Will AI make law better? Yes. And those who remain lawyers will enjoy more inspiring working conditions and thrive amid an abundance of artificial intelligence.
Will AI make law better?
Yes.
For whom?
For many on both sides of the legal profession’s moat.
I’ll be brief.
(If you’re looking for verbosity, see my other writings. Links to some decorate this one.)
The Pains and Pleasures of Law Practice
Many lawyers find practice dispiriting. More than a few live lives of quiet desperation. They face tedious work, punctuated by alarming stress and pathways to burnout. There’s been a lot of human wreckage. Yet others thrive on the liberating pleasures of artisanship. For them, “flow” is frequent, and they enjoy good work-play balance.
Given all the recent progress, it’s a reasonable bet that AI will finally dent the shiny fenders of our professional monopoly.
Much useful legal work will be done by software, and the authorities will no longer defensibly suppress it as “unauthorized.” (See “Liberty, Justice, and Legal Automata.”)
With much of the underlying work being doable less expensively, there will be a virtuous circle of lower prices and higher demand. The economics of private practice will become more challenging, but the ergonomics — understood in the broadest sense as the character of our work environments — may well improve.
Those lucky enough to remain employed as lawyers will enjoy more inspiring working conditions.
They will thrive — calmly confident, dedicated to craft and excellence — amid an abundance of intelligent assistance.
At least some lawyers are in for more artistry and professional satisfaction. The rise of intelligent machines could usher in an age of greater humanism, even if that involves commanding battalions of nonbiological paralegals, some with expertise greater than the commander. How to buy, rent or build your posse of artificial agents will be the subject of popular, if not required, law school courses.
On the consuming side of legal services, there’s never been much pleasure.
Most folks can’t afford — and don’t need — artisanal solutions to their legal problems. But artificial ones will be increasingly effective.
Routine
We sometimes assume that “routine” legal work will be taken over by indefatigable machines.
In 1990, I speculated about the relationship between AI and legal routines (see “Computational Intelligence and the Paradoxes of Legal Routine“). The basic idea was that most notions of routineness quickly lose coherence upon careful consideration. Our uses of that term are full of vexing contradictions. Routine work is not necessarily simple, or straightforwardly automated. Some highly non-routine work might be more algorithmic than appears. Aspects of both could effectively be reproduced in software; aspects of both remained quite elusive. I suspect that remains true in our present era, when “good old-fashioned AI” has been outshone by second-wave, new-fangled, deep learning techniques.
Togetherness
Human effort and machine assistance echo and support each other. “Toward a Phenomenology of Machine-Assisted Legal Work” charts the vast landscape of human experience that has yet to fall within the cognizance of our artificial assistants. “Dancing with Cognitive Exoskeletons” imagines a future in which people don metaphorical robotic carapaces to interact with others who are also equipped with intellectual armor.
“The Centrality of Choice in Legal Work” focuses on one area in which explicit human attention appears indispensable — namely, reaching decisions in which tradeoffs of values and evaluative perspectives play central roles. (One can argue that human involvement is an essential ingredient in any true “judgment,” at least when human goals and concerns are in play. But machines can facilitate good decision-making. See “On Balance” in the International Journal of Artificial Intelligence and Law.)
Bridges
Practice tools, like document automation and expert systems deployed in law firms and legal departments, have long bridged the realms of human and machine cognition. Large-scale nonprofit services like LawHelp Interactive, leveraging similar technologies, have served millions of lawyer-less people without charge over the past decade.
Even though we’re coming to expect computational magic without the necessity of much traditional coding, optimal tool-making still requires both design thinking and good machine tools, such as those emerging under the banner of domain-specific languages. (See “Knowledge Tools for Legal Knowledge Tool Makers” and “Domain Specific Languages and Legal Applications.”)
Flowering
The broader implications of recent AI developments pose fundamental challenges for the legal profession, the legal academy, and society. We have a historic opportunity to extend access to justice and legal wellness to the vast part of humanity presently excluded from them. But we are still in the earliest stages in terms of system quality, educational institution support and regulatory readiness. (For my thoughts on these topics, see “Substantive Legal Software Quality: A Gathering-Storm, Lawyering in an Age of Intelligent Machines,” and “Safe Harbors and Blue Oceans.”)
We’re in a liminal period — between practice as we’ve understood it and practice as it will be.
I’m optimistic that, on balance, things will be better in the world of law — for most lawyers, clients, and folks who can’t afford, or choose not, to be either.
Will AI make law better?
One thing is clear: Flourishing will require deep collaborations among artisanal humans and artificial co-workers. We’re in this together.
This article is dedicated to the Centre for Computational Law (CCLAW) in Singapore.
Collaborations with friends there have helped shape the ideas here.
Image © iStockPhoto.com.
Don’t miss out on our daily practice management tips. Subscribe to Attorney at Work’s free newsletter here >
