# Utah Lawyer, Fabricated Citations, and the "The Clerk Used ChatGPT" Defense Short answer: **It will not be a clean shield.** It may reduce the severity of the sanction, but it is very unlikely to eliminate discipline entirely. Here's why, walking through the Utah rules that matter. --- ## The Core Problem: Rule 5.3 (Nonlawyer Assistants) Utah RPC 5.3 is the rule that directly governs this scenario: - **5.3(a):** The lawyer must *exercise reasonable supervision* over the clerk's work and make *reasonable efforts to ensure* the clerk's conduct is compatible with the lawyer's professional obligations. - **5.3(b):** A supervisory lawyer must give *appropriate instruction and direction* concerning the subject matter. - **5.3(c):** The lawyer is **responsible** for the nonlawyer's conduct if the lawyer ordered it, ratified it, or *knew of it and failed to take reasonable remedial action*. The critical point: the duty in 5.3(a) is **continuous and structural**, not a one-time "don't use AI" memo. A bar investigator will ask: - Did the firm have any written policy on legal-research tools? - Did the lawyer review the filing *before* submission? - Was there a second-person check on citations? - Had the lawyer ever inquired *what tools* the clerk used? If the answer to those is "no," the "I didn't know" argument collapses into "I didn't exercise the supervision the Rules require," which is itself a Rule 5.3 violation. --- ## Rule 3.3 (Candor Toward the Tribunal) Utah RPC 3.3(a)(3) prohibits knowingly offering evidence the lawyer *knows* to be false, and 3.3(b) obligates the lawyer to make reasonable efforts to rectify a violation. Fabricated case citations are treated by courts and bar committees as a direct breach of candor. The "knowingly" language in 3.3 is the lawyer's best textual hook: *the lawyer didn't know the citations were fabricated.* But: - Rule 8.4(c) (dishonesty, fraud, deceit, misrepresentation) and Rule 1.3 (diligence) fill the gap. A lawyer who submits a brief without verifying citations, relying on a clerk's unverified research output, has a **negligence/recklessness** problem even without actual knowledge. - Bar committees have treated unverified, AI-generated citations as a *derivative* 3.3 violation because the lawyer "presented" the authority to the court. --- ## Rule 1.3 (Diligence) A lawyer must act with "reasonable diligence." In Utah's current post-2023 landscape—where multiple bar committees have flagged AI hallucinations as a known risk—submitting a filing full of unverified citations without a reasonable verification step is hard to defend as "reasonable diligence." The bar will point to the widely publicized Mata v. Avianca (S.D.N.Y. 2023), the Ninth Circuit's 2024 sanctions, and similar Utah cases as evidence that the risk was *known and foreseeable* in the legal community. --- ## What "Without My Knowledge" Can Actually Do It won't make the violation disappear, but it can **mitigate**: | Factor | Effect on Sanction | |---|---| | Prompt self-disclosure to the court and bar before the client or opposing party raised it | Strongly reduces severity; shows good faith | | Voluntary correction / motion to withdraw the bad citations before any ruling | Reduces "harm" element | | No prior disciplinary history | Argues against a public sanction | | The lawyer did in fact review the filing (just missed the fake cites) | Shifts the issue from "no supervision" to "reasonable-care lapse" | | The firm quickly adopted an AI-use policy, retrained staff, added a citation-verification step | Demonstrates remedial action under 5.3(c)(2) | | The lawyer cooperates fully with the disciplinary committee | Favorable discretionary factor | A realistic outcome in a *first-offense, good-faith* case might be: a **private reprimand**, a mandatory **ethics-continuing-education requirement** (often including an AI-in-practice module), a court **cost order**, and a **written warning on file**. A public reprimand or suspension is more likely if the lawyer did *not* review the filing at all, or if the fabricated citations actually affected the client's case (e.g., a summary-judgment motion where the opposing party relied on the fake authority). --- ## What the Lawyer Should *Not* Do - **Say "I had nothing to do with it."** Rule 5.3 makes the lawyer the accountable party for the clerk's work product. Disclaiming responsibility reads as a lack of candor in the disciplinary proceeding itself. - **Wait for the other side to catch the fake cites.** Voluntary correction and disclosure are the single most important mitigating step. - **Blame the clerk exclusively.** The committee will treat that as a Rule 5.3 violation in its own right (failure of supervision) and may view the blame-shifting as a credibility problem. --- ## Practical Takeaway The "the clerk used ChatGPT and I didn't know" explanation is a *reasonable factual account* that a committee will credit as truthful, but it does not map neatly onto a "no violation" defense because **the lawyer's own duties under Rules 5.3, 1.3, and arguably 3.3/8.4 are independent of the clerk's state of mind.** The question the committee will ask is not *"Did the lawyer know?"* but *"Did the lawyer exercise the supervision and verification that Utah's Rules require?"* A lawyer who can document a reasonable review process and who corrects the problem quickly will almost certainly avoid the harshest sanctions. A lawyer who can't will face discipline even with a fully honest "I didn't know" explanation. *This is a legal-ethics analysis, not legal advice. A Utah-licensed attorney facing an actual AI-citation inquiry should retain independent counsel experienced in Utah bar discipline (and, if the matter is in federal court, the relevant federal district's local rules on AI-assisted filings, several of which now require disclosure of AI use).*