AI misuse reported in over 2000 cases worldwide. What's a lawyer to do? (Part 2)

In August 2026, the Los Angeles Times reported on litigation involving State Farm Insurance Company concerning insurance coverage for a house damaged by fire in Carson. State Farm's motions in limine for an October 2026 trial date were riddled with fabricated cases. The insurance company's lawyer fell on her sword and apologized profusely to LA Superior Court Judge Elizabeth Bradley.

The case? Fa'alagilagi Meni-Siliga v. A's Contractor, Inc, 24-CMCV01105 (Cal. Super. Ct. L.A. County). Yesterday, we shared three tips to prevent this type of mortifying error by counsel. Here are tips four through seven.

4.                  Verify cases and double-check all legal work

New technology does not obviate a lawyer's duty to verify the contents of a pleading. Always check your work. Asking AI to check your cases is not verification. Some AI platforms ingest user prompts which can then be transformed into faulty AI legal propositions. In the same vein, some AI models have used court orders concerning a lawyer's use of false or fabricated cases to report, in response to a verification prompt, that a fake case cited in the court order actually exists. For example, in Mata v. Aviance, Inc. (S.D.N.Y. 2023), the court singled out and listed six fabricated cases cited to by counsel who had used ChatGPT. Subsequent user prompts in AI made it appear that the fabricated cases existed. They did "exist" in an order; however, they were not real cases, and AI did not distinguish between them.

Verification means verifying that the case is real and stands for the proposition cited.

5.                  Have a written firm policy on use of AI

Law firms and other institutions have written policies for a variety of topics on which to inform and educate staff and for use to shield the entity in litigation regarding the topic. For example, institutions often have employee manuals with provisions on drug use at work, sexual harassment and/or discrimination.

As a general rule, where a company has a written policy on a topic and the dictates are enforced throughout the institution--from the C-suite and management to the workers on the loading dock--the institution is better positioned to argue that errant behavior by an employee should not be attributed to the company under a vicarious liability theory.

Legal malpractice carriers often exclude coverage for court-imposed sanctions, penalties and attorney's fees. Some are now warning lawyers about exclusions of malpractice coverage for AI hallucinations. Some are cautioning lawyers who use AI to reduce liability and policy fees by requiring attorney review of pleading contents, implementing citation-checking protocols, using two-case research tool checks for all cites and banning blind reliance on AI-generated case law.

6.                  Ensure that any AI used by the firm protects client confidentiality

AI platforms typically use LLMs (Large Language Models) that train on material that is submitted to the platform. When lawyers use such platforms, they can easily violate their duty of confidentiality to the client if the information submitted comes from the client. When lawyers use AI, they must ensure that the terms of service provide confidentiality. Lawyers must also evaluate their AI vendors for expertise in the use of security measures used to protect client data and confidentiality. The American Bar Association, in Formal Opinion 512, advises law firms to read and understand the terms of use, privacy policy, and related contractual terms and policies of any AI tool they use. This is so the lawyer can understand who will have access to the information that the lawyer inputs into the platform. Thomas Reuters, in an article entitled "Building Trust in AI to Keep Firm and Client Data Safe," cautions law firms to avoid the use of AI purveyors which have "vague security answers, unclear data deletion policies, or reluctance on third-party sharing." See also, "Evaluating AI Solutions for Legal Professionals."

According to Professor Lynn Dickinson, author of two books on the use of AI in education, "there is a difference between LLMs that are free or 'public facing' and closed pay-for-fee versions (which are referred to in the IT industry as 'enterprise solutions'). The former typically have no confidentiality restraints, the latter represent that they have and implement confidentiality protocols." As a result, it is advisable for law firms and lawyers to purchase and use enterprise solutions specifically programmed for their firm to keep data, including prompts and responses, private. The terms of service would necessarily include provisions such as (i) no LLM training with submitted material, (ii) no sharing with third parties and/or (iii) timed retention/deletion of materials.

It is the lawyer's duty to ensure that the platform does not train with or store client data. LexisNexis, Westlaw and Harvey, among others, represent that they do not train with or store client data. Do your research and read the AI vendor service agreements before you use AI on any legal matter.

7.                  Address AI use in lawyer and law firm fee and retainer agreements

While the ABA's Formal Opinion 512 warns that use of generative AI tools which train on inputted data implicate confidentiality concerns, it has not adopted language like California's SB 574, which, if signed into law by Gov. Newsom, would mandate disclosure of AI use to courts regarding all submitted documents. While California is on the verge of having legislation which requires disclosure to courts, neither California nor the ABA have specifically directed that lawyers must inform their clients about the use of AI.

But since ABA Model Rule 1.6 directs lawyers to "not reveal information relating to the representation of a client--unless the client gives informed consent," it can easily be argued that clients should be informed about the disclosure of their case information into AI platforms. In fact, the ABA issued a guide for lawyers on April 7, 2026, urging--but not requiring--lawyers to be transparent with clients about AI use. See, ABA's "When Should Lawyers Disclose AI Use? A Practical Compliance Guide." The retainer agreement is the place where the client can be asked to sign off on the affirmative use of AI by counsel.

Many of the court and bar rules on AI concern--reflected in LexisNexis' Generative AI Tracker-- not just on disclosure, but the issue of how a client should be billed for work that now takes a fraction of the time. All are clear that there should be truth-in-billing and that it is improper to bill for the time it would or could have taken rather than the actual time it did take for the task.

Here is an example of one of several lawyer-to-client AI disclosures propounded by the State Bar of Texas:

·         The firm may use AI tools to assist with tasks such as legal research, document review and drafting. These tools are used under attorney supervision to improve efficiency and quality. The firm will not disclose identifiable client information or input it into any AI system that stores or uses data for future training, unless the system is deployed in a private or secure environment. All AI use will comply with applicable ethical rules, confidentiality obligations and data protection requirements.

The Texas AI fee agreement language could be supplemented with the addition of the following:

·         Additionally, it is intended that AI platform ____ (fill-in-the-blank) will be used in the client's case. The client will be billed for the actual time in which AI was used in the case and for attorney verification and review of AI sources.

Follow the adage: "You make it; you break it; you pay for it."

AI can easily be misused in the practice of law. Misuse can result not just in sanctions or lack of malpractice coverage but also:

        i.            Lawyers can be removed as counsel, as happened in federal district court in Alabama, see Johnson v. Dunn, 792 F. Supp. 3d 1241 (N.D. Ala. 2025);

       ii.            The case itself can be dismissed, as happened in a federal health care qui tam case in Utah, in U.S. ex rel. Khoury v. Intermountain Healthcare Inc et al., where a fabricated expert report resulted in dismissal of the case; and/or

     iii.            The lawyer or law firm can be referred to state bar committees for disciplinary proceedings.

Bonnie Shucha, an expert on ethical use of AI in law, who serves as an associate dean and law library director at the University of Wisconsin Law School, gives advice to law students that attorneys should take to heart. Shucha urges those in the legal profession to "understand what these tools can and cannot do, what the risks are and to understand one's ethical obligations in the use of AI." Shucha states that "if you put your name on something, you own it completely." The LA State Farm lawyer recognized this adage. In her declaration about AI abuse filed on Aug. 11, 2026, at page 6, paragraph 13, the lawyer stated, "I accept responsibility for those inexcusable errors, and offer no excuse, only an apology."

On Aug. 28, 2026, Judge Bradley issued an OSC as to why State Farm's lawyer should not be required to pay $5,000 in sanctions to the court for the lawyer's lack of diligence. The hearing was held last Friday. The court issued sanctions in the amount of $999.99. The State Farm lawyer is to pay that amount to the Los Angeles Superior Court by Oct. 11, 2026.

Let the now-sanctioned lawyer's experience inform your own use of AI. To recap: Be transparent with your client and potential clients about AI usage and risks, including mentioning its use in the retainer agreement); know its limitations; have in-office training on AI use and have office-wide AI policies; stay abreast of AI bar rules and laws; and protect client confidentiality in all matters, including the use of AI. Also, vigorously verify all content in court submissions.

AI is now part of the practice of law. It is your duty is to use it in a healthy fashion. Think of it this way: healthy AI use is akin to consumption of a salad with triple-washed lettuce. Such a dish can contribute to a healthy lifestyle if contaminants in the leaves are removed. If not, there's a significant risk that you will, at a minimum, endure an intestinal bug, or worse yet, some dire medical consequences. If you do not triple-check your AI-generated work, you could be jeopardizing your professional health and suffer a range of career-damaging repercussions.

Stay healthy. Keep your law license. Your clients are counting on you.

Postscript: No AI was used in the creation of this article, either today's portion or yesterday's. All propositions and cites have been verified by each of the coauthors.