Monday, July 10, 2017
Following up on Paul's follow up on Whittier Law School, my summer project has to do with the limits of artificial intelligence in the process of lawyering, and its impact on how and what we teach in law school. (The Savannah Law Review is holding a provocative symposium on September 15, 2017 entitled The Rise of the Automatons. The call for papers provoked me — see below.)
I want to connect three things here: (a) the role of "machine lawyering" referred to in the discussion about Whittier, (b) the essay by Jay Finkelstein on "skills training" particularly in transactional law to which I provided a link a couple days ago (plus the comments to the post), and (c) what I think about law schools of the future. The punch line is that I hope to reveal just how radical I am as a futurist on the last point, even if, as I descend into my own dotage, I probably don't have the energy to carry off the revolution.
Here's a teaser to get you past the break. Law school curricula, by and large, are the educational versions of the Winchester Mystery House. Sarah Winchester was the widow of the inventor of the Winchester repeating rifle. It's been disputed, but some say she believed that unless she kept building rooms onto her house in San Jose, California, she would not appease the ghosts that were haunting her. The result was a bizarre mansion, built haphazardly with no planning at all, containing doors and stairs that go nowhere, windows opening into other rooms, and stairs with irregular risers. That is an apt metaphor for standard legal education, at least beyond the first year. So if we were to redesign it from the ground up, in light of the possibility of robot lawyers, the cry for skills training, and the dilemma of law schools like Whittier, what would we do?
Robot lawyering. The connection between Whittier's dilemma and robot lawyering, as highlighted in the USA Today piece, is the impact of machine thinking, or artificial intelligence, on the number of lawyer jobs and, derivatively, the (declining) demand for law school seats. Two points.
First, I don't think the impact of automation is going to be limited to "low-level" tasks. Our dean, Andy Perlman, has it right when he says that (a) lawyers' work is going to change significantly, and (b) it ought to be more interesting. For much of my career, I did pretty sophisticated mergers and acquisition work. Some of it, quite frankly, was robotic. Other portions perhaps weren't, but I would gladly have let them been. The norm (unless it's an auction in which case the buyer bids by way of a price and a markup of the seller's agreement draft) is that the buyer drafts the purchase agreement. Nothing spoiled a day for me, as a seller's lawyer, like having to read and mark up the buyer's one-sided draft, doing things like inserting the word "material" in every representation and warranty. In a heartbeat, I would have purchased a reliable program into which I could have inserted the draft, and got back an annotation of each instance in which it was "buyer-oriented," "seller-oriented," or "fair." Indeed, that's what sophisticated AI will do: crunch trillions of bits and bytes of data based on algorithms that one could write that would identify phrases and provisions that way. Then all I'd have to do is look at the annotations and check off which ones mattered to me. (I've been out of the trenches for a while. Maybe this now exists?)
Second, far be it for me to put words in Andy Perlman's mouth, but I'm pretty sure what he is thinking about, particularly for our school, are the jobs of legal engineering that it will take to create both high-level and low-level legal technology, and to manage it. That is honorable work for those who do and those who teach it, but I confess it thrills me little more than teaching the law of secured transactions, the gun-jumping rules under the Securities Act of 1933, or the law of negotiable instruments. (I will come back to my rant about subject-matter classifications later.) What provoked me about the Savannah symposium was a teaser quote from John von Neumann. From the first footnote of my current draft (if you are interested in seeing it, contact me offline):
In short, while others focus on what robots can do, I'm going to continue focusing on what they can't. Because by the very nature of the math involved in being a computer of any kind, there are indeed things they can't do.
"Skills training." When I was conversing with Jay Finkelstein via email about his piece, I observed in effect that I thought he was right for the wrong reason about so-called skills training. To recap, Jay runs a transactional simulation offered widely across the country in which a team at one law school will negotiate an international business transaction with a team from another law school.
Here's why I think he's wrong and right in his own words. The suggestion is that these kinds of courses, offered by practitioners, will make students more "practice aware" (if not "practice ready"). The skills "thing" is the usual "drafting substantive contracts and ancillary agreements, locating optional/alternative clauses, negotiating contracts and salient provisions and, among large firms, reading a balance sheet or basic financial statements." (p. 133) To me, that is all goodness, but in the long-run, only a matter of degree removed from how we usually go about teaching the law of secured transactions, the gun-jumping rules under the Securities Act of 1933, or the law of negotiable instruments.
No, what's really transformative about Jay's program are underscored by thoughts like this: "The practice of law ... is ... like a kaleidoscope where the multifaceted pieces blend together to form a complex interactive mosaic" and "the manner in which a transactional lawyer approaches a complex business problem, including the multi-subject and multi-disciplinary nature of transactional practice which integrates legal, social, economic, and political perspectives to create win-win transactional results." Those sound to me more like something transcending skills in the same way that they transcend the pieces of legal doctrine taught in the legal academy's Winchester Mystery House course catalog.
Legal education. So we come back to legal education. I'm not going to theorize about industry consolidation here. I do remember the 1980s in the law firm business, however. The first law firms to implode, at least in our town, were the twenty-five lawyer "full service" firms that couldn't really offer the full service of the mega-firms, and didn't have the sharp focus and expertise of the boutiques. The middle got crunched. One of the anonymous commenters to my post about Jay's article asked why the highest paying law jobs are still filled with grads of the "elite" schools that pay the most attention to theory. The assumption was that it's a matter of pedigree. I teach at a school that probably lacks for pedigree but regularly sends a segment of its grads to exactly the same high-paying jobs as those elite schools. I think what distinguishes those grads is indeed their aptitude for doing what theory does. Namely, "integrating legal, social, economic, and political perspectives to create win-win transactional results" (or something like that in whatever field the grad chooses to enter).
As a long, long, long time practitioner and generalist, I continue to be amused (or something like that) by the buckets of legal education (the rooms of the Mystery House). For example, it took returning to academia to find out that "commercial law" (i.e. the UCC) is a different area than "corporate law." Within business law, there are corporate camps and "uncorporate" camps, with the latter seemingly most interested in demonstrating why the area in which they happen to write and teach is normatively superior to the other (my friend and co-author, the late Larry Ribstein, being a prime example of the latter).
In the long, long term, I think the crunchable middle will be both doctrine, as traditionally taught, and what today pass for "skills." Both, to a large extent, have the potential of being robotic. The long game is in doing and teaching what robots really can't do, or in managing the robots. I'll put aside both trial and appellate litigation and focus on everything else lawyers do. In the interim, I'd do away with a lot of classes that are merely more yammering away at segments of doctrine by way of litigated cases, reverse the classroom, and make classes ones in which you merely bring doctrine to the party along with all the other theories. (In my own area, I'd do away with the traditional business law courses, and combine with the business school to teach "Law & Finance of Business Entities" with J.D. and M.B.A. students intermingled.)
But I have the luxury of being old enough not to be around in 2070 or so when my long-term thesis gets put to the test.