Utah: AI Ethics Guidance for Law Firms

Bottom line for a 5-50 attorney Utah firm: No formal ethics opinion exists, but the operative framework is binding: ABA Formal Opinion 512 applied through Utah RPCs 1.1, 1.4, 1.5, 1.6, 3.3, 5.1, and 5.3, plus the Utah Artificial Intelligence Policy Act disclosure obligation for high-risk client interactions, plus the Garner v. Kadince sanctions standard for court filings. The signing attorney’s duty to verify AI-generated citations is non-delegable in Utah appellate courts.

Start with: a written AI use policy, a citation verification protocol tied to Garner v. Kadince, and a vendor due-diligence record for each AI tool in use.

Using ChatGPT in Our Practices: Ethical Considerations (Utah State Bar, 2023)

Citation: Utah State Bar, “Using ChatGPT in Our Practices: Ethical Considerations” (2023). Primary source PDF.

Status: Informal educational guidance. Not a formal ethics opinion and does not constitute a safe harbor. The article predates ABA Formal Opinion 512 (July 2024), the Utah Artificial Intelligence Policy Act, and the 2025 Kadince sanctions ruling.

What it states

• Client disclosure and consent (Utah RPCs 1.2, 1.4): attorneys should tell clients they intend to use AI in the representation; informed consent is recommended before use.
• Confidentiality (Utah RPC 1.6): share with AI only information you would share with others; omit client names and identifying information.
• Fees (Utah RPC 1.5): charge clients only for time actually spent using AI, not for the time the task would have taken to complete manually.
• Verification obligation (Utah RPCs 1.1, 3.3): treat AI-generated results like a draft from a law clerk; verify any propositions or citations not confirmed accurate. Asking ChatGPT to verify its own citations does not reliably catch hallucinations.

Other Utah State Bar publications

The Utah State Bar AI Standing Committee (Chair: Bennett Borden; Co-Chair: Nick Hafen) produces educational content but does not issue formal ethics opinions. Two publications stand out from November 2025 and January 2026: Hafen’s Five Ethical Principles video and article and Borden and Carrington’s AI Tools Utah Lawyers Can Use Today. In February 2026 the Bar also republished the ALPS AI coverage article. Keith Call’s “A Cowboy’s Guide to AI Ethics” in the Utah Bar Journal (July/August 2025) applies ABA Formal Opinion 512 to Utah RPC duties.

What federal courts in Utah require for AI use in filings

No district-wide AI standing order has been confirmed for the U.S. District Court for the District of Utah. Individual judges may have adopted AI certification requirements through chambers standing orders. Firms should review the assigned judge’s standing orders at the opening of each federal matter. The operative federal default is Fed. R. Civ. P. 11.

What Utah state courts require for AI use in filings

The Utah Supreme Court has not issued a statewide AI standing rule. The Kadince ruling (below) is the operative state-court precedent on AI-hallucinated citations.

Garner v. Kadince, 2025 UT App 80

Citation: Garner v. Kadince, 2025 UT App 80; No. 20250188-CA (Utah Court of Appeals, opinion filed 2025-05-29). Primary source via Justia. Utah’s first published appellate ruling on AI-hallucinated citations; binding precedent in Utah courts.

Attorney Richard Bednar signed and filed a petition for appeal drafted by a law clerk using ChatGPT. The brief cited cases that did not exist or were substantively misrepresented, including the fictitious “Royer v. Nelson.” Bednar acknowledged he had not reviewed the citations before signing. The court held that “attorneys must verify each source cited in their legal filings” and that the signing attorney’s duty to verify is non-delegable.

Sanctions: Bednar was ordered to pay respondents’ attorney fees for time spent responding to the petition and attending the OSC hearing. He had to refund all client fees for preparation of the petition within seven days. He also had to pay $1,000 to “and Justice for All” within fourteen days. Rules implicated (applicable, not expressly stated): Utah RPC 1.1, RPC 3.3(a)(1), and RPC 5.3.

Significance for Utah firms: filing AI-hallucinated citations is sanctionable in Utah appellate courts. Reliance on a law clerk’s AI use is not a defense. The “I didn’t know the clerk used AI” explanation is itself an RPC 5.3 supervision failure.

Utah Artificial Intelligence Policy Act

Citation: Utah S.B. 149, Artificial Intelligence Policy Act (signed 2024-03-13; effective 2024-05-01); amended by S.B. 226 and S.B. 332 (2025), both effective 2025-05-07. Primary source S.B. 149; primary source S.B. 226. Utah was the first US state to enact a law specifically regulating private-sector generative AI interactions.

What the 2025 amendments did: narrowed the disclosure obligation to “high-risk artificial intelligence interactions.” That term covers generative AI use involving two factors. First, collection of sensitive personal information. Second, significant decision-making in medical, mental health, legal, or financial contexts.

Application to Utah firms: consider a firm running a chatbot or AI assistant that both collects sensitive client information AND provides substantive legal guidance, initial consultations, or document generation. The firm must disclose AI involvement clearly and conspicuously at the outset and throughout the interaction. Internal-only AI use does not appear to trigger Act disclosure. RPC 1.4 communication duties still apply independently. Firms whose AI systems clearly disclose non-human identity throughout interactions are shielded from certain Division of Consumer Protection enforcement actions under the 2025 safe harbor.

Enforcement: Utah Division of Consumer Protection. Administrative fines run up to $2,500 per violation, with a separate cap of $5,000 per violation of an existing administrative or court order. UAIPA non-disclosure does not automatically constitute an RPC violation. A pattern could also be characterized as a failure of client communication under RPC 1.4.

Utah legal services regulatory sandbox

The Office of Legal Services Innovation (OLSI), housed within the Utah Supreme Court, was the first US jurisdiction to authorize AI-assisted and AI-driven legal services at the operational level. Phase 2 (effective 2024-09) narrowed the program substantially. As of 2025-04-30, 11 authorized entities remained active, down from 39 in 2022. The sandbox is relevant as a regulatory signal. A 5-50 attorney firm is not in the sandbox and is not subject to OLSI oversight.

What AI-related rules are pending in Utah?

The Utah Attorney General’s bipartisan AI Task Force (announced Nov 2025, co-led with NC AG Jeff Jackson; partners include OpenAI and Microsoft) signals ongoing state-level enforcement interest. It is not an attorney-regulation body. Utah firms should monitor Task Force outputs that could shape future consumer protection enforcement under UAIPA.

How do malpractice carriers in Utah treat AI use?

ALPS Insurance is the Utah State Bar’s endorsed legal malpractice carrier and the nation’s largest direct writer of lawyers’ malpractice insurance. ALPS has published AI coverage warnings, republished by the Utah Bar. Identified risks include four categories. First, standard policy intentional-act exclusions for blind AI reliance. Second, UPL exclusions for client-facing AI without attorney oversight. Third, confidentiality exclusions for data input into public or unsecured AI platforms. Fourth, emerging “absolute” AI exclusions appearing in some E&O and management liability lines (not yet standard in LPL products).

What does my Utah malpractice carrier ask about AI at renewal?

ALPS has not published the specific items it will request on the application; its Bar-republished risk advisory establishes the substantive standard. The framework that documentation should map to is ABA Formal Opinion 512 applied through Utah RPCs, plus the Kadince verification standard, plus UAIPA disclosure for any high-risk client interaction. At renewal, request written confirmation from the carrier that AI-assisted work product errors are covered under the professional services definition. Identify which policy (cyber vs. malpractice) responds to consumer-tier AI data exposure. Document the carrier’s response.

What documentation should a Utah firm keep on file?

Month one (foundational)

1. (Owner: managing partner + firm administrator) Written AI use policy (Utah RPCs 5.1, 5.3; ABA Op 512). Approved tools, prohibition on submitting unverified AI output to courts or clients, mandatory attorney review before any AI work product is filed or sent to a client, and which tools may receive client-identifying information. The Kadince ruling makes the supervision policy a litigation-risk document, not merely best practice.
2. (Owner: litigation lead) Citation verification protocol (Utah RPCs 1.1, 3.3; Kadince). Written procedure requiring every AI-generated citation be verified against a primary source before any filing or client deliverable, documenting tool used and verifying attorney. The signing attorney is the gatekeeper; delegation to a law clerk does not transfer the obligation.
3. (Owner: matter lead attorney) UAIPA high-risk interaction assessment (Utah S.B. 149 as amended). Identify any client-facing AI tool that collects sensitive client information AND provides substantive legal guidance; document the disclosure mechanism and confirm it meets the 2025 safe harbor.

Months two and three (operational documentation)

1. (Owner: firm administrator + outside IT) Vendor due diligence record (Utah RPC 1.6; ABA Op 512). For each AI tool: vendor training-on-input status, retention terms, and whether the firm has an enterprise agreement with data protection commitments. Consumer-tier ChatGPT and standard Gemini without enterprise agreements present RPC 1.6 risk.
2. (Owner: firm administrator) Attorney and staff training log (Utah RPCs 1.1, 5.1, 5.3). Date, attendees, content covering hallucination risk, RPC 3.3 candor, the Kadince holding, UAIPA disclosure triggers, and firm policy.
3. (Owner: managing partner + billing partner) Engagement letter AI provision (Utah RPCs 1.4, 1.5, 1.6). Disclose categories of firm AI use, data-handling commitments, and the firm’s billing approach for AI-assisted work consistent with the Bar’s 2023 reasonableness guidance.

Months four to six (per-matter discipline)

These are recurring practices, not one-time projects: each new matter exercises them.

1. (Owner: matter lead attorney) Judge-specific AI order check (Utah RPC 3.3; Fed. R. Civ. P. 11). At each new D. Utah matter, review the assigned judge’s standing orders and chambers practices for AI certification requirements; document the check.
2. (Owner: managing partner) Periodic review of Utah Bar AI publications (Utah RPC 1.1). Monitor the AI Standing Committee feed at utahbar.org/tag/ai/ and any updates to UAIPA enforcement posture.

We are building practitioner resources for firms working through this list. The monthly update covers new resources as they ship.