By Nicholas Bartels
“Always Appealing” is a column addressing current issues in appellate practice and recent appellate cases written by the lawyers of Smith Goodfriend, PS, a Seattle law firm that limits its practice to civil appeals and related trial court motions practice.
Regular readers of Always Appealing know that last August, I was asked by Catherine Smith and Valerie Villacin to help them do some cursory research on the applicability of generative artificial intelligence (AI) in practice. That “cursory research” quickly turned into a deep dive, which resulted in an article in the September 2025 edition of the Bar Bulletin[1] and a presentation at the KCBA’s 2025 Family Law Hot Topics. For those who have not read the article or did not attend Hot Topics, our takeaway was that while AI may be helpful, it is not a replacement for a competent human. Although we are all still of that opinion, I write now to bring you some updates on the use of AI and offer what I believe to be the silver bullet to avoid any misadventures in AI.
When we initially reported on our (mis)adventures in AI, there was not yet any formal guidance from a Washington court or the Washington State Bar Association on the use of AI.[2] However, the WSBA recently released a formal ethics opinion discussing AI, the Court of Appeals has issued a handful of opinions warning several litigants against using AI, and at least two superior courts have promulgated local rules dealing specifically with AI in pleadings.
In December 2025, the WSBA’s Committee on Professional Ethics released Advisory Opinion 2025-05, which addressed “the ethical implications of using emerging artificial intelligence-enabled tools in law practice.”[3] The Opinion uses several illustrations to discuss the ethical duties of competence, diligence, confidentiality, communication, candor, supervision, and reasonable billing implicated by the use of AI. The conclusion of the opinion is that “[a]though [AI tools] can assist lawyers in delivering legal services, they do not relieve lawyers of the[ir] core duties.”[4]
This conclusion is perhaps most clearly illustrated in the case of “Amanda, a junior associate at a mid-sized law firm” who, when “[f]eeling overwhelmed by the volume of documents and tight deadlines, [decided] to use a newly released AI legal research tool to assist with her work.”[5] In her attempt to lighten her load, Amanda inputs sensitive client information and case details into the AI program without investigating the AI program’s capabilities and limitations, blindly relies on the AI program for research without independently verifying the accuracy or relevance of the cited cases, and uses the unverified research to draft a “crucial motion.”[6] In doing so, Amanda’s actions implicate both her duties of competence and diligence.[7] The Opinion warns that “[w]hile AI can enhance efficiency, it cannot replace the critical thinking, judgment, and ethical obligations of a competent and diligent attorney” and goes on to state that “lawyers cannot cede either their professional judgment or their responsibility for work to AI tools.”[8]
The Court of Appeals has also addressed the use of AI in several unpublished opinions. In the only case discussing AI use by an attorney, Division One in Put a Bird on It, LLC v. Seattle Arena Holdings, LLC chastised appellant’s counsel for their use of “AI-based search engines,” stating that counsel’s use of erroneous citations that were generated by AI “falls below our expectations of counsel, particularly since it is now well known in the legal community that AI resources can generate erroneous citations and false quotations.”[9] The Court did not issue sanctions because counsel had “taken steps to remediate the situation, . . . apologized to the court, and . . . represented that they are implementing additional quality-control protocols (as all lawyers should) to ensure this does not happen again.”[10]
Divisions One and Three have also each issued two opinions addressing a pro se litigant’s improper use of AI which resulted in fake citations.[11] In Richmond (Division Three), West, and Esterra Commons Venture (both Division One), the Court recognized that the various citations in the pro se parties’ briefing were likely the result of AI, and that any litigant, pro se or represented, had a duty to ensure that the cited authority was real and that it supported the proposition for which it was cited.[12] The Court, however, did not issue sanctions against the pro se litigants in those cases. On the other hand, in Carey, Division Three declined to sanction the pro se litigant “for his use of fictitious case law and failure to comply with the RAPs,” but did order sanctions under RAP 18.9(a) for filing a frivolous appeal, and required the pro se litigant to pay fees and costs.[13]
Superior courts have also started adapting to the existence of AI. For those who practice in Lincoln County (immediately west of Spokane County), the Lincoln County Local Civil Rules require an AI “disclosure footnote” for “[a]ny legal brief or memorandum submitted to the Lincoln County Superior Court that has been materially drafted or supplemented using artificial intelligence (AI) tools.”[14] If only a portion of the pleading was drafted using AI, then the disclosure must specifically state in a footnote which portion of the pleading was developed using AI.[15] If the entire pleading was developed using AI, then the rule states that “a footnote to that effect shall appear on the first page of the document.”[16]
Kitsap County has taken a different approach. If you practice on the Kitsap Peninsula, then you are subject to a modified version of CR 11, which:
1) serves as a warning to litigants that AI “may produce factually or legally inaccurate content and should never replace the lawyer’s independent legal judgment”;
2) states that the filing of a pleading that “contains references to caselaw or statutes that do not exist shall be subject to sanctions”; and
3) sets a “presumptive sanction of $2,000 per violation” for attorneys who violate the rule which “shall be entered as a judgment against the attorney personally and not their client.”[17]
As the recent developments in Washington and the growing list of sanctions cases demonstrate,[18] AI can be a significant trap for lawyers. However, there is a simple trick — a proverbial silver bullet[19] — for lawyers to disentangle themselves from the trap that is AI: check your citations to ensure that the cited material supports your stated assertion.
This may seem like a “no-brainer” as CR 11 requires that an “attorney has read the pleading, motion, or legal memorandum,” and is making arguments that are “grounded in fact” and “warranted by existing law.”[20] Further, RAP 10.3 explicitly requires an appellate brief to have a “[r]eference to the record . . . for each factual statement” and an argument containing “citations to legal authority and references to relevant parts of the record.”[21] But each of these recent examples suggests that some consider AI a substitute for cite checking.
To be clear, there is nothing inherently wrong with using AI in law practice, but blindly relying on AI creates ethical problems.[22] Cite checking every reference in a document, be it a pleading, letter, or anything else produced by a lawyer, is the only surefire way to avoid unethical use of AI. However, despite the growing weight of authority, attorneys continue to misuse AI by blindly relying on it and not cite checking their work product.[23] I’ll be the first to admit that cite checking is a time- and labor-intensive, fastidious task, but it nevertheless remains imperative whether a lawyer is using AI or not.
Nicholas Bartels is an associate at Smith Goodfriend and a graduate of the Seattle University School of Law. While at Seattle U he was a lead article editor for the Seattle University Law Review. Before being admitted to the bar, Nicholas worked as a law clerk at Smith Goodfriend. He can be reached at nicholas@washingtonappeals.com.
[1] Catherine Smith, Valerie Villacin, and Nicholas Bartels, Always Appealing: (Mis)Adventures in Appellate AI, King County Bar Association (Sept. 1, 2025).
[2] The ABA had issued Formal Opinion 512 on Generative Artificial Intelligence Tools. See ABA Comm. on Ethics and Pro. Resp., Formal Op. 512 (2024).
[3] WSBA Comm. on Pro. Ethics, Advisory Op. 2025-05, at 1. (2025); see also Sandra Schilling, Navigating the Ethical Use of AI in Law Practice, NW Sidebar (Jan. 20, 2026), (summarizing the “key themes” of the Opinion). Interestingly the WSBA appears to be leading by example as the NW Sidebar article contains a note that “[s]ome material in this publication was generated using Microsoft Copilot”—Microsoft’s AI platform—“and was reviewed for accuracy by the author, a member of the WSBA Advancement department, before publication.”
[4] Advisory Op. 2025-05, at 12.
[5] Advisory Op. 2025-05, at 4.
[6] Advisory Op. 2025-05, at 4.
[7] Advisory Op. 2025-05, at 4-5.
[8] Advisory Op. 2025-05, at 5.
[9] Put a Bird on It, LLC v. Seattle Arena Holdings, LLC, No. 87756-9-I, 2026 WL 579399, at *10, n.2 (Mar. 2, 2026).
[10] Put a Bird on It, 2026 WL 579399, at *10, n.2.
[11] Estate of Carey, No. 40344-1-III, 2026 WL 586674 (Mar. 3, 2026); Esterra Commons Venture, LLC v. Norton, No. 87320-2-I, 2026 WL 734359 (Mar. 16, 2026); Richmond v. City of Newport, No. 40954-6-III, 2026 WL 746700 (Mar. 17, 2026); West v. Lower Duwamish Waterway Grp., No. 87289-3-I, 2026 WL 1069752 (Apr. 20, 2026).
[12] Richmond, 2026 WL 746700, at *3, n.3; West, 2026 WL 1069752, at *6, n.13; Esterra Commons Venture, 2026 WL 734359, at *1, n.1.
[13] Carey, 2026 WL 586674, at *6.
[14] Lincoln County Local Civil Rule 7(b)(6)(A).
[15] Lincoln County Local Civil Rule 7(b)(6)(A)(i).
[16] Lincoln County Local Civil Rule 7(b)(6)(A)(i).
[17] Kitsap County Local Civil Rule 11(a). The local rule was originally adopted on an emergency basis, effective September 1, 2025, and has been renewed on an emergency basis multiple times, most recently on June 1, 2026. See In re Local Rules of the Superior Court of Washington for Kitsap County, Emergency Order Amending Local Rules (June 1, 2026), (available at )
[18] The law firm Ropes & Gray and the researcher Damien Charlotin have interactive trackers of recent decisions and orders dealing with the use of AI in courts. Ropes & Gray, Standing Orders, Local Rules, and Decisions on the Use of AI, (last visited June 10, 2026); Damien Charlotin, AI Hallucination Cases, (last visited June 20, 2026).
[19] The use of em dashes has become a litmus test for whether something was produced by AI, but despite AI’s affinity for em dashes, according to one commentator, “[i]f you dig into actual research, it’s clear em dashes aren’t a real giveaway.” Brent Csutoras, The Em Dash Dilemma: How a Punctuation Mark Became AI’s Stubborn Signature, Medium, (Apr. 29, 2025), . Unlike the WSBA, I did not use AI in drafting this article as my use of em dashes may suggest.
[20] CR 11(a).
[21] RAP 10.3(a)(5)-(6).
[22] Malkeet Lnu v. Blanche, __ F.4th __, 2026 WL 1587554, at *6 (9th Cir. June 3, 2026) (“While not inherently unethical or irresponsible, using generative AI without rigorously checking its output does present a higher risk of violating certain ethical and procedural rules.”).
[23] Indeed, the seminal case dealing with AI in pleadings involved a lawyer who did “not read[] a single case cited” in his pleading and who took “no other steps on his own to check whether any aspect of the assertions of law were warranted by existing law.” Mata v. Avianca, Inc., 678 F. Supp. 3d 443, 464 (S.D.N.Y. 2023).
