Bench & Bar May/June 2026

gets to bill more. This is not a reasoned approach for determining the propriety of lawyer professional conduct; the Rule should be amended to address this disparity. A CRITICISM OF ABA FORMAL OPINION 512 The American Bar Association’s Formal Opinion 512 (Generative Artificial Intel ligence Tools) addresses the use of AI in billing, but its apparent adherence to the time-and labor model exacerbates the problem. The Opinion states that “GAI tools may provide lawyers with a faster and more efficient way to render legal services to their clients, but lawyers who bill clients an hourly rate for time spent on a matter must bill for their actual time.” 6 Hence, as most lawyers bill by the hour they may not bill a client for time saved by AI. 7 This position is consistent with the historical reading of Rule 1.5, but it only reinforces the very flaw that makes the rule obsolete as it assumes that value is derived from sweat, not judg ment. The lawyer’s value is in wisdom, discernment, and trust—knowing which words the machine got wrong or standing behind the machine when it is accurate. By mandating a fee reduction propor tional to time saved, the Opinion creates an economic disincentive to utilizing tech nology that improves the quality, speed, and accuracy of the client’s result. The point of professional ethics is to preserve trust, and trust is maintained when the human lawyer masters, supervises, and stands in front and behind the machine. In the AI era, a reasonable fee should be measured not by inputs but by outputs : accuracy, speed, reliability, and client impact. A lawyer’s unique value lies in what we might call “good legal judg ment”—trustworthiness, sound judgment, clear communication, empathy, verification, confidentiality, and accountability—ele ments not captured by the old factors. A “ value-assurance model ” that assesses rea sonableness by the quality of legal judgment and risk management, not the quantity of time expended is needed. THE NEED FOR A VALUE AND OUTCOME BASED PARADIGM

Value-assurance pricing—an approach that sets fees according to the value and impact of the outcome delivered to the client— offers a more reasonable and responsive approach. This model can include: • Flat or fixed fees for discrete ser vices (such as document drafting or basic contract preparation) that allow clients to know costs up front. • Capped fees , which provide a maximum limit for costs while still enabling hourly or hybrid billing. • Hybrid or flat-fee-plus models , blending flat rates for predict able work and hourly or perfor mance-based rates for contingen cies or unusual complexity. • Success or performance incen tives , where additional fees are paid upon achieving predefined client outcomes, such as favorable verdicts or negotiated resolutions. An experienced lawyer, at the outset of the representation, can often assess the likely outcome and the expected complexity of the task. These factors—not the hours devoted—are the crucial elements in deter mining whether a fee is reasonable. I agree with the Virginia State Bar in its Opin ion on this subject. The Virginia proposal addresses whether and how lawyers can bill for work done with the assistance of AI using a value-based or other non-hourly billing method. The Virginia Opinion makes the following valid points. • While AI can dramatically reduce the “time and labor required” for certain tasks, the “skill requisite to perform the legal service properly” might actually increase, as effec tive AI use could require special ized knowledge to prompt, verify, supplement, and integrate AI out puts into competent legal work product. The lawyer’s judgment in determining when and how to deploy AI tools, and the exper tise needed to critically evaluate AI-generated content, represent valuable services for which the lawyer reasonably can be compen sated.

in a modern age of science. Furthermore, professional norms that require lawyers to adopt technologies that improve efficiency and access to justice (see Comment [6] of Rule 1.1) 5 but embed an economic disin centive for doing so is not ethics, it is an economic absurdity. Rule 1.5(a) not only emphasizes a time-and labor approach to billing, but it also requires that reasonableness be determined by the fee customarily charged in the locality for similar legal services, regardless of whether other lawyers in that very same locality are efficiently using technology. Specifically, when AI tools generate well-reasoned drafts or locate precedent in seconds, the human lawyer’s “skill” is redefined neither in locating nor reciting law, but in curat ing, interpreting, and verifying the outputs of an intelligent system. The Rule, however, offers no mechanism to value this oversight skill —the judgment to know when AI is dependable and when it is dangerously wrong. Hence, the lawyer who responsibly automates is penalized, while the lawyer who does not adopt technology and spends days on a task that could take three minutes

21 bench & bar