Generative AI and the Practice of Law- What Are My Ethical Duties? An On-Going Discussion

April 2025

By Dianne Jackson McLean, Esq.

As the terms AI or Artificial Intelligence (AI) and more notably Generative Artificial Intelligence (GAI) are being used frequently in today’s world, I question its meaning in the context of the practice of law, which typically involves human beings analyzing case law, statutes and procedures; and applying such knowledge to the facts in client’s matter to provide competent representation to the client.  There is much that is being written about AI and the ethical standards that should apply[1].   However, at present, there are  no settled standards, including in the practice of law, for its use.  Lawyers continue to be bound by the California Rules of Professional Conduct (“CRPC” or “Rules”), statutes and case law when determining their ethical obligations.  This article looks at three recent sources of guidance that California attorneys can consider in using GAI: the California State Bar, Practical Guidance for the Use of Generative Artificial Intelligence in the Practice of Law, adopted by the Board of Trustees on November 16, 2023; ABA Formal Opinion 512, issued on July 29, 2024; and the California Lawyers Association, Task Force on Artificial Intelligence Report on AI in the Practice of Law issued September 7, 2024.

First, what is AI?  ABA Formal Opinion 512, footnote 1-states that “there is no single definition of artificial intelligence.  At its essence, AI involves  computer technology, software, and systems that performs tasks traditionally requiring human intelligence” At a recent panel discussion on the topic it was suggested that it is just “Math doing things with Data”.[2] However, even though the term may appear ambiguous,  computer generated assistance will continue to impact society, and particularly the practice of law.  Therefore, attorneys must stay abreast of the ever-changing technology of AI and how they can continue to meet their ethical obligations under Rule 1.1, Comment [1].[3]  This article focuses on GAI, which requires the user to input relevant data and information, which might include confidential information, and which is covered in the three resources discussed below.

California State Bar- Guidelines (adopted by the Board of Trustees on 11/16/23)

The first resource, adopted by the Board of Trustee in November 2023 was prepared by the State Bar of California Standing Committee on Professional Responsibilities and Conduct and is entitled Practical Guidance for the Use of Generative Artificial Intelligence in the Practice of Law (the “Guidelines”).  The Guidelines provide applicable Rules that attorneys should consider when using AI.

The benchmark of client representation is the duty of confidentiality.  Thus, the Guidelines stress the importance of ensuring that the use of AI does not violate Business and Professions Code 6068(e)(2); Rule 1.6, and Rule 1.8.2.  The use of GAI, which as noted requires the user to input relevant information, can cause concern.  Once data consisting of confidential information is inputted, the attorney has an obligation to ensure that it is not disclosed to a third party, in any format, without the client’s consent.  To meet this obligation, the attorney must be familiar with the software and its limitations and risks, e.g., attorneys need to ascertain whether the software contains adequate confidentiality and security protections.  This might require an attorney to consult with IT professionals or cybersecurity experts.

Equally important is the duty to supervise lawyers and nonlawyers in regards to the lawyer’s ethical obligations (See Rules 5.1, 5.2 and 5.3).  There may be circumstances where associates or paralegals assisting an attorney may employ GAI as an expedient means to meet their deadlines. However, lawyers have a duty to supervise subordinates to ensure that the information (output) is accurate, that there are no “hallucinations”[4] and that no confidential information has been inputted to get the result without the consent of the client. 

Other issues in the Guidelines include whether there is a duty to communicate to the client regarding the use of GAI.  According to the Guidelines, there probably is not.  However, depending on the nature of the representation, it may be prudent to convey to the client that the attorney intends to use GAI.   In addition, the Guidelines, recognizing that GAI can permit a lawyer to work more efficiently, explains that under Rule 1.5, a lawyer may charge for the time required to craft and refine the GAI information inputs, but may not charge for the “time saved” by using GAI.

Another important issue noted in the Guidelines is the fact that GAI is based on algorithms and its output is based on the information inputted.  Therefore, if the data includes implicit bias or is based on discriminatory assumptions, the results can be very troubling and may run afoul of Rule 8.4.1.  Thus, lawyers have an obligation to understand the underlying assumptions of the data that are being inputted, whether such assumptions are valid and/or reflect systematic bias, etc. 

Lawyers must have a basic understanding of the GAI product, its benefits and disadvantages, and the risks involved in its use.  Lawyers still must be competent and cannot assume that a GAI tool will competently perform the legal task for which the lawyers employ it. The GAI tool’s output must be reviewed for accuracy.

ABA Formal Opinion 512 (“Generative Artificial Intelligence Tools”) (July 29, 2024).

ABA Formal Opinion 512 (the “Opinion”) covers many of the same issues addressed in the Guidelines, including the issues of competency, confidentiality, and what is a reasonable fee.  However, it places more emphasis on whether an attorney must disclose the use of GAI to the client, contrary to the advice provided in the Guidelines and suggests that client consent may be required[5].  The Opinion provides “to competently use a GAI tool in a client representation, lawyers need not become GAI experts.  Rather, lawyers must have a reasonable understanding of the capabilities and limitations of the specific GAI technology that the lawyer might use.” [See Opinion at pgs. 2-3].

The Opinion concludes that it is probably permissible to use GAI technology  to generate an initial draft of a document the lawyer is preparing, but warns against relying solely on a GAI tool to perform tasks that call for the exercise of professional judgement. [See Opinion at pg.3]  As is known, a GAI can produce results that are not accurate, i.e. “hallucination”.  Thus, it is important the lawyers confirm any conclusion and/or document that is based on the GAI product. The Opinion, similar to the Guidelines, stresses the importance of protecting client confidentiality.

CLA- Taskforce Report (September 7, 2024)

The last publication I will discuss is the “Report on AI in the Practice of Law” (“Report”) by the CLA-Task Force on Artificial Intelligence.[6]  The Report reviews the importance of the Duties of Competence and Diligence (Rules 1.1 and 1.3) , the Duty of Confidentiality (Rules 1.6 and 1.8.2 and Bus. and Pro. Code §6068(e)), and understanding the risk of the use of the tool. 

Uniquely, the Report recommends the formation of a partnership consisting of standing committees, the state bar, GAI developers, vendors and stakeholders, which could among other matters, establish a state bar certification for vendor, as a means to certify that the GAI product would meet certain levels of disclosure and safety.  The Report also recommends1) that legislation be adopted which requires GAI vendors to obtain licenses to work with the legal professional, or (2) that the California State Bar provide optional certification for lawyers.  Such alternatives could provide assurance that base line protection and security are included in the GAI product.  These types of alternatives could assist attorney in complying with their ethical obligations under the Rules, particularly as GAI develops and attorneys must ensure that they stay abreast of the ever-evolving product in compliance with Comment [1] of Rule 1.1.

In conclusion, before a lawyer uses GAI, he or she must be aware of the limitations and risks and ensure that the software has security and protections to ensure that confidentiality is protected.  Lawyers still have a duty to do their own due diligence and to be competent.  GAI is another tool to assist lawyers to remain competent, but GAI cannot represent the client.  Lawyers cannot become over-reliant on GAI.  However, to continue to provide competent representation, lawyer must stay abreast of technology, including GAI. 

Dianne Jackson McLean, Esq., partner at Goldfarb & Lipman LLP.  Ms. Jackson McLean is a member of the Ethics Committee of the California Lawyers Association.  The views expressed herein are her own.

---

[1] The discussion is not limited to the legal profession, and is a topic that being discussed in every major sector of our society, including the government.

[2] CLA 2024 Annual Meeting and Leadership Conference, The Landscape of Artificial Intelligence n the Legal Profession:  Trends, Implications and Ethical Considerations, September 7, 2024.

[3]Comment [1] to Rule 1.1 (Competence)-states: 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.”

[4] AI hallucinations are when a large language model (LLM) perceives patterns or objects that are nonexistent, creating nonsensical or inaccurate input. (IBM webpage). Terminology regarding AI is not consistent. However, Generative AI has been defined as “A type of AI technology that generates content such as text, images, audio, and video.  Also, sometimes referred to as a generator.  LLM is defined as “A complex model trained on vast amounts of data that generates language that resembles human-generated language”, Stanford University, Teaching Commons, Teaching Guides, Artificial Intelligence Teaching Guide, Defining Al and chatbots.

[5] See Opinion at pg. 7.

[6] The Report is available at https://calawyers.org/california-lawyers-association/california-lawyers-association-task-force-on-artificial-intelligence/ (last visited 9/19/24).