The analysis draws on ABA Model Rule 1.1 (competence), Rules 5.1 and 5.3 (supervision of lawyers and nonlawyers), and ABA Formal Opinion 512 to establish that deploying AI does not eliminate a lawyer's duty to understand and oversee the tool. It references vendor liability concepts, audit logs, human-in-the-loop checkpoints, and governance controls as practical frameworks for managing autonomous systems. The core tension remains unresolved: whether existing tort and professional-responsibility rules adequately address multi-step AI agents that can initiate actions across workflows with limited human involvement, or whether liability should scale with the degree of autonomy the system exercises.
The issue has moved from theoretical to urgent. Businesses are now embedding AI agents into live workflows, creating immediate questions about accountability when something fails. Regulators, courts, clients, and insurers will likely scrutinize the autonomy level of deployed systems and the safeguards in place. Attorneys should audit their AI governance structures now—specifically, the control mechanisms, decision logs, and human checkpoints—before liability questions land in discovery or a malpractice claim.