Susskind organized his keynote around two concepts he believes will lower the cost of legal services and increase access to justice: the commoditization of legal services and the development of disruptive technologies. The public wants free services, he says, and clients want cheaper services.
I find it difficult to rebut these presumptions, but I think Susskind skips happily over a number of problems as if they were mere speed bumps on the way to a future filled with AI lawyer software constructs.
(If you want to see Susskind’s presentation, you can watch it here.)
In the present, Susskind thinks lawyers ought to embrace commoditization. He does not, fortunately, argue that all legal services should be commoditized. Rather, he argues that our goal should be “eliminating lawyers when they are not really needed.” Or put more softly, “make sure lawyers are only involved when their deep expertise is needed.”
Many legal services are simple. An affidavit of service, for example, could easily be prepared with a short form requesting names, dates, addresses, and document titles. You probably wouldn’t even need more than 50 variations (one for every state). Employment agreements are more complicated, but a pre-defined set of questions and a bit of if-then logic could produce a good first draft, at least. (You’ll notice, by the way Susskind was careful to talk primarily about first drafts during his presentation.) Although to serve even a single state’s client base, you would have thousands of possible outputs from a single input form. Even time-consuming tasks traditionally performed by low-level associates may be able to be passed off to computers. Document review is headed down this road right now, for example.
These are the sorts of things lawyers ought to be willing to commoditize, according to Susskind. Some should even be made free. Susskind’s tacit threat is that if lawyers don’t, others will.
I think we can all agree that some legal information and advice can be commoditized. And if it can be commoditized, it probably should be. It’s hard to justify charging a client for something that a bit of software can do at the push of a button. The trick is helping people get to the right document or determine whether their legal need matches up to a commoditized solution. Plus, in the law there are often problems that will fall outside of the programmed logic of a commoditized solution. And there are usually good-better-best options from which a client must make an informed selection.
Then, as I learned from John Mayer, the executive director of CALI and the man behind the A2J Author, the interface is often as important as the legal information and the logic behind the software. Unsophisticated users need to be able to understand what is expected of them and what to do with the output.
It is theoretically possible for skilled lawyers, UI designers, and programmers to build systems that meet these needs for complicated legal services like making and deciding whether to accept a plea offer or drafting a summary judgment brief. But this is not a trivial problem. LegalZoom has a lot of money behind it, and it is so far from replacing lawyers that it has started partnering with them. Susskind skimmed right over this problem, and while it is probably not insurmountable at some point in the future, it seems to be a much larger problem than he is willing to admit.
If we are ever going to commoditize legal services on a large scale, it is going to take more than simple web forms and online document assembly. It is going to take, as Susskind puts it, disruptive technology.
I’ve already stated my feelings about using the term disruptive in relation to the law. As long as lawyers, courts, or government are involved, technology is never disruptive in the law. It takes years to accomplish what might take months in other industries, and decades to accomplish what might take years.
In any case, the gist of Susskind on technology is this: technology will eventually lead to big changes in the way legal services are delivered. I don’t think that is controversial, actually.
If Moore’s Law holds, according to Susskind, in 2020 the average desktop computer will have processing power comparable to the human brain. By 2050, all of humanity. (This does not mean computers will be able to think, but artificial intelligence at some level is at least likely, if not inevitable.)
We are going to be able to do some awesome things with that kind of processing power.
So of course computers will eventually replace some lawyers performing some legal tasks. They already do, with varying degrees of success, and their levels of influence will only grow. The complexity of the tasks we can perform will grow with it. But we are nowhere near full-on AI lawyers, yet, and real lawyers will not have any easier time assembling the logic for, say, a universal settlement agreement. No matter how powerful the computers, they will still be limited by the ability of the human beings behind them. The forms in the appendices of CLE materials will be turned into software and built on, but it will take a long time to get anything that can cope with any situation.
I’m willing to accept that AI lawyers will eventually replace some — perhaps many — lawyers. Of course, we may be well on our way to interstellar travel by that point, as well. So while it’s an interesting topic for discussing around a water bong, I’m not sure professionals need to worry too much about. We should be more worried about people attempting to make money from bad web forms drafted by lawyers who don’t know what they are talking about.
Since the UK’s Legal Services Act took effect, some banks have begun providing legal services at branches. Lawyers provide the services, and Susskind says a survey of the public indicates they are at least as interested in getting legal services from a bank or supermarket as from a law firm.
I’ve said before I think allowing non-lawyer ownership could increase the affordability of legal services. That post was not without its detractors, of course, but I don’t think we will get large-scale innovation without non-lawyer capital. I also think it is possible to overcome the ethical concerns that come with non-lawyer capital without stifling that same innovation. But I also think the UK’s Legal Services Act has not been around long enough to provide an instructive example, and I would like to see where it goes.
As a sort of corollary to this line of reasoning is this: non-lawyers could do a lot of things better than lawyers, and lawyers shouldn’t be doing those things. Susskind used a senior partner whose job is more project manager than litigator. He says we should stop training lawyers to do 1-on-1 consultative representation and start training them to provide legal services the market wants, in the way it wants.
So what does all this mean for lawyers who are going it alone or in small firms? Susskind has the answer for the near term, of course. In the UK, where 40% of lawyers are solo practitioners, he says he thinks most will just have to join the banks and supermarkets providing legal services from the teller’s desk or checkout lane. In the future, some litigation associates may be lucky enough to keep on lawyering, but the rest will, I suppose, be junior associates to AI partners, or else administering AI server farms. There is no future for the solo, in other words. Or for human lawyers in general, to take his talk to its logical conclusion. The end of lawyers, indeed.
Before we go much further along the road Susskind wants the legal industry to follow, we have a lot of important questions to resolve, and Susskind is not as interested in those. He focuses primarily on the technology, and not at all on the “little details” like, to start with, ethics. Not just formal professional ethics, which can be changed, but big-picture ethics like whether we should send people to jail or hold them to wills based on information output by an algorithm.
Currently, a lawyer provides two things: a theoretical guarantee of quality through bar passage, and an actual person to sue if something goes wrong. And, of course, a license to practice law. At some point, technology-based legal information/advice will have to be considered the practice of law. When that happens, who is licensed and liable for the advice? Will AIs have to sit through law school, then take the bar exam?
Another question: who is clamoring for these changes? I’ve long argued that LegalZoom’s success is a poor indication of demand for less-expensive legal services. It is a great indication of the number of people who will not pay for a lawyer, but LegalZoom’s products are not comparable to the services of a lawyer. People who will not pay are only a compelling measurement of demand if someone else will pay for their attention.
Susskind did not talk about who would provide all the free legal services he envisions, or how they would make enough money to do it. Surely non-profits will, since that is just why many are here at this conference, and many are already doing it. And lawyers are likely to continue using free forms and legal services as a marketing tool. Could advertising subsidize an online legal services provider? Surely it is possible. Bar associations?
I don’t disagree with any particular point Susskind makes. His argument for the commoditization of tasks that don’t really require lawyers is sensible, and people are already doing it. I mentioned CALI’s A2J Author project, which is a great example of legal forms built by competent lawyers to assist pro se litigants. It’s a good idea, and there’s no reason why solos and big firms alike couldn’t use similar tools to reduce costs and better serve their clients.
But when Susskind starts talking about what will happen in the future, his talk would be more at home at a sci-fi convention. That doesn’t mean IBM won’t build Watson 2 in a couple of years and put it to work on document review, but AI replacements are little more than fantasy.
Read the next post in this series: "Clay Shirky’s #LVI2012 Keynote: Authority in an Age of Open Access."