Listen to this article
Artificial intelligence is en vogue. Lawyers use AI for legal research, briefs, searching and summarizing documents, and for background searches on witnesses and parties, among other tasks. For those of us who are old enough, the AI revolution is similar to the evolution from books and West Digests to computer research through Westlaw/Lexis. When Westlaw and Lexis first came on the scene, lawyers faced the question of whether and how computer research services could be billed. The same question has now arisen for generative AI technology.
While Maryland’s various ethics committees have not yet opined on the billing of AI technology, other states and the ABA have. See VSB LEO 1901; ABA Opinion 512 (2024); North Carolina Ethics Opinion 2024-1. For example, the Virginia State Bar recently adopted Legal Ethics Opinion 1901 Generative AI and Billing. LEO 1901’s goal was authored to provide ethical guidance to Virginia lawyers on billing for generative AI services, especially in flat fee or alternative billing arrangement cases. The committee and authors recognized that while AI saves time on certain tasks, creating the generative AI models takes time, skill and effort which can be billable. Importantly, the VSB’s conclusion differs from ABA Opinion 512 which concluded that the time savings may mean the fee is not reasonable.
In focusing on ethical billing for generative AI, the authors, as expected, reviewed Rule 1.5’s requirements that fees be reasonable as required by Rule 1.5(a). (The Maryland Rule is 19-301.5(a)). According to LEO 1901, flat fee or alternative billing is permissible under Rule 1.5. Virginia did not necessarily equate reduced time with proportionally reduced fees because such an approach failed to account for the investment lawyers make in developing Al expertise and the continuing value of their professional judgment. Instead, a proper analysis recognizes that reasonable non-hourly fees can reflect efficiency gains, the specialized skill of effectively incorporating technology, and the value of the relevant services and output. Accordingly, just because a lawyer used generative AI does not mean that the client is entitled to a reduced fee for less time spent.
While it remains unclear how Maryland will approach billing for generative AI legal services, there is ample guidance already in place. First, as with all fee agreements, the fee must meet Rule 19-301.5’s reasonableness requirement. Second, the fee needs to be explained fully to the client, preferrable in writing. Indeed, under Rule 19-301.15, nearly all fee agreements should be written, regardless. Should a lawyer have a specific charge for generative AI, the engagement letter should specifically reference the fee.
Please also only use generative AI as a tool. Failure to use AI properly, including not cite checking cases, may be considered violations of other rules. Potential concerns include Rule 1.1 because competence includes knowing how to use technology properly, Rule 1.6 confidentiality if the AI model is not in-house, Rule 3.1 and 3.3 for pursuing only meritorious claims and defenses and being candid with the Court, and Rule 8.4 misconduct for prejudicing the administration of justice. It is only a matter of time before we see public charges in Maryland for improper use of AI.
Finally, as a reminder, Maryland has amended Rule 19-301.15, and the change can impact charges and billing for generative AI technology. The amended Subsection (a) requires all fees to be put in trust until the fee is earned, even for flat fee and other alternative fee arrangements. When setting fees that include use of generative AI, remember to put the unearned fee into your IOLTA account, and transfer only after the fee has been earned.
Craig Brodsky is a partner with Goodell, DeVries, Leech & Dann LLP in Baltimore. His legal ethics column appears monthly. He can be reached at [email protected].
