As of October 26, 2026, the three states govern lawyers' use of AI in three different ways. California has the most: a new statute signed in 2026 that requires lawyers to disclose generative AI use in court filings starting next year, State Bar guidance, proposed rule comments, and a published sanctions opinion. Utah has no formal ethics opinion, but it has a published Court of Appeals sanctions decision and informal Bar guidance. Nevada has no formal opinion and no statewide court rule, but its Rules of Professional Conduct apply as written, and at least one federal judge in Nevada requires an AI certification on filings.
In all three, the duty that gets lawyers sanctioned is the same: you must read and verify every authority you cite. The details, state by state, follow. Each instrument is linked to its source; this is a summary, not legal advice, and the area is moving fast.
Utah
| Instrument | Binding? | What it requires | Source |
|---|---|---|---|
| Utah Rules of Professional Conduct (Rules 1.1, 1.4, 1.5, 1.6, 3.3, 5.1, 5.3) | Yes | Competence, client communication, reasonable fees, confidentiality, candor and supervision apply to AI use as to any other tool | Utah State Bar ethics opinions index: no AI opinion as of Op. 2025-02 (Aug. 13, 2025) |
| Garner v. Kadince, Inc., 2025 UT App 80 (May 22, 2025) | Yes, published precedent | Counsel who signed a petition citing a case ChatGPT invented "fell short of their gatekeeping responsibilities." Sanctions under Utah Rule of Appellate Procedure 40: the other side's fees, a refund to the client, and a $1,000 payment by the signing attorney to a legal-aid nonprofit | Opinion (PDF), Utah courts |
| "Using ChatGPT in Our Practices: Ethical Considerations" (May 2023) | No, informal guidance posted by the Bar | Tell the client and let them decide; strip identifying information; bill only time actually spent; treat output like a law clerk's draft | Utah State Bar (PDF) |
| Utah Code Title 13, Chapter 77 (generative AI consumer disclosures) | Yes, but its professional-disclosure duty is drawn narrowly | A supplier must disclose AI to a consumer who clearly asks. The separate duty for "regulated occupations" is defined by reference to occupations regulated by the Department of Commerce; lawyers are regulated by the Utah Supreme Court, so on the statute's text that duty likely does not reach them. A firm's client-facing chatbot is the place to look twice | § 13-77-103 · § 13-77-101 (definitions) |
What it means for a Utah lawyer. There is no Utah opinion to cite for permission, and none is needed: the rules you already practice under govern. Garner is the case Utah judges will reach for, and it puts the duty on the lawyer who signs, not on whoever drafted. Our deeper guide is AI ethics for Utah lawyers.
California
| Instrument | Binding? | What it requires | Source |
|---|---|---|---|
| SB 574 (2026) | Yes, statute; the disclosure requirement starts in 2027 | Per the Judicial Council, "California attorneys will be required to disclose when they use generative artificial intelligence" in documents submitted to courts. Read the enrolled text for the exact duties before you rely on a summary | California courts newsroom (Oct. 1, 2026) · Bill text |
| Noland v. Land of the Free, L.P., No. B331918 (Cal. Ct. App. Sept. 12, 2025) | Yes, published precedent | No brief should contain a citation the responsible attorney has not personally read and verified. $10,000 sanction, referral to the State Bar, and the opinion served on the client | Opinion (PDF), California courts |
| Practical Guidance for the Use of Generative AI (2026 version, approved May 14, 2026, replacing the 2023 guidance) | No, advisory | Do not input confidential information without informed client consent and adequate safeguards; competence and independent judgment over inputs and outputs; bill actual time spent; consider telling the client; never let AI file on its own | State Bar of California (PDF) |
| Proposed AI comments to Rules 1.1, 1.4, 1.6, 3.3, 5.1 and 5.3 | Not yet: proposed, after two public comment periods | Among them: under Rule 3.3, check that no cited authority is fabricated, misstated or taken out of context; under Rules 5.1 and 5.3, firm AI policies and supervision of staff AI use | State Bar public comment page |
| California Rules of Court, rule 10.430, and Standard 10.80 (effective Sept. 1, 2025) | Binding on courts (rule 10.430); advisory for judicial officers (Standard 10.80) | Governs courts' and judges' own AI use, not lawyers' | Rule 10.430 · Standard 10.80 |
What it means for a California lawyer. California is the only one of the three moving AI duties into binding text, and the disclosure requirement is the change to plan for now: put a disclosure step into your filing checklist before it takes effect. The verification standard is already binding through Noland. Our deeper guide is California AI ethics rules for lawyers.
Nevada
| Instrument | Binding? | What it requires | Source |
|---|---|---|---|
| Nevada Rules of Professional Conduct | Yes | Rules 1.1, 1.4, 1.5, 1.6(c), 3.3, 5.1 and 5.3 apply as written. Nevada's Rule 1.1 has no technology comment, but Rule 1.0A says the ABA's comments may be consulted for guidance | Nevada Rules of Professional Conduct (Nevada Legislature) |
| State Bar of Nevada formal ethics opinions | n/a | No AI opinion has issued; the newest is Formal Opinion 61 (Feb. 18, 2025) | State Bar of Nevada ethics opinions |
| "Navigating AI with Competence and Care," President's Message, Nevada Lawyer (Sept. 2026) | No; the article says it is not formal ethics guidance | Maps AI practices to the rules: understand the tool (1.1); verify before filing (1.1, 3.3); keep confidential information out of tools whose terms you have not read (1.6); supervise and set firm policy (5.1, 5.3); check AI-assisted billing (1.5) | Nevada Lawyer (PDF) |
| Second Amended Standing Order, Magistrate Judge Couvillier, D. Nev. (effective Aug. 17, 2026) | Yes, in his civil cases | A filing with AI-generated content must carry a certification above the signature; omitting it may draw sanctions, and so may failing to verify AI-generated citations | Standing order (PDF), D. Nev. |
| "Artificial Intelligence: A Guide for Judicial Officers" | Advisory, for judges | Verify output, protect confidentiality, watch for fabricated evidence | Nevada courts (PDF) |
What it means for a Nevada lawyer. With no Nevada opinion, ABA Formal Opinion 512 is the persuasive authority Nevada's own rules point you toward (our ABA 512 explainer walks through it). In federal court, read each judge's standing orders before you file: certification requirements are judge-specific, and the District of Nevada's local rules have no general AI provision.
What all three states require
Strip away the differences in form and the same five duties run through every instrument above:
- Verify every citation before you sign. Garner, Noland and the D. Nev. standing order all put it on the signing lawyer. This is the duty behind every order on our running list of AI hallucination sanctions cases.
- Protect client confidences. Know where a tool sends your input and whether its terms let the provider keep or train on it, before client facts go in.
- Supervise. Associates, paralegals and contract lawyers who use AI are your responsibility under Rules 5.1 and 5.3; a written firm AI policy is the practical way to discharge it.
- Bill honestly. Charge for time actually spent, not for the hours the tool saved.
- Communicate. Consider telling clients how you use AI, and in California plan to tell the court. Our guide to whether lawyers have to disclose AI use to courts covers the wider picture, and the state-by-state AI guidance map covers every other state.
The first duty is the one you can make routine. Our citation checker for briefs explains what CaseRead's free check does with a draft: each citation looked up, then read against the sentence it supports. It does not replace reading the authority, but it tells you which ones to read first.
Frequently asked questions
Does Utah have an AI ethics opinion for lawyers? No. As of October 2026 the Utah State Bar's Ethics Advisory Opinion Committee has issued no opinion on generative AI; the newest opinion is 2025-02. The Bar has posted informal guidance from 2023, and the Utah Court of Appeals sanctioned counsel in Garner v. Kadince, 2025 UT App 80, for a petition citing a case ChatGPT invented.
Do California lawyers have to disclose AI use to the court? Starting in 2027, yes. SB 574, signed in 2026, requires California attorneys to disclose when they use generative AI in documents submitted to courts, according to the California courts newsroom. Read the enrolled bill text for the exact scope. Until then, check each court's local rules and each judge's standing orders.
Does Nevada have rules on lawyers using AI? Nevada has no AI ethics opinion and no statewide court rule on AI as of October 2026. The Nevada Rules of Professional Conduct apply as written, including competence, confidentiality, candor and supervision, and some federal judges in the District of Nevada require a certification on filings that contain AI-generated content.
Can a lawyer be sanctioned for AI-generated citations in these states? Yes. The Utah Court of Appeals did so in Garner v. Kadince (2025), and the California Court of Appeal imposed a $10,000 sanction and referred counsel to the State Bar in Noland v. Land of the Free (2025). In both, the court held the signing attorney responsible for citations no one verified.
Is the California State Bar's AI guidance binding? No. The State Bar's Practical Guidance for the Use of Generative AI, updated in 2026, is advisory. Proposed AI comments to six Rules of Professional Conduct have gone through public comment but had not taken effect as of October 2026. SB 574 is the binding instrument, and the existing Rules of Professional Conduct apply to AI use now.