Sanctions Imposed for Improper Use of Generative Artificial Intelligence in Federal Civil Litigation: Analysis of the Order Issued in Lacey v. State Farm

Order of the Special Master (Hon. Michael R. Wilner, Ret.), U.S. District Court, Central District of California, May 6, 2025

‍

1. Background of the Proceeding

In the case of Lacey v. State Farm (C.D. Cal., No. 2:24-cv-05205-FMO-MAA), former Los Angeles District Attorney Jacquelyn “Jackie” Lacey brought suit against her insurer, State Farm, concerning the denial of coverage under a professional liability policy. A discovery dispute arose concerning the scope of the attorney–client privilege invoked by the insurer.

To assist with pretrial discovery, the Court appointed the Hon. Michael R. Wilner (Ret.) as Special Master, pursuant to Rule 53 of the Federal Rules of Civil Procedure. In the course of the proceedings, the Special Master identified serious deficiencies in a brief filed by Plaintiff’s counsel—deficiencies which led to the imposition of financial and procedural sanctions.

‍

2. Misuse of Generative Artificial Intelligence in a Legal Submission

On April 14, 2025, Plaintiff’s counsel filed a brief opposing the assertion of privilege by the defendant. That submission cited several purported federal decisions that were either non-existent, inaccurately cited, or contained fabricated language.

Following inquiries from the Special Master, it was revealed—through sworn declarations—that one attorney had used various generative AI tools, including CoCounsel (by Casetext), Westlaw Precision Drafting (by Thomson Reuters), and Google Gemini, to prepare the legal draft. This draft was then circulated internally and incorporated into the filing without proper review or verification by supervising attorneys. Several lawyers involved were unaware that the content had been generated using AI.

Even more troubling, the errors were not remedied in a subsequently revised brief filed after the Special Master had already raised concerns, thereby confirming a persistent failure in oversight and human verification.

‍

3. Findings and Legal Basis for Sanctions

In a detailed order issued on May 6, 2025, the Special Master concluded that the conduct amounted to “tantamount to bad faith”. The key elements of his reasoning were:

• Under Rule 11 of the Federal Rules of Civil Procedure, attorneys who submit filings to the court certify that they have conducted a reasonable inquiry into the facts and law. The delegation of research or drafting tasks to an AI tool does not discharge this obligation.
• The cited cases and quotations were not merely erroneous, but fabricated or misrepresented, and those inaccuracies persisted even after a formal warning from the Special Master.
• The conduct reflected a systemic failure of human oversight, leading to a waste of judicial resources and unnecessary prejudice to the opposing party
  

The Special Master characterized the violations as egregious and “frightening,” underscoring the danger posed by overreliance on AI without adequate professional scrutiny.

‍

‍4. Sanctions Imposed

The following sanctions were ordered:

• The faulty brief was stricken from the record and the motion to compel was denied.
• The law firms Ellis George Cipollone and K&L Gates were held jointly and severally liable for the sum of $31,100, representing:
  
  • $26,100 in fees incurred by the Special Master, and
  • $5,000 in attorneys’ fees awarded to the defendant for its opposition to the motion.
• The Court issued a clear warning that any further conduct of this nature would result in additional sanctions, including potential referrals to disciplinary authorities.‍

‍

5. Legal and Ethical Implications for Practitioners

This case forms part of a growing body of U.S. federal decisions addressing the risks of unverified use of generative AI in litigation, including Mata v. Avianca (S.D.N.Y., 2023) and Park v. Kim (C.D. Cal., 2024).

The Lacey decision affirms several key principles:

• Generative AI tools may assist with legal research or drafting, but lawyers remain fully responsible for the accuracy and reliability of any content filed in court.
• Disclosure is not sufficient: verification is mandatory. Failing to verify citations, factual assertions, or quotations—regardless of the source—may be sanctionable.
• Law firms must implement internal safeguards and protocols to ensure that AI-generated work product is systematically reviewed by qualified attorneys before submission.

The decision concludes with the Special Master’s stark warning:

“The AI use misled the Court. That’s frightening.”

‍

6.Comparative Perspective: What About French Courts?

As of the date of this publication, no French court has been called upon to sanction the improper use of generative AI in legal proceedings. However, the widespread adoption of such technologies by legal professionals makes the emergence of such litigation highly likely in the near future.

French procedural and professional standards offer a legal framework that could support similar sanctions:

• Article 1.3 of the Règlement Intérieur National (RIN) of the French Bar mandates that lawyers act with competence, diligence, and integrity.
• Article 55 of the Code de procédure civile requires that all claims and defenses be set out in law and in fact, with clarity and precision.
• Established case law recognizes the principle of loyalty in procedure, and French judges may sanction litigants (or their counsel) for submitting inaccurate or deceptive pleadings—whether the result of negligence or misuse of technology.

‍

In future cases, the use of fabricated citations or unverified AI-generated content in French court filings could lead to:

• Striking of the pleadings or disregard of the submission;
• Civil fines under Article 32-1 of the Code de procédure civile;
• Ethical sanctions or disciplinary referrals for breach of professional conduct.