Ethics Corner - AI Ethics: Because Your First Rulebook Didn't Cover Robots

Advertisers in this issue:
Fox Law Corporation
Krycler Ervin Taubman & Kaminsky
Law Offices of Lawrence C. Noble
Lawyers' Mutual
Lewitt Hackman
One Legal
Scheideman Design
Valley Community Legal Foundation

Download this issue of Valley Lawyer.

Ethics Corner, by Parker Shaffie, LLP
By Adam J Telanoff and Joel Osman

The explosion of Generative Artificial Intelligence (GenAI) tools is often touted as the latest, greatest innovation in the practice of law. Unfortunately, this innovation also comes with a significant downside—GenAI can make it easier for lawyers to not only continue to make rudimentary mistakes, but, sadly, can amplify those mistakes to entirely new levels. With the help of GenAI, it is easier than ever to abrogate your obligations to your client and to the court, while also violating the Rules of Professional Conduct.

Much has been made in the press of lawyers in other jurisdictions being caught out in problematic reliance on GenAI. While California is often a leader in pushing legal frontiers, we have been slow getting a published appellate opinion that addresses the improper use of GenAI in legal briefs, something that has been addressed by appellate courts in several other jurisdictions. Finally, with the recent certification for publication of Noland v. Land of the Free, L.P. 2025 WL 2629868, California now has its own reported case in this issue.

The facts and arguments in Noland are, as the Court of Appeal noted, nothing special. What made the case worth publishing was the quality of the arguments presented by Appellant’s counsel. In the opening brief there were 23 citations, 21 of which were, in the words of the Court of Appeal, “fabrications.” Counsel’s reply brief was similarly replete with fabrications. The fabrications in both briefs were typical of GenAI “hallucinations”; quotes attributed to cases that did not contain the quotes, holdings attributed to cases that did not address even the same subject matter, and, the classic, citation to cases that did not even exist.

Appellant lost the appeal, and because of the AI hallucinations in its brief, was required to pay substantial sanctions to the Court of Appeal.  Appellant’s counsel was referred to the State Bar for possible disciplinary action. Appellant could have faced additional liability for Respondent’s attorney’s fees, except Respondent’s counsel did not discover the fabricated citations – the Court of Appeal itself  discovered Appellant’s fabricated citations. Thus, counsel for both sides failed to meet their ethical obligations to their clients and to the court, but the failure of Appellant was worse.

It is not just small firm and solo practitioners who fall victim to AI hallucinations. In a recent high-profile case against the Church of Scientology, Boies Schiller Flexner, an AM200 firm, filed an appellate brief that cited hallucinated GenAI citations. The responsible partner fell on his sword and has requested leave to file a new brief that cites actual case law. The Church of Scientology has asked for the appeal to be dismissed.^[1] There are other, similar cases across the country. There have been close to one hundred documented instances of AI hallucination in court filings.^[2] Our personal favorite is Kohls v. Ellison, from the District Court of Minnesota, Case Number 24-cv-3754, in which an expert declaration from Jeff Hancock, Professor of Communication at Stanford University and Director of the Stanford Social Media Lab, on the topic of AI Deepfakes and their associated dangers was stricken for, you guessed it, citation to GenAI generated hallucinations.

The underlying concepts implicated by inadvertent reliance on AI hallucinations are not new. Business & Professions Code §6068(b) requires attorneys to “maintain the respect due to the courts of justice and judicial officers.” Part of maintaining that respect includes only citing to actual, not hallucinated, authority. Historical cases of improper citation have generally been limited to citation to facts that are unsupported by the appellate record. GenAI makes citation to plausible, but false, case citations are becoming more common.

Rules of Professional Conduct already exist to address issues raised by AI hallucinations. RPC 1.1 requires that an attorney act with competence. A basic tenet of competence is that the attorney actually understands the law that they are using to support their arguments. It should be axiomatic that the law must first exist to be understood. And, as stated in Comment 1 to Rule 1.1, “The duties set forth in this rule include the duty to keep abreast of the changes in the law and its practice, including the benefits and risks associated with relevant technology.” Attempting to claim ignorance of GenAI hallucinations is, in and of itself, a violation of Rule 1.1.

Similarly, RPC 3.1(a) requires lawyers to ensure that their arguments are meritorious, and either are warranted under current law or under a reasonable extension of existing law. Fabrications that are not based on actual law violate Rule 3.1(a).

RPC 3.3 prohibits lawyers from knowingly making a false statement of fact or law to a tribunal. While the “knowingly” requirement may mitigate, using GenAI without verifying citations should be de facto knowingly making a false statement of law to a court, given that GenAI may fabricate a response 30-50% of the time. [Noland p.23].

Using GenAI without verification of the “citations” that it provides not only impacts the immediate matter in which it occurs, it also impacts the integrity of the legal profession as a whole. In this context, inappropriate reliance on the unverified product of GenAI may warrant a finding of a violation of RPC 8.4(c), which classifies dishonest, fraudulent, deceitful, or misrepresentation (intentional or reckless) as professional misconduct. Citing to cases that do not exist is, at the least, reckless.

Avoiding the trap of GenAI hallucinations is not difficult. The State Bar Standing Committee On Professional Responsibility And Conduct has issued Practical Guidance For The Use Of Generative Artificial Intelligence In The Practice Of Law.^[3] These guidelines include additional reference to the RPC, in particular RPC 5.1, 5.2, and 5.3 which govern review of work performed by subordinates. GenAI is analogous to a precocious new attorney or law clerk; they may produce good arguments, but their work needs to be reviewed by a senior attorney.

This all makes us reminisce about our law school legal writing classes in which the single most important message was the admonition to verify every citation in every document that you file – and in every document filed by the opposition. That one simple rule avoids the quicksand of relying on GenAI with the added benefit of supporting better quality arguments.

About the Authors:

Adam J. Telanoff is a Senior Counsel to Parker Shaffie LLP. He concentrated his practice on litigation with a particular focus on professional responsibility matters including legal malpractice.

Joel A. Osman is a Partner in and General Counsel to Parker Shaffie LLP. He concentrates his practice on professional responsibility matters including legal malpractice, ethics counseling, serving as an expert witness and representing attorneys in response to State Bar complaint investigations.