A Summary of the LACBA Litigation Section CLE Program
By Jamison Gilmore, Weinberg Gonser LLP
WG Associate Jamison Gilmore recently spoke at a Los Angeles County Bar Association panel titled AI in Civil Litigation: What Judges and Lawyers Need to Know Now. He was joined by Hon. Michael Wilner (Ret.) of JAMS, David Ter-Petrosyan of Dykema LLP, and Jeff Koncius of Kiesel Law.
Program Overview
The program addressed four core areas. First, Judge Wilner drew on his thirteen years on the federal bench in the Central District of California to discuss the ethical framework governing AI use, including the duties of competence (RPC 1.1), confidentiality (RPC 1.6), candor to the tribunal (RPC 3.3), supervision of subordinates (RPC 5.1 and 5.3), communication with clients (RPC 1.4), and billing integrity (RPC 1.5). He also addressed the growing landscape of judicial standing orders and local rules requiring AI certification of court filings, and what attorneys should assume a judge expects even in the absence of a standing order.
Second, the panel explored the practical benefits of AI in litigation: accelerating initial case assessment, issue-spotting in novel or unfamiliar legal areas, reducing research and drafting time, and even promoting civility by running communications to opposing counsel through AI before sending. The panel also discussed AI’s role in expanding access to justice for pro se litigants.
Third, the panel addressed risks—most notably the release of trade secrets or confidential information, the dangers of hallucinated citations (as illustrated by the sanctions in Mata v. Avianca), and the fundamental failure that underlies most AI-related disciplinary issues: attorneys not checking the work. The panel emphasized that AI output should be treated as a draft, never as filing-ready work, and that practitioners should use only approved platforms with appropriate safeguards for confidential information.
Privilege, Work Product, and AI: The Emerging Case Law
Jamison addressed emerging case law that addresses whether a client’s or attorney’s use of AI waives attorney-client privilege or work product protection. Three recent federal decisions provide guidance as to the effect of AI use on waiving privileges.
In United States v. Heppner, 820 F. Supp. 3d 292 (S.D.N.Y. Feb. 17, 2026), the defendant used a free AI platform to outline defense strategy after he already had counsel—but without his attorney’s direction or involvement. The court held that communications with the AI platform could not qualify as communications between a client and his or her attorney because the privilege doctrine presumes a human relationship. No attorney-client relationship existed between the defendant and the platform. Critically, the court also rejected a work product argument, finding that documents not prepared at counsel’s behest and not reflecting counsel’s strategy fall outside the doctrine. The court further noted that the platform’s privacy policy disclosed that it collects user inputs and outputs, uses data to train its models, and may disclose data to third parties—including government authorities.
By contrast, in Warner v. Gilbarco, Inc., 820 F. Supp. 3d 292 (E.D. Mich. Feb. 10, 2026), the court held that materials created using AI can qualify for work product protection under FRCP 26(b)(3). The court reasoned that AI systems are tools, not persons, and that disclosing information to an AI tool is not disclosure to an adversary and therefore does not trigger a waiver.
Similarly, in Morgan v. V2X, Inc., 2026 WL 864223 (D. Colo. Mar. 30, 2026), the court confirmed that AI-assisted materials can qualify for work product protection and emphasized that applying this protection to pro se litigants is especially important in the AI context, because AI is one of the most powerful knowledge tools ever to become available to the masses. The court distinguished Heppner on the grounds that it involved a criminal case and a client acting independently from counsel’s direction.
The throughline of this emerging case law appears to show that protected and unprotected AI use turns on whether the AI is used at the direction of, and in the course of, the attorney-client relationship. A client’s independent, unsupervised use of a public AI tool—as in Heppner—risks falling outside privilege and work product protections entirely. Use directed by counsel and integrated into the representation is far more likely to retain those protections. Further, use of an AI tool as a pro per party may provide work product protections under Warner and Morgan, but the risk of waiver is so great, it is highly advised that no party use AI to seek legal advise without first consulting counsel.
Jamison also addressed the trade secret implications of AI use. Both the California Uniform Trade Secrets Act (CUTSA) and the federal Defend Trade Secrets Act (DTSA) require “reasonable efforts” to maintain secrecy as a precondition of protection. Many courts do not confer trade secret status based purely on nondisclosure agreements. Under the reasoning of Heppner, there is a plausible argument that inputting trade secrets into public AI platforms may defeat the “reasonable efforts” requirement—regardless of what the platform’s terms of service say about confidentiality. Again, consult your counsel before inputting your valuable trade secrets into any AI platform.
The Key Practical Recommendations – Consult Counsel
The single most important takeaway from my portion of the program is that clients should refrain from using AI tools for legal work outside the advice and direction of their counsel. Independent, unsupervised client use of public AI tools can forfeit attorney-client privilege and work product protection and can jeopardize trade secret rights. Counsel should affirmatively advise clients of these risks and establish clear protocols for any AI-assisted work within the scope of the representation.
CLE Recording Availability
A recording of this CLE program will be available for purchase on the LACBA website once posted. Readers may access it at: