney. Id. at 258–59. Although “embarrassing and certainly negligent, perhaps even grossly negligent” and evidencing “extreme carelessness,” the court declined to impose sanctions as no showing was made that the attorney’s citations to non-existent cases were generated or submitted in bad faith. Id.

Similar to counsel in U.S. v. Cohen, the admitted conduct here was the result of extreme carelessness and negligence, done under tragic personal circumstances. The explanation proffered to the Court that counsel failed to check the citations generated by her clerk who used Google for research does not evince bad faith or willfulness, but rather severe carelessness. (See ECF No. 29 at p. 1.) Unlike in Mata, counsel here had no reason to believe that the clerk who drafted the Opposition was unfamiliar with notice of claim requirements under New York state law. Further, counsel for Plaintiff’s reason for neglecting to check this work was the sudden, unexpected death of her spouse. (Id.) Indeed, Plaintiff explained that “[t]he shock and grief resulting from my husband’s death has had a profound impact on all aspects of my life,” and explained that her husband’s death “has affected [her] ability to attend to the practice of law with the same focus and attention as before.” (Id. at pp. 1–2.) As such, counsel for Plaintiff has not only repeatedly apologized to the Court and opposing counsel, assuring both that “[t]his will not happen again,” she also indicated her intentions on seeking medical and mental treatment. (See id. at pp. 1-2.)3 Notably, contrary to the circumstances in Mata where the offending attorneys misrepresented their use of ChatGPT to the court, counsel for Plaintiff outright admitted and took “full responsibility for failing to check the citations in the Opposition.” (Id. at p. 1.) Counsel also stipulated to the relief sought on the anticipated motion by ultimately withdrawing certain claims.4

While the Court is aware of the serious implications that the misuse of AI-generated non-existent caselaw presents, it is also mindful of the circumstances here which do not support any

finding of bad faith, as well as the remorseful explanations proffered by counsel who experienced one of life’s unspeakable tragedies. It is for these reasons that the Court in the exercise of its discretion, declines to impose sua sponte monetary sanctions upon counsel for Plaintiff.

For the reasons stated herein, the Court declines to exercise its inherent power to impose monetary sanctions upon counsel for Plaintiff. Counsel is, however, admonished and this Order should serve as a forewarning. Counsel is further directed to serve Plaintiff with a copy of this Memorandum and Order. Dated: Central Islip, New York

August 7, 2025

S O O R D E R E D: /s/ James M. Wicks

JAMES M. WICKS

United States Magistrate Judge