Clients Using AI - What It Means For Lawyers
Clients are increasingly bringing AI-generated material to their lawyers.
What does their use of AI means for the work that lawyers then have to perform, the risks they assume, the fees they need to charge and ultimately the business model of legal practice.
I recently wrote a post online about the growing frustration among lawyers dealing with clients who arrive with AI-generated legal research, draft documents, suggested arguments and sometimes, a fairly firm view about what the law is and what should happen, even before they have even spoken to a lawyer.
The response to that post resonated with a lot of people, as I suspect it touched on a problem that many lawyers are already experiencing but have not yet learnt how to articulate clearly to clients, let alone process properly for themselves.
I don't think taking a completely binary pro-AI or anti-AI position really helps in dealing with this issue.
Ban it, adopt it, AI is good, AI is bad. Unfortunately, none of those positions deals particularly well with what actually happens when the client arrives with a tangled mass of material produced with AI. And they are and will continue to do so.
Some of what they bring may be misguided. Some of it may be quite plainly wrong. However, some of it may actually be right, and there may also be something useful buried amongst it.
I also think as a lawyer you owe it to the client to respect at least the fact, that they have done some research into their own problem. Even where the material turns out to be wrong, it may give you an indication of where the client is coming from, what they think the problem is and where they think the matter is going, and that has to has some ultimate value.
All of this is reality being played out in lawyers' offices around the world right now, in real time, and that reality as I said before isn't going away.
Clients are going to use AI whether you like it or not
It is clear that clients are going to use AI before, during and after they retain lawyers.
AI is also going to get better and clients are going to continue to use it at scale because access is cheap, or almost free and it is user friendly.
I think the legal consumer, the consumer of legal information who then becomes a client, is increasingly going to see themselves to be in some form of ongoing partnership with the lawyer as well as AI. There will be the client's AI and, increasingly, the lawyer's AI operating throughout the matter, before, during and potentially after the transaction or matter itself.
The more interesting and pertinent question will increasingly be what the lawyer actually does with this client-created information when it arrives, and to the increasing realisation that there is a partnership of sorts that has not previously existed of this nature.
Additional invisible and visible legal work
I think it has to be recognised upfront, in light of the above, that dealing with the product of the above is additional legal work.
It is also hard legal work and complex legal work by its very nature.
It isn't as simple as just sifting through what the client has produced, and it is certainly not necessarily the sort of routine work that can itself be easily automated.
Ironically, I think it is creating a new category of legal work that is largely invisible to the client, but highly visible to the lawyer.
On the one hand, the client's use of AI may save the lawyer a lot of time. On the other, it may actually create a lot of additional time and expense, because the lawyer is now spending considerable time using their experience and professional judgment reviewing material the client has produced using AI, some of which may frankly be just plain unreliable, but some of which may be very helpful. It is not a straight offset however.
The lawyer still has to read it all, digest it all and assess it all. They have to identify what is wrong, what is misleading and what is incomplete, as well as what might actually be useful. Then they have to understand what the client has relied upon and why. Finally, they have to exercise their own professional experience and judgment over it.
The reality is that this is work. It is, as I said before, legal work and, in many cases, it is complex legal work that has to be charged for - and charged a lot more than average for.
I think validation, correction, professional experience and professional judgment are all becoming an increasingly important part of the modern lawyer's toolkit when reviewing AI-generated legal information.
That also raises another problem which I think is going to become much more significant - the distinction between legal information and legal advice is starting to become very blurred, particularly from the consumer's and then ultimately the client's perspective. AI goes to pains to say this is not advice but the consumer sees it as just that unfortunately as courts all over the world are experiencing first hand with pro se litigants by-passing lawyers altogether.
The commercial question
A client can now obtain enormous amounts of seductive ad persuasive sounding legal information almost instantly and for almost nothing. They can ask questions, produce documents, develop arguments and reach conclusions about their legal position before they ever speak to a lawyer. Understandably, some will then arrive at the lawyer's office believing that quite a lot of the legal work has already been done. Others will have the bravado to go solo themselves without one.
That is where the commercial question for lawyers becomes interesting. How do you charge for this sort of work and, probably more importantly, how do you pitch it to the client? Also, what happens to the economics of legal practice when clients increasingly bring this material into the lawyer-client relationship potentially at all stages of the retainer?
That is a real-time commercial question lawyers are facing as we speak.
Reviewing client-generated AI material can make the lawyer's job more difficult and more risky. Frankly, it may not be particularly pleasant work either, as you might be arguing with your client in the very early part of any retainer with them about who is right the AI or you as the lawyer, let alone even before engaging with the other side.
But again, I don't think you can be indifferent or arrogant and simply dismiss it all. It has come from your client, they may have placed considerable reliance upon it and there may actually be something useful in it. You need to review it and explain it to them irrespective of how you feel about it - oh, and charge for it.
Fees and professional responsibility
There is also an important connection here between fees and professional responsibility.
Whatever the client has generated before seeing you, the lawyer still has to take proper instructions. The lawyer has to understand the client's problem, verify the material that matters, exercise their experience and independent professional judgment and ultimately be responsible for the advice that follows.
The client actually wants this with their AI-generated material. Indeed, if that advice is wrong, the client isn't going to blame AI. The client is going to blame the lawyer.
The above reality has to be reflected in how the work is pitched and how it is charged. The lawyer isn't charging more simply because the client used AI. The lawyer is charging for the work required to review it, verify it, correct it where necessary and exercise professional judgment over it, together with the responsibility and risk that ultimately follows regarding acting on it.
It's also about the changing nature and the changing role of being a lawyer.
It may actually all be creating more and not less legal work, albeit more work of a different and potentially more sophisticated kind, in a net sense. Ironically, it may not necessarily be about lower fees either. It could entail higher fees where the work requires more sophisticated advice, greater judgment and greater professional responsibility from the lawyer. I think most of this Ai generated client work sits in that category. I know the client will not think the same though.
There is a real tension here, because lawyers are going to have to reconsider how they charge for this sort of work but, more importantly, how they sell and explain their value to the client.
The client is effectively bringing material to the lawyer and saying, "I've already done this work." The lawyer then has to explain why their processing of that work has a value of its own.
This can be high-risk work and, where the complexity, time and risk justify it, as I said before the fee should reflect that.
"Can you just have a quick look at this?"
Experienced lawyers already understand the danger in the apparently innocent request - "Can you just have a quick look at this?"
I still remember when the internet came into existence, my clients cobbling stuff off it and asking me to take a quick look.
There is often nothing quick about properly reviewing somebody else's work. AI potentially magnifies that problem because clients can now produce an enormous amount of apparently sophisticated material very quickly and at virtually no cost.
We may almost be reaching the point where lawyers have to ask every client whether they have used AI in relation to their matter (and maybe at every stage).
Once you ask that question, however, another whole group of issues appears. There is the material the client has supplied and the need to verify it. There are confidentiality and privilege issues. There is almost an issue around supervision of the client's use of AI. There is the engagement itself and what is actually within the scope of the retainer. There also again that vexed issue of the pricing of the additional work.
The business model of the law practice
It is the client's use of AI that is changing the business model of the law practice.
That, to me, is a big issue.
What does all of this mean for the work lawyers actually do and how we charge for it?
How do we explain the value of that work to the client?
What is actually within the scope of the retainer?
What does the client now expect from the lawyer?
And do any of us really understand the complexity and impact of all of these issues that are coming to the fore in real time?
I think the answer is probably no.
We are all struggling with it in real time.
