Reframing the AI Question in Law
“Here Come the Robot Lawyers” proclaimed a recent headline. Similar headlines abound these days as journalists highlight both the fears and hopes attending the increasingly widespread use of Artificial Intelligence (AI) in law. The looming question (read fear) in all this concerns whether AI “robots” can possibly replace the judgment and expertise exhibited by trained and seasoned attorneys.
But that question and the arguments that ensue over whether any machine can possibly supplant human intellect miss both the point and the opportunity of AI in law.
Lawyers now serve legal markets in the same manner that the old Bell monopolies served the telephone markets. Yes, your telephone operator may have been nice, but you needed a telephone operator to make long distance calls. Innovation amounted to adding a few color options to the basic black model And call rates occupied a stratosphere that only monopolies inhabit.
Let’s start with the opportunity. Clients have been telling us for years that the law firm service model is abysmal. Association of Corporate Counsel (ACC) surveys highlight year after year that lawyers don’t return phone calls, that they are difficult to reach and unresponsive. Altman Weil, BTI Consulting and others capture the repeated complaint that lawyers don’t trouble themselves to learn their clients’ businesses. And as though that weren’t bad enough, only a small fraction of chief legal officers surveyed by the ACC believe firms are willing to change their service models.
But that’s not the whole story. These complaints are about lawyers failing to act in accordance with their clients desires. But what about the legal infrastructure that underlies that? We are getting signals now that it’s woefully inadequate as well.
Take that most fundamental instrumentality of law — the contract. It has operated in exactly the same manner for centuries. Two or more parties come together on a subject and agree to an undertaking. Consideration is exchanged and, most often, that agreement is recorded, often at extraordinary length in unintelligible legalese. Every possible risk is inscribed, whether by quill and ink or word processor — and that written record becomes the only vehicle for enforcement of the parties’ undertakings.
Even a simple corporate asset acquisition is likely to generate hundreds of pages of various agreements for the transfer, for financing and security. And something as simple as obtaining a mobile phone? Dozens of pages of user agreement legalese.
If ever there was a cause for dissatisfaction, our mechanism for contracting is it. Creating and enforcing contracts is profoundly inefficient, even where the parties possess equal bargaining power. And in contracts of adhesion, where only one party truly has bargaining power, contracts only serve to obfuscate rights.
Which brings us to enforcement. Contracts are written on the assumption that the courts are available to help parties enforce their agreements. But if the courts were ever an efficient means of enforcing the trustworthiness of agreements, that day is long past. Only the most desperate and well-funded litigants now take contracts to court. The delay and expense inherent in modern litigation — where eDiscovery expenses are likely to dwarf the cost of resolving the underlying dispute — is an almost insurmountable barrier to contracting parties.
It was in this context that lawyer-computer scientist Nick Szabo began in the 1990s to describe an alternative method for securing agreements dependent on trust. He coined the term smart contract for such agreements and began to define cryptographic protocols (which have come to be called blockchain) to serve as an infrastructure for “trustless” agreements. And these would circumvent the limitations and frustrations inherent in our other means of recording and securing agreements.
Billions of dollars a year are now pouring into the fleshing out of blockchain and smart contract infrastructure. The point here is not that blockchain is a fait accompli or that it is even the right way to remedy the inefficiencies inherent in contracting. Rather, it is that even our most basic of legal instrumentalities are so woefully inadequate as to prompt clients (e.g. the big banks now funding a lot of blockchain development) to undertake a huge DIY effort to fix the problem.
The shortcomings just described, which are large enough to cause an uproar among corporate clients, are small compared to exponentially larger scale problems besetting the poor, middle class and small business individuals who seek justice in our system. The World Justice Project measures the effectiveness of the rule of law in countries the world over. A functioning legal system is essential to the delivery of the other necessities of a civil society. Without the rule of law, citizens lose access to health care, education and economic security.
WJP measures thirty seven variables in assessing how firmly the rule of law is established in various countries. In many of these, the US shines. Its governments are open and transparent, government powers are limited constitutionally, and laws are stable and accessible to the public.
But in access to civil justice — the ability of all citizens to secure representation and enforce their rights in a timely manner — the US lags far behind other developed nations. This is particularly evident with the poor, but similar problems plague even the middle class and small business. Lawyers are too expensive and too busy and courts are too clogged to serve our citizenry.
And therein is the opportunity for AI, indeed for any advanced technology. Lawyers, law firms and courts need to fundamentally alter and improve their service models. Lawyers now serve legal markets in the same manner that the old Bell monopolies served the telephone markets. Yes, your telephone operator may have been nice, but you needed a telephone operator to make long distance calls. Innovation amounted to adding a few color options to the basic black model. And call rates occupied a stratosphere that only monopolies inhabit.
It may be that corporate clients, for all their complaints about the disconnect between legal fees and value provided, have not reached the pitchfork and fire brand stage yet. They’re not burning down firms. But radical changes in the service model are nonetheless afoot. On the corporate side, the emergence of blockchain technology in aid of large financial institutions is probably the most significant change in the legal service model. It may not be have captured much market share yet, but it’s as important to the legal trade as automated looms were to weavers in Ned Ludd’s day.
On the consumer side, even more is happening. Entities like LegalZoom have invented a new infrastructure for the delivery of legal services. Simple contracts, wills, leases, and other legal instruments are available online and a small fraction of the cost of consulting a main street lawyer. And LegalZoom, RocketLawyer and others have set their sights on small businesses too.
So, where is AI in all this? Here’s the thing: for all the worrying about robot lawyers, AI is but one tool in repaving the entire legal service world. Every one of the so-called “exponential” technologies will come into play in some way — even 3-D printing (which should come to bear in patent modeling, at a minimum). In the same way that we don’t need to worry how VR headsets will play out in law practice, we don’t need to worry exactly how Watson, ROSS and other AI practice aids will be used.
Does it matter that LegalZoom may not, strictly speaking, use AI to deliver its services? No, the real transformation it presages is in the legal service model. That’s what needs our focus.
Rather than focusing on the particulars of AI, lawyers, legal technologists and others with impact on how services are delivered need to be developing a real intention to shift the practice model. Once that intention is in place, then all available technologies, as well as other tools such as modern project management techniques, should be employed to radically change the user experience of clients.
If corporate clients’ most frequent complaint is that lawyers don’t return calls and can’t be found when needed, then maybe that is an opportunity for AI, or for any other more pedestrian technology. The reception desk that served well in Alexander Graham Bell’s day may not meet the needs of today’s clients. So automate it. Maybe with AI, maybe with something more mundane.
And stop worrying about robot lawyers. Instead, start doing something about your service model. Or we might actually need robot lawyers.
