Most of this section is about building things. This one’s about what you owe the person on the other end of them.
Two questions, and they don’t get the same answer. Do you have to tell the client? That one’s genuinely unsettled, and reasonable lawyers disagree. What happens to the bill? That one’s not unsettled at all, and the answer is the one lawyers least want to hear.
The fee question is settled
Start with the bill. It’s the clear one, and it’s where the money is.
The ABA’s Formal Opinion 512, issued July 29, 2024, takes up generative AI under the rules you already know. On fees under Model Rule 1.5 it makes two points, and neither leaves much wiggle room.
You can’t bill a client for time spent learning a technology for general use in your practice. Learning your tools is overhead. Same reason you can’t bill a client for learning a new research database or a new document management system. (The opinion treats it differently when a client specifically asks you to use a particular tool on their matter.)
And if the work took less time, you bill less time. A two-hour task that now takes twenty minutes is a twenty-minute entry. The efficiency belongs to the client.
Sit with that second one. Everything else in this section hangs on it. Reasonable-fee rules have always keyed to time actually and reasonably spent. When a tool squeezes the work down, the entry shrinks with it. There’s no reading of the rules where the lawyer pockets the gain by billing the old number.
That leaves hourly practice with some ugly math: the better you get at this, the less you bill for the same work. The rules aren’t broken. They’re doing exactly what they were built to do. And it’s the best argument I know for moving compressible work to fixed fees. The firm keeps the efficiency it created without billing for time it didn’t spend, and the client gets a number up front.
The dishonest fix, same hours for less work, would quietly undo everything this section is for. If AI-assisted practice just means the same bill for less effort, nobody outside the firm is better off, and the access argument falls apart. The whole reason to publish any of this is that the gain shows up as a smaller bill, or as a case somebody couldn’t otherwise afford to bring.
The disclosure question isn’t
Now the hard one. Do you have to tell the client?
Rule 1.4 requires a lawyer to “reasonably consult with the client about the means by which the client’s objectives are to be accomplished,” to keep the client reasonably informed, and, at subsection (b), to “explain a matter to the extent reasonably necessary to permit the client to make informed decisions regarding the representation.”
No rule says “disclose AI use,” and I don’t expect one that says it in those words, for a reason I’ve laid out elsewhere: in a few years that’ll describe every document every lawyer produces, the way “used a computer” does now. A disclosure requirement that covers everything tells the client nothing.
But 1.4 isn’t a checklist. It turns on the facts. Here’s how I sort it.
You clearly have to disclose when client data leaves your control in a way the client would care about. If a client’s confidential information is going out to a third-party service, that’s a fact about the representation the client has a right to know and, depending on the arrangement, to consent to. That’s the strongest case, and it isn’t close. Formal Opinion 512’s confidentiality analysis runs through the duty to protect information relating to the representation absent informed consent, and a client can’t consent to something nobody told them about.
You clearly have to disclose when the client asked. Obviously. And more clients are asking.
You have to disclose when it bears on a decision the client has to make. If your choice of approach changes the cost, the timeline, or the risk in a way that matters, 1.4(b) reaches it. The client can’t make an informed decision about something they don’t know about.
You probably don’t have to disclose ordinary internal tooling that never sends their information anywhere. A local model sorting the incoming mail is office equipment. We don’t disclose our word processor, our research database, or our calendaring system. What decides it is the question “does the client’s information go anywhere” and not the label on the tool.
And some disclosure is just smart, required or not: institutional clients with their own AI policies, matters where the work product is unusually sensitive, and any engagement where you’d squirm if the client found out later from somebody else. That last test has served me well on plenty of disclosure questions that had nothing to do with technology.
Put it in the engagement letter
Most of this gets solved with a paragraph, not a conversation.
An engagement letter that says plainly what technology the firm uses, what happens to the client’s information, and what stays under lawyer supervision does the 1.4 work up front, in writing, at the moment the client is actually deciding whether to hire you. A judgment call you’d otherwise make over and over becomes a settled term.
Here’s what that paragraph should cover, in plain English and not in the language of a software license:
- That the firm uses technology, including AI tools, to work more efficiently.
- Whether client information goes outside the firm and, if it does, under what confidentiality terms.
- That a lawyer reviews all work product and is responsible for it.
- That the client can ask about it any time.
Don’t let it get long. A dense technical disclosure nobody reads is worse than a short one somebody does. It looks like consent without being consent, and preventing exactly that is why informed-consent doctrine exists.
Watch the opposite mistake too. A clause broad enough to authorize anything isn’t informed consent either. “Client consents to the firm’s use of technology” tells the client nothing, and it wouldn’t survive a serious challenge. If the point is letting them make an informed decision, the paragraph has to actually inform them.
What clients actually ask
This isn’t a rule. It’s what I’ve heard, having had this conversation a number of times now.
Clients rarely ask whether you use AI. They ask two other things.
“Is a real lawyer looking at this?” They want to know somebody’s on the hook. Right question, and it has a clean answer: yes, and here’s what that means. Nothing goes out unreviewed, a lawyer signs it, and the responsibility for the work hasn’t changed.
“Am I paying less?” Also the right question, and it decides whether they believe your first answer. Tell a client their lawyer is dramatically more efficient, then send the same bill, and they’ll draw the obvious conclusion. They’ll be right.
So the two halves of this piece turn out to be one. In practice the fee answer is the disclosure answer. A firm that passes the efficiency through has an easy conversation about technology, because the client can see what they got. A firm that doesn’t will find that conversation getting more uncomfortable every year. It should.
The line that doesn’t move
One thing here isn’t a disclosure question, a fee question, or a policy question, and I’ll put it flatly.
The lawyer is responsible for the work. Not “responsible for supervising the tool.” Responsible for the work, in the same way and to the same extent as if you’d typed every word yourself at two in the morning. A bad citation in your brief is your bad citation. A missed deadline is your missed deadline. An error in a document you signed is yours.
That isn’t going to change, and I wouldn’t want it to. It’s what makes the rest of this safe to do. Everything in this section, the verification gate, the specialist chain, the refusal to guess, the human sign-offs, exists because somebody has to answer for the output, and that somebody is a person with a license.
What you owe the client you automated is what you owed them before: competent work, an honest bill, enough information to make their own decisions, and a name on the document.
Sources
- ABA Standing Committee on Ethics and Professional Responsibility, Formal Opinion 512 (July 29, 2024)
- Minn. R. Prof. Conduct 1.4 (Communication), Rule 1.5 (Fees), Rule 1.6 (Confidentiality)
Rules quoted were verified against the Revisor of Statutes. This is general commentary, not legal advice and not ethics advice. Formal Opinion 512 is advisory and not binding in any jurisdiction; disclosure obligations depend on your jurisdiction, your engagement terms, and the facts of the representation.