I’ve been getting ready to give the opening keynote at ABA TECHSHOW in Chicago next week. I plan to address what I think are the two critical questions facing the legal profession right now: “As AI displaces lawyers from a growing share of legal task performance, what will be left for lawyers to do?” and “How are we going to develop lawyers when we don’t know what we’re training them for, and when we can’t count on law firms to do the training anymore?”
I’ve argued that the answer to both questions is to re-establish the importance of lawyers’ human attributes and personal capabilities, and to apply those strengths to the three highest roles lawyers can play: advocate, advise, and accompany. Lawyers won’t be able to match AI on legal knowledge or technical legal skills, but they can still “out-human” the machines — if they reconfigure themselves as trustworthy guides and allies with legal insight and good judgment.
In the last couple of months, though, a new candidate for future lawyer value has started to come into sharper focus. That candidate is accountability, also known as risk transfer — and if anything, the emergence of legal AI actually strengthens its value proposition.
The central idea here is that clients want more than a legal answer or product. They want the assurance that a licensed, insured, regulated professional stands behind the legal service and accepts responsibility for the risks it might carry. Those risks can be reputational (a person to hold responsible if it all goes wrong) or financial (funds to compensate for any monetary loss). Whether as a source of confidence or a means of protection, the lawyer imprimatur has real value, especially if the lawyer’s prestige is high and their pockets are deep.
Several authors have made a compelling case for the risk transfer theory. Alex Su argues that in high-stakes matters, clients often are paying not so much for legal output as for the credible transfer of consequential risk to a trusted institution. Mark Barrett describes it as the “assurance layer” of legal services: verification, judgment, and defensibility, with a regulated and insured professional behind it all. And Mike Whelan Jr. notes that lawyers’ accountability has to be rooted in real expertise and judgment, not just a rubber stamp or a “moral crumple zone” for machine-generated work.
The rise of AI increases the importance of accountability. As AI commoditizes legal reasoning and artifact production, most lawyer output will become vulnerable to low-cost substitution. In that scenario, differentiation — not just from other lawyers, but also from AI offerings — shifts from “what I can do” to “how much assurance I can credibly provide.” The more trusted and well-established the provider, the more valuable are their materials and opinions — and the more the provider can charge for them.
If this is correct, it seems to me that one implication would be an eventual migration of competitive advantage from lawyering excellence to institutional credibility — and to the strength of the associated brand. If what the client is really buying is the reputation that stands behind the work, more so than the work itself, then already-famous lawyers and successful firms have a formidable built-in advantage over less well-known rivals, and certainly over newcomers to the profession who’ll never get the chance to prove themselves. That’s not an inspiring thought.
Nonetheless, I can see two ways in which the risk-transfer theory of future lawyer value, strong as it is, might not entirely pan out. Both have more to do with clients than they do with lawyers and law firms.
The first is that the willingness to pay lawyers’ premium price for accepting a transfer of risk varies widely across the market. It might be true, as the old saying went, that nobody ever got fired for buying IBM, because its reputation for reliability was sterling. But how many people could afford to buy IBM in the first place? Equally, Kirkland & Ellis might have a powerhouse brand, but the market for that brand doesn’t extend much past the Fortune 50.
Most of the “risk transfer” conversations tend to focus on large corporate/institutional clients, or on matters involving significant amounts of money or strategic importance. In these cases, accountability does have immense value, especially to whoever on the client side is signing their name to the purchase order. But this is also a very narrow and ferociously competitive market, and if the weapon of choice is brand power, there will be only a very few winners.
More to the point, risk transfer is only an attractive feature when you have a lot of risk to transfer. But most people who need legal help aren’t dealing with risks that justify paying a premium for elite assurance. They’re not looking for the strongest available reputational backstop, or for plausible deniability in case something goes off the rails; they’re looking for help in a tough situation that they’d like to get dealt with, as quickly and affordably as possible.
Most legal needs are common rather than exceptional. They arise from the little collisions and complications of everyday life and business: unpaid invoices, expiring leases, difficult neighbours, punitive policies, withheld entitlements, and a thousand other matters that, while profoundly important to the people involved, are not usually existential. For these issues, and for the people dealing with them, the gilt-edged assurance package is a luxury. They just want something that’s good enough, and they want it quickly and affordably.
I suspect lawyers have a tendency to value “professional accountability” beyond what many of their clients are willing or able to pay for it. People aren’t indifferent to the quality of their legal services, of course, and they’d be happy to receive the added protection and peace of mind that a powerful brand provides. But they also have to ask themselves: Do I really need that much protection, given the stakes and given my budget? Do I buy the platinum-level extended warranty, or do I take a calculated risk that it’ll all work out?
For most people, the alternative to hiring an expensive lawyer isn’t hiring a cheaper one. It’s doing nothing, or doing it themselves, or asking a friend, or downloading a form — taking their chances with the imperfect options in their price range. Legal AI is going to look increasingly attractive to this segment of the market. It’s not perfect, and it’s not even always reliable; but it’s cheap, immediate, and surprisingly good at a wide range of tasks. Sure, the lawyer’s opinion comes with greater accountability. But it’s also, for that reason and others, well out of budget.
The second problem with counting on accountability to sustain lawyers’ post-AI practices has bigger implications. To understand it, we need to draw a distinction between two different types of risk.
If the value you’re seeking, as a client, lies in the assurance that comes from a licensed, insured, regulated professional exercising judgment and taking responsibility, then you’re going to want a good, well-respected lawyer for that. But if what you really want is to cover any financial losses that might result if the legal assistance turns out to be flawed, that’s a different story. Because it’s not only lawyers who can absorb financial risk.
The major AI vendors, for example, are starting to offer limited contractual protection for their users. Microsoft’s Copilot Copyright Commitment will cover some copyright claims related to paid commercial Copilot services, while Google’s Generative AI Indemnified Services offering is similar. Meanwhile, some insurance companies are starting to develop AI-specific financial products: Armilla, for instance, is marketing both AI liability insurance and an AI performance warranty for specified model failures or underperformance.
It’s at least plausible that similar mechanisms could eventually be attached to legal AI tools, especially those aimed at individuals and small businesses with relatively straightforward and financially modest matters. Suppose an AI recognizes that a user’s request has a legal dimension or could have legal consequences. The AI might open a popup window offering a defined amount of protection, for a small additional fee, should the AI’s output lead to a specified category of loss.
To be very clear, this isn’t a development I would welcome. There are public-policy reasons not to encourage the spread of flawed AI-generated legal instruments, even if the parties involved are indemnified for any damages. But my larger point is that if what the user really wants is financial insurance, not professional assurance, there are other ways to arrange that. If the transfer of risk is really about transferring responsibility for monetary losses, that’s not an exclusively legal issue.
Like everything else to do with AI and the law, these are uncharted and potentially hazardous waters. The sudden appearance of legal service capability from non-lawyer (and non-human) sources means we have to figure out quickly where the public interest lies — how to balance the widespread availability of legal assistance with the equally wide-ranging consequences of its misapplication. But for present purposes, I think we can safely draw two conclusions.
One is that risk transfer will be a highly valuable asset in some markets, but a much less decisive one in others where the stakes aren’t that high and the budgets aren’t that big. And if the risk being transferred is addressed less by the security of highly regarded professional advice, and more by the promise of monetary compensation if things go wrong, then the importance of a lawyer’s involvement is reduced yet again.
And the other conclusion is that, at the end of the day, what lawyers think about the value of their various offerings will matter less than what clients think. AI empowers demand in many ways, and one of those ways will be to give people more choice when trying to solve their legal problems, and more freedom to take responsibility for what might flow from those choices. The real risk to lawyers, in fact, might be that clients might soon choose to keep more risk for themselves.
As usual, I agree with you and Mr. Carr as well (valuable reads, always). In addition, and related to his query about pricing, QUALITY has entered the chat. It's part of value, to be sure, but with complete lack of quality standards in the profession/business of law, the time has come, surely, for us to get on (or at least closer to) the same page about how we define "excellence."
Regardless, all this does is confirm for me that the People, Process, Platform sequence remains the right approach to continuous improvement, irrespective of whatever AI or other solution is being considered.
Compellingly insightful, as usual