The Ethics of Generative AI – More of the Same?

Posted by kvinson on

By: Laura Hennessey

As generative AI has taken the world by storm, its role in the legal profession has only just begun to take shape. Initial skepticism and words of warning have given way to a tentative embrace of the new technology in legal research and writing. With the advent of legal-specific generative AI tools, generative AI seems destined to become embedded in legal practice. But the risks of relying on generative AI tools are ever-present, especially as the technology remains in a nascent stage. So how should law students and new lawyers think about the ethics of using products powered by generative AI in school and in practice?

The American Bar Association (ABA) weighed in on July 29, publishing its Formal Opinion 512 on Generative Artificial Intelligence Tools. ABA Standing Comm. on Pro. Ethics & Resp., Formal Op. 512 (2024), https://www.americanbar.org/content/dam/aba/administrative/professional_responsibility/ethics-opinions/aba-formal-opinion-512.pdf [hereinafter Formal Op. 512 or Formal Opinion]. The Formal Opinion posits that the ethical consequences of generative AI tools are best assessed in the same way that lawyers already judge the ethics of their actions—with reference to the Model Rules of Professional Conduct. Just like the name suggests, the Model Rules of Professional Conduct is a set of ethical rules that the ABA recommends lawyers follow in their practice. Every jurisdiction in the U.S. has adopted ethical rules based at least in part on the Model Rules of Professional Conduct. The ABA’s Formal Opinion proposes that generative AI could interface with at least six categories of the Model Rules of Professional Conduct: (1) competence, (2) confidentiality, (3) communication, (4) candor toward the tribunal, (5) supervisory responsibilities, and (6) fees.

  • Competence

Model Rule 1.1 requires lawyers to provide competent representation to their clients, including possessing the requisite “legal knowledge, skill, thoroughness and preparation,” as well as understanding the tradeoffs associated with all utilized technologies. Model Rules of Pro. Conduct r. 1.1 & cmt. [8] (Am. Bar Ass’n 2024). As it relates to generative AI, the ABA professes that “technological competence presupposes that lawyers remain vigilant about the tools’ benefits and risks.” Formal Op. 512 at 3. The risk that generative AI tools will deliver false information requires lawyers to engage in sufficient “independent verification or review” to ensure accuracy. Id. at 4. Likewise, because generative AI tools are becoming enmeshed in the profession, use of these tools may become necessary to deliver competent legal representation. For this reason, the ABA says lawyers have an affirmative obligation to become familiar with generative AI tools. Id. at 4-5. In this way, lawyers will continue to be able to provide competent representation to their clients while utilizing the latest technology.

  • Confidentiality

The Model Rules of Professional Conduct require lawyers to keep confidential information related to representation of current, former and prospective clients, unless the client specifically approves of the disclosure. See Model Rules of Pro. Conduct r. 1.6, 1.9(c), 1.18(b). Before using a generative AI tool, therefore, lawyers should assess the likelihood that the information disclosed to the tool could be accessed outside their firm or other employer. Self-learning AI products make this a real risk, as information input for one client’s problem could be repeated back by the AI tool when working on another issue. In the ABA’s view, lawyers are obligated to educate themselves on the terms of use of the generative AI tools they are using, and to obtain informed consent from their clients when necessary. Formal Op. 512 at 7. Legal-specific AI tools may offer more protection and transparency surrounding the subsequent use of the information that is input, hopefully making it easier for lawyers to utilize generative AI tools without the risk of disclosing client confidences.

  • Communication

The ABA Formal Opinion encourages lawyers to communicate openly with their clients about their use of generative AI tools during their representation. Formal Op. 512 at 8-9. This is in line with Model Rule 1.4, which prescribes lawyers to “reasonably consult with the client about the means by which the client’s objectives are to be accomplished.” Model Rules of Pro. Conduct r. 1.4. Essentially, lawyers should not utilize generative AI unless they would be comfortable discussing that use with their clients. Moreover, lawyers should consult the Model Rules and the Rules in their own jurisdiction surrounding the specific circumstances when disclosure may be required.

  • Responsibilities to Courts

The use of generative AI also implicates a lawyer’s ethical responsibilities to courts and other judicial bodies. The relevant Model Rules concern a lawyer’s obligation to tell the truth to the court—avoiding any frivolous claims or other fraud or misrepresentation. Model Rules of Pro. Conduct r. 3.1, 3.3, 8.4(c). These obligations underscore the need for attorneys to independently verify the accuracy of their work product—especially when generative AI is involved—before presenting it to the court, including checking all citations and discussions of authority. Generative AI, at least in its current form, is well-known for making mistakes and generating “hallucinations”—all of which is the lawyer’s responsibility to catch and rout out.

  • Supervisory Responsibilities

The use of generative AI tools in legal practice is analogous to a lawyer’s reliance on a junior associate or paralegal to conduct relevant legal work. The Model Rules impose an obligation on managerial lawyers to ensure that subordinates comply with the rules of professional conduct. Model Rules of Pro. Conduct r. 5.1, 5.3. In the context of generative AI, this means that supervisory lawyers should create clear guidelines on the use of generative AI at their place of work and should provide appropriate training to junior associates and other staff. Formal Op. 512 at 10.

  • Fees

Finally, the use of generative AI tools impacts the fee that is appropriate for lawyers to charge for a given work product. Because generative AI tools are designed to increase efficiency, use of generative AI should typically reduce legal fees. Formal Op. 512 at 12. As always, providing clients with transparency regarding the basis for the fee charged is paramount. Model Rules of Pro. Conduct r. 1.5(b). In the context of generative AI, that means being upfront about the use of generative AI and reflecting the efficiency gains of those tools in the fee charged.

In short, by using the traditional framework of the Model Rules of Professional Conduct to assess the risks of generative AI, the ABA makes it a bit easier to bring this new technology into the fold. Generative AI may be best thought of as simply the next iteration of legal technology, in line with the revolutionary digitization of legal research that began in the 1970s. By continually referencing the rules of professional conduct in their jurisdiction, law students and new lawyers can be assured they are utilizing generative AI in an ethical manner.