Business lawyers adopting artificial intelligence must handle an expanding set of ethical and practical boundaries. A panel at the American Bar Association’s Spring Meeting in Atlanta addressed these challenges. Monika McCarthy of CrossCheck Compliance led the discussion, which included compliance officers, law firm partners, and in-house counsel. They examined how generative AI intersects with long-standing professional standards.
No “AI exception” to legal ethics
The ABA’s Formal Opinion 512, released in 2023, confirms that lawyers using generative AI remain subject to the same ethical duties that have long governed the profession. Competence, confidentiality, honesty with courts, and fair billing still apply. Panelists noted these obligations become even more pressing because the technology amplifies both effective and flawed legal work.
Related: Canada Proposes Legislative Changes Affecting US Border
César Escovar of Capital One and Sarah Gatti of Zappi explained that competence under Model Rule 1.1 does not require lawyers to master coding. It does, however, require understanding how AI tools process data, where errors may occur, and when human oversight is necessary. Gatti stated that consulting AI specialists alone is insufficient. Lawyers must also involve information-security and privacy teams to track how client data moves through these systems.
Data handling remains a key concern. Model Rule 1.6, which safeguards client confidentiality, applies to all information entered into generative-AI tools. Lawyers need to know how inputs are stored, anonymized, or shared with outside vendors. The panel advised implementing formal review processes for any vendor whose AI platform handles client data, even if only in aggregate form. Courts now view AI as an extension of the firm, meaning weak vendor security can create legal exposure for lawyers.
Courts are still determining AI’s role in litigation
Recent cases like Warner v. Gilbarco and U.S. v. Heppner show how courts are grappling with how traditional privilege and work-product rules apply to AI-generated materials. The panel presented these cases to demonstrate that lawyers have professional responsibilities to ensure accuracy, informed judgment, and supervision of AI-generated content.
Related: AI Hiring Bias Threatens Employers More Than Expected
Billing practices are adapting as well. Model Rule 1.5 mandates reasonable fees. The panel emphasized that some responsibilities cannot be outsourced. Drédeir Roberts, a solo practitioner, stated, “The client hires you, not the algorithm. If the AI overlooks a jurisdictional detail or misstates a law, the complaint goes to you, not the vendor.”
This risk makes ongoing oversight essential. The panel outlined governance measures firms should adopt, including regular audits of AI outputs, error-reporting procedures, and training beyond basic prompts. They suggested red-team exercises to test the tool’s reliability. Without these controls, efficiency gains can create ethical problems.
Related: E-Cigarette Explosions: Understanding the Risks and Prevention
The technology advances faster than regulations, but the panel’s point was clear: AI is a tool, not a substitute. Lawyers treating it as the latter may face disciplinary action over a chatbot’s confident but incorrect filing.
One area of agreement among speakers was that the discussion is ongoing. As courts issue more decisions and bar associations update guidance, the boundaries of acceptable AI use will continue to shift. For now, the best approach is to assume every AI-generated document will face scrutiny and act with that in mind.
