Williams | Robinson | Wiggins

What Missouri clients should know about AI, their lawyer, and the rules that govern its use.

Artificial intelligence is now built into everyday legal research and drafting tools, and the question for most lawyers isn't whether they'll encounter it — it's whether they'll use it responsibly. Courts across the country, including in Missouri, have sanctioned lawyers for filing briefs containing case citations that an AI tool simply invented. Using AI does not change any of a lawyer's professional obligations — it can make the work faster, but it cannot exercise judgment, and it cannot carry the lawyer's duties. Here is what the rules actually require, and why it matters to your case.

The cautionary tales. The problem is not hypothetical. In the widely reported case of Mata v. Avianca, 678 F. Supp. 3d 443 (S.D.N.Y. 2023), a New York attorney used ChatGPT to research a brief, and the tool invented several court decisions out of whole cloth; the attorney and his firm were fined $5,000 for submitting them. Missouri has its own example. In Kruse v. Karlen, 692 S.W.3d 43 (Mo. Ct. App. 2024), a self-represented appellant filed a brief in which 22 of 24 case citations were fabricated — drafted by an online consultant using AI. The Missouri Court of Appeals, Eastern District, dismissed the appeal and imposed $10,000 in damages for a frivolous appeal under Rule 84.19. The court's holding is the part that should stick with anyone relying on AI-assisted work: citing nonexistent authority is a false statement to the court — a breach of the duty of candor — even when someone, or something, else created the error. Researchers who track this problem nationally have documented hundreds of court decisions worldwide involving AI-hallucinated content, with the volume accelerating rather than fading.

Where the rules come from. Missouri lawyers look primarily to two sources for guidance. Missouri Informal Opinion 2024-11, issued by the Office of Legal Ethics Counsel and the Advisory Committee of the Supreme Court of Missouri in April 2024, is non-binding but is issued by a body affiliated with the Supreme Court of Missouri, and the Missouri Rules of Professional Conduct it cites are binding. It organizes a lawyer's obligations around three pillars: understand the tool, protect confidentiality, and follow court rules. ABA Formal Opinion 512, issued by the ABA Standing Committee on Ethics and Professional Responsibility on July 29, 2024, is persuasive rather than binding in Missouri, but it is the most comprehensive framework available and covers competence, confidentiality, communication, candor, supervision, and fees. Because Missouri's Rules of Professional Conduct closely mirror the ABA Model Rules, Opinion 512 maps cleanly onto a Missouri lawyer's actual obligations.

What the rules require of your lawyer. Several existing Rules of Professional Conduct apply directly to AI use, without any need for a new "AI rule":

• Competence (Rule 4-1.1). A lawyer must have a reasonable understanding of what an AI tool can and cannot do. Nobody needs to be a computer scientist, but a lawyer does need to know a tool's limits before relying on it.
• Confidentiality (Rule 4-1.6). Before entering any client information into an AI tool, a lawyer must confirm the tool keeps that information private and does not use it to train the underlying model, and must obtain client consent where the situation calls for it.
• Candor (Rule 4-3.3). A lawyer must independently verify every citation, quote, and statement of law an AI tool produces before filing it with a court. Fabricated authority in a filing can mean sanctions — as Kruse v. Karlen shows.
• Communication (Rule 4-1.4). A lawyer should inform clients about AI use when the engagement calls for it or when the client's confidential information is involved.
• Supervision (Rules 4-5.1 and 4-5.3). A supervising attorney must treat AI output like work product from a brand-new law clerk — useful, but always checked — and must make sure staff follow the same rules.
• Reasonable fees (Rule 4-1.5). A client should be billed for time actually spent. If AI makes a task faster, the client benefits from the shorter time; a lawyer should not bill for the time spent learning to use an AI tool.

A real example of why verification matters. Hallucinated authority is most dangerous when it looks authoritative. In one documented exchange, an AI tool was asked to draft Missouri trust language giving a trustee power over a decedent's digital assets. It returned confident, properly formatted language citing a "Missouri Revised Uniform Fiduciary Access to Digital Assets Act" at "§§ 436.1100–436.1190, RSMo." That statute does not exist. Only a follow-up question — asking the tool to confirm the citation — surfaced the error; the tool then correctly identified Missouri's actual law, the Missouri Fiduciary Access to Digital Assets Act, §§ 472.400–472.490, RSMo (effective 2018). Nothing about the first, fabricated answer looked wrong on its face. The only reason the error was caught is that someone tried to verify it. Had it gone unchecked, a client could have signed a trust citing a statute that does not exist.

What this means for you as a client. None of this should discourage the responsible use of AI — used well, it can make legal work faster and more thorough. But it should shape what you expect from your lawyer. A responsible firm will have vetted the tools it uses for client work, will protect your confidential information from being used to train a public model, will independently verify every citation and legal proposition before it goes into a filing or a document you sign, will tell you when AI is involved in your matter if the circumstances call for it, and will bill you for the actual time your matter took — not for a lawyer's time learning a new tool. If you're ever uncertain, it is entirely fair to ask your attorney whether, and how, they use AI on your matter.

Summary. Artificial intelligence has not changed a lawyer's ethical obligations under the Missouri Rules of Professional Conduct — it has only raised the stakes of getting verification wrong. Confidence is not accuracy: every sanctioned filing in every reported case looked authoritative right up until someone checked it. The lawyers who avoid these problems are the ones who treat every AI-assisted draft as exactly that — a draft — and who verify the underlying authority themselves before they sign their name to it.

Did you know? The above information is presented by Williams | Robinson | Wiggins as a public service and to generally outline the law in a particular area. It is not provided and is not intended as legal advice tailored to you or to your unique situation. Every legal matter depends upon specific facts which an attorney hired by you must consider in forming legal opinions and advice.

Need more information? If you have questions about how our firm uses AI, or about any legal matter, you may contact us at(573) 341-2266 to obtain more detailed assistance.