In recent years, the topic of artificial intelligence (“AI”) has quickly dominated conversations across various industries, institutions, and sectors. The legal profession is no exception; AI usage is embedded in nearly every aspect of modern legal practice.
At the ABA Business Law Section’s Spring Meeting in Atlanta, GA, Monika McCarthy, Managing Director & General Counsel at CrossCheck Compliance LLC, moderated a CLE program offering actionable insights for this era, titled “AI, Esq.? Legal Ethics and Practical Considerations for Business Lawyers.” The panel included speakers César Escovar, Senior Manager of Compliance Advisory at Capital One; Sarah Gatti, Head of Legal at Zappi; Tammy Malvin, Partner at Akerman LLP; and Drédeir Roberts, Founding Member at Drédeir Law. Drawing on their combined experience from governance, regulatory, litigation, and transactional practice, the panel explored how lawyers can responsibly integrate AI into their practices in compliance with legal ethics obligations, while also emphasizing certain nondelegable aspects and limitations of AI.
As background, ABA Formal Opinion 512 provides that “to ensure clients are protected, lawyers using generative [AI] tools must fully consider their applicable ethical obligations, including their duties to provide competent legal representation, to protect client information, to communicate with clients, to supervise their employees and agents, to advance only meritorious claims and contentions, to ensure candor toward the tribunal, and to charge reasonable fees.” Formal Opinion 512’s guidance confirms that lawyers’ use of generative AI does not relieve them of existing ethical rules and that there is no “AI exception” to professional responsibility.
The panel presented examples and key takeaways drawn from ABA Model Rules of Professional Conduct to demonstrate the various instances where lawyers, as well as judges, have professional responsibilities to ensure accuracy, informed judgment, and supervision of AI-generated content.
As the panel’s slides noted, “AI tools are increasingly used for legal research, drafting, discovery, litigation strategy, transactional work, and enterprise governance.” While Model Rule 1.1 (competence) means it is crucial for lawyers to know how generative AI works conceptually, it is not necessary for lawyers to understand the technical aspects of it, such as coding. By having a working knowledge of AI tools, lawyers can show that they are continuously staying abreast of how technological advancements are impacting the way they serve their clients and their ability to provide effective representation. However, consulting or relying solely on AI experts is not enough; the panel encouraged lawyers to also consult with information security and privacy experts as part of their AI literacy training in order to draw on new perspectives and establish greater competency to meet enterprise or client needs.
From a compliance standpoint, the panel noted that while AI is not considered a legal entity, courts have a tendency to treat it as an operational agency acting on behalf of the company. When inputting data into generative AI, lawyers are bound by strict confidentiality (Model Rule 1.6) and must protect against third-party data usage. They should be able to map the data flow and understand how pass-through data are being handled, including aggregate or anonymized data. To ensure proper handling of AI data, lawyers must establish vetting protocols to ensure the third-party vendor’s management processes and data policies do not pose risks to the enterprise and client.
When communicating with clients (Model Rule 1.4) or submitting legal documents to the courts (Model Rule 3.3), it is important to note that AI “amplifies both good and bad lawyering,” the panel noted. Presenters highlighted two AI-assisted litigation cases, Warner vs. Gilbarco, Inc. and U.S. vs. Heppner, to demonstrate that courts are still grappling with how traditional privilege and work-product doctrines apply to AI-generated materials. One thing that has been clear is that AI is viewed as a nonlawyer assistant, and if a lawyer or judge directed the AI use, then their supervisory obligations (Model Rules 5.1 and 5.3) apply. Otherwise, lawyers and judges run the risk of letting inadvertent citation or reliance on bad law to result in cascading effects if they are not vigilant about their AI use.
With the increased use of AI tools in the modern legal practice comes the lawyers’ responsibility to also adopt new billing practices that reflect their obligation under Model Rule 1.5 to charge reasonable fees. Although lawyers are not required to describe hourly work done with granular specificity, the panel encouraged lawyers to develop a standard practice of including an AI section in the client engagement letter, describing the scope of AI use and providing full disclosure regarding their handling of client data generated by AI.
In summary, generative AI is a powerful tool that has spread through modern legal practice, but when using it to build efficiency, lawyers and courts must be vigilant in complying with the ABA Model Rules of Professional Conduct. Lawyers must adhere to their ethical responsibilities when using AI by exercising informed judgement, establishing rigorous verification processes, adopting clear AI governance, and ensuring continuous human oversight over AI-generated content.
