How Generative AI will change what lawyers do
As we enter the Age of Accessible Law, a wave of new demand is coming our way — but AI will meet most of the surge. What will be left for lawyers? Just the most valuable and irreplaceable role in law.
It seems the “trough of disillusionment” stage of Generative AI’s evolution is upon us. Goldman Sachs — which once predicted that 44% of work tasks in the US legal profession could be automated by AI — now says that AI is costing too much money with too little to show for it (kind of ironic coming from Goldman Sachs, IMO). Similar expressions of disappointment can be found elsewhere in the legal profession.
This is all very much to be expected, though — the early excitement over Generative AI was so intense that some letdown was inevitable. It doesn’t change my view that Gen AI is basically an Industrial Revolution In A Box, and that it will prove to be a transformative force in the legal sector over the next decade-plus.
For one thing, the Gen AI “Big Bang” is still exploding. New versions of frontier models are coming at breakneck speed — the latest AI releases from Google, Anthropic, Meta, and Midjourney have all appeared in just the last six weeks. These models arrive with little or no documentation, so nobody knows what they can really do until people start experimenting with them. It’s hard to assemble a library of proven legal use cases under those conditions.
For another, unlike most technology, Generative AI wasn’t built to do something; it was built to be something. AI researchers have been remarkably upfront that they’re trying to create superintelligent AGI — “to build a thinking machine,” even though they’re not really sure what that will actually look like.
I belong to the generation that watched T2: Judgment Day so many times we get PTSD flashbacks when people say things like that. But the point is that Gen AI’s current capabilities are, from their developers’ point of view, basically just byproducts of a more important ongoing project. Gen AI, therefore, is not “work technology”; but it’s technology you can use for work, or for play, or for anything else you like:
But for all that, even at this early stage of the Big Bang, we can start to trace the outlines of how the transformation of the legal sector likely will proceed. Anoop Singh, in a guest post at Artificial Lawyer, got me thinking along these lines with a recent post about Jevons Paradox and why Gen AI will increase legal demand:
In the post-AI world, attorneys will be getting their work done far more quickly and thoroughly. However, attorneys will be serving their clients far more deeply, serving more clients, and tackling many more legal questions. More individuals and companies will ask for more legal advice, especially in such a litigious society as the US. Firms that typically have not engaged attorneys will do so. Questions that today are not brought to attorneys will be put forth to them.
I think this is broadly correct: In due course, AI will increase demand for legal services by making those services easier and cheaper to procure, unleashing latent demand that had been pent up by the scarcity and inaccessibility of supply. But what will that look like in practice?
As fast machines replace slow humans as the primary performer of many legal tasks, productivity will rise and production costs will plummet. The result should be a dramatic increase in the accessibility of legal help — both from lawyers who adopt this technology and adjust their business models, and from new providers created or enabled by deep-pocketed global tech giants that will steamroll over UPL protectionism on their way into the market.
As we enter this era of Accessible and Reliable Law, we’ll see a mass migration of people and businesses from the latent legal market of unmet needs to the active legal market of met needs. Where once only 100 people or businesses could afford legal help, closer to 1,000 will be able to do so. At long last, legal demand will start to be matched by legal supply.
So the primary beneficiary of AI’s adoption in the legal sector will be overall economic productivity. More people will have valid wills, so there will be fewer people wasting time and energy fighting over intestate assets. More small businesses will get legal help, so more enterprises will successfully launch with strong foundations and face fewer setbacks, creating more jobs and more economic activity. The legal uncertainties and restrictions that held people back from fulfilling their potential and creating value will dwindle and fade.
This, I need hardly add, will be a Really Good Thing. I think it’s the most underrated benefit that legal services liberalization would produce, and I wish regulatory reformers and A2J advocates would cite it more often.
But when we look at the impact on lawyers, the picture changes. Remember, increased legal demand will flow directly from falling legal production costs. All those people and businesses will escape the latent legal market because they’re accessing AI-powered products and AI-accelerated services. The new surge in demand will be met mostly by low-cost, tech-driven supply. The first big wave of AI-enabled legal demand will crest short of the legal profession.
What about second-order effects? These newly empowered people and businesses will soon need more sophisticated legal help: More new wills create more wealth planning, more new businesses create more litigation, and so on. Lawyers, the theory goes, will simply upgrade to this higher-value work, climb the “value ladder” to heights that the AI can’t reach.
That would be a very happy ending — if Generative AI’s capacity were to peak at “basic” legal work. But as noted, AI researchers are spending billions of dollars to develop AI that’s better than humans at almost every cognitive task. They don’t have to achieve that goal, or even come particularly close, to still eclipse the majority of services lawyers offer.
AI can already provide actionable professional advice; within the next ten years, if it takes that long, I believe it will offer acceptable legal advice. No one really wants “AI courts,” but soon enough, we’ll have AI-enabled mediation and arbitration, which will have a much greater impact on everyday dispute resolution.
I think it’s dangerous to assume that AI will never be able to do something that lawyers now do. “Never” is a very long time. And AI doesn’t need to replicate the complete arsenal of the most gifted lawyer out there. If a Legal AI can replicate 80% of what a middling lawyer can do, for 10% of the cost, in 1% of the time, that’s all the revolution you’ll need.
So if you’re doing any kind of strategic planning in the legal space right now, one of your scenarios ought to be a world in which:
almost every transactional and documentation element of lawyers’ work has been commoditized by technology, and
most instances of advice and judgment are at least influenced or augmented by artificial intelligence.
If lawyers are going to survive in that world, or one much like it, we’ll need to redefine our role in the legal sector. We’ll need to come up with a new economic model for our profession. I’ve addressed this topic on a couple of previous occasions, here at Substack last summer and originally in a 2019 post at my Law21 blog. I’d encourage you to read both pieces, but here’s the core of my argument from five years ago:
The truly disruptive impact of advanced technology in the law will be to reduce the incidence and volume of traditional legal work given by clients to lawyers. … The old legal economy consisted of paying lawyers by the hour to do every legal task that needed to be done. In the new legal economy, systems, software, and structures are going to integrate, automate, delegate, and eliminate countless legal tasks by which lawyers once made a living.
Now, I want to be very clear about this. I am not saying that AI will replace lawyers, that lawyers will disappear because the machines will take away everything we can do. I suppose there’s a distant dystopian scenario in which that happens, but if we wind up with AI that super-intelligent, we’ll have bigger problems on our hands than keeping lawyers employed.
What I am saying is that we will need to change our understanding of what lawyers do. We will not be personally carrying out tasks like drafting motions, negotiating contracts, researching legal positions, reviewing merger terms, assessing litigation probabilities — all those activities by which lawyers have traditionally pursued client outcomes. Machines will be doing most of that work — much faster than we could, and eventually, as well as and better than we could.
This will not be a catastrophe; this will be the legal sector finally catching up with most other industries in which technology is routinely used to improve productivity. Doing all the tasks ourselves, and billing our time, will no longer be a feasible business model — and let’s be honest, it was an exhausting and miserable way to work.
Lawyers, instead, will be overseeing and managing all this machinery. It will “report to us.” We will evaluate the results, certify the documents, and synthesize the findings — and as the technology gets better and proves itself reliable, we will trust it with more advanced tasks. We will reallocate our time and redirect our focus where our value-add is highest: towards personal relationships with our clients and in furtherance of the outcomes they seek. We will summarize and analyze all the legal dimensions of our client’s situation, align them with and assess them against our client’s objectives, and confidently recommend the best path forward.
When you go see a doctor, she does not personally draw your blood and look at it under a microscope and test the blood herself for anomalies or deficiencies and cross-tabulate the findings against your information on file. Other people and devices do all that. Your doctor looks you in the eye and says, “I have your test results,” and she tells you what they mean, and she recommends the treatments you should follow, and she answers your questions and allays your fears. She is the point of contact between you and your health problem, and she is your personal guide towards a cure. That’s what a professional does in the 21st century.
And that’s what lawyers will do. We will be providers of good counsel and stalwart advocacy — powered by reliable technology, informed by sound legal knowledge, trained in legal risk assessment, and ultimately rooted in good character and judgment. We will sell clients the experience of successfully achieving their objective, not the time it took for us to assemble all its component parts. Our clients will trust us, and pay us, to advise them, advocate for them, and accompany them.
Don’t spend too much time worrying about how you’ll make this all happen — how you’ll switch from the billable hour to flat fees, or train your lawyers to be great counsellors and advocates, or change your ownership structure and financial model. We’ll figure this out soon enough, collectively, as a profession — trying and discarding different approaches, upgrading and fine-tuning as we go. We will all get there, because we’re all heading in the same direction.
No, what you need to be thinking about is this: When the new legal economy arrives, what will set you apart from everyone else? We’ll likely all have access to the same astonishing technology and adopt the same basic model of “trusted legal solutions guide.” What will make your offerings stand out to the people and businesses you want to serve? How will you come to their attention? What will draw them to you rather than to your competitors? What will make them want to stay with you?
This is why I advised young lawyers last time out: Learn how to cultivate relationships with clients. That’s going to be the one thing, the only thing, that will make you truly unique and irreplaceable as a lawyer: the person you are, the professional you are, and the relationships you create that will be unlike and better than anyone else could create with these particular clients.
And that’s the kicker, the twist ending to this whole story. In the new legal economy, after all this incredible new technology has changed everything around us, you will be amazed and delighted to discover that your killer app turns out to be you.
How do you see the AI revolution affecting the future of government lawyers like district attorneys, US Attorneys, or attorneys for the many government agencies at the state level like licensing and regulations, staff at city attorney offices, etc. where there are no external clients directly?