Machines are Coming to Law (and it’s a Good Thing)

Machines are coming to displace lawyers, and bar regulation will not stop them. The results will be good for consumers but mixed for lawyers. Superstars may be helped, but journeymen lawyers face a less favorable future.

Russ Pearce and I have detailed the present and future effects of machine intelligence in a just published article, The Great Disruption.  In this post I will summarize the way machine intelligence is going to encroach on lawyers’ practices.

Machine intelligence is powered by Moore’s law—the doubling of computer power every eighteen months. For a long a time, computers were not powerful enough to have wide ranging effects on law, but that is now changing. And once computer power invades a domain,  things can change very fast. As we note:

It is important to recognize two central propositions about the progress of machine intelligence. First, before the combination of hardware, software, and connectivity progresses to a certain point, machine intelligence represents no substitute for human activity. For example, decades after computers were invented, they presented no challenge to an average chess player, let alone grand masters. But once machine intelligence reaches a level where it becomes competitive with humans, it continues to improve, soon surpassing human skills. Second, because increases in the power of computing are exponential rather than linear, computers may be able to undertake complicated legal tasks relatively sooner than it initially took computers to do simpler legal tasks. For instance, in 2004, not a single autonomous vehicle drove farther than eight miles on a course through the desert. But before the middle of the next decade, researchers predict that driverless cars will transport passengers in highway and urban driving.

Specifically, five areas of legal practice will be transformed by machine intelligence. First, document review for litigation will be conducted almost entirely by machines using predictive algorithms. Second, legal search will get better, telling lawyers what cases to cite in which courts. Third, Moneyball will come to law: lawyers will use predictive analytics rather than hunches to foretell the course of litigation. Fourth, documents will be increasingly generated by machines not persons. Fifth, in the more distant future, machines will provide the first drafts of briefs, as they are already providing simple articles for newspapers.

The effects will be wide ranging and good for consumers, if not for most lawyers.

The areas of legal practice on the cusp of change through legal search—discovery, search, document generation of both forms and briefs, and predictions of case outcomes—comprise the bulk of tasks in many legal practices. As a result, those who engage in the routine elements of such services will face increasing competition from machines. Moreover, as machine intelligence commoditizes many aspects of law, information technology will accelerate greater transparency that will, in turn, accelerate such lawyers’ loss of market power over legal services. Most obviously, the transparency will come in the form of consumers’ increased ability to compare the prices of legal services

The larger lesson here is the technological advances generally reduce inequality in consumption by  helping the middle class and poor.   Legal services were too expensive for many of their needs previously. Innovation through machine intelligence will soon bring a quality of service that they could not previously afford. In the future, it will bring a level of service  now available only to the rich.  In many ways, we are become more, not less equal.

John O. McGinnis is the George C. Dix Professor in Constitutional Law at Northwestern University. His recent book, Accelerating Democracy was published by Princeton University Press in 2012. McGinnis is also the co-author with Mike Rappaport of Originalism and the Good Constitution published by Harvard University Press in 2013 . He is a graduate of Harvard College, Balliol College, Oxford, and Harvard Law School. He has published in leading law reviews, including the Harvard, Chicago, and Stanford Law Reviews and the Yale Law Journal, and in journals of opinion, including National Affairs and National Review.

About the Author

Comments

  1. gabe says

    And the best part is, perhaps:
    Machines will not be driven by either “faculty lounge” bias or the need to “stand apart” and “make one’s bones” (apologies to the Godfather saga, of course).
    I wonder, however, will their programmers program in sufficient “cleverness” such that we may continue to distort the plain text of the Constitution.

  2. dr. james willingham says

    And the machines, along with the computers, automation, robotics, and etc., have made man superfluous. We must also remember that calls for the reduction in the population of the world dates back to, at the very least, Malthus, in the late 18th century. O yes, and does this mean that the Guidestones in Georgia are a real warning, pointing the way to the end of 5.5-6.5 billion people? In the Summer of ’83 our son attended a school supposedly for bright kids on a two week scholarship. during those two weeks he worked on a computer, mentored by a twenty year old African American College student. One day there was a quiz on the computer in which this question was asked: “If you were an official in a world government and had an overpopulation problem with a country in Africa, how would you handle it? a. Have a war and kill them off. b. Use an infectious agent, germ or disease and kill them off. c. Let them starve.” Our son had little interest in that kind of stuff, and we were utterly dumbfounded. The school offered him a scholarship (and why not with an IQ of 142?), but he was not interested. Neither were we. In fact, we were outraged. Putting questions like that to an eleven year old? Later, we would hear the question was on a State Department Exam. In any case, we have known of a conspiracy for many, many years, but it was still a surprise to find a school that recruited and trained beginning grade school and high school. Actually, I googled the kind of school and found that there were twelve in as many states.

    The folks who run things are about to the point, apparently, where they feel it is necessary to get rid of the excess population. Getting rid of the “useless eaters” as H.G. Wells called them in his book, The Open Conspiracy, seems to lead to the road of security for some folks wealth, etc. In fact, the reverse is true, Society becomes what it reflects from inside man. One Puritan identified it well in a sermon on the text, Eccles.9:3 which he titled, Mystical Bedlam, or as we would call it a Spiritual Psychiatric Ward or A Spiritual Insane Asylum. When children with the best of advantages and raising and wealth, etc., can murder parents out of pique and leave their bodies rotting in the house, when spouses murder one another and even their children, when people in positions of responsibility take advantage of their authority to satisfy their sexual addictions with reference to children, holding them as slaves for years, when folks promote and support war in order to keep their power or to increase their wealth, we know we have a situation that only God can handle with a visitation of grace and justice.

  3. djf says

    This wave of automation that is about to wash over us just underscores how very, very much America needs . . . more immigrants!

  4. Devin Watkins says

    Some of the areas you talk about, I agree computers are going to be helpful in. Document review will stop being document review as we know it today, algorithms will be created to do more of it, faster, and more accurately. No doubt computers have already made legal search easier, and that will only continue. And predictive analytics of computers will be used more and more often (as they are in all businesses today).

    Document creation itself is hard, the problem is knowing which questions to ask. Machines can do the simple stuff (like wills and simple contracts). Its hard to say if computers will be able to expand on this or not.

    The actual creation of briefs, in their entirety is almost a AI-complete problem (you almost have to be able to simulate the entire human mind to solve it). Maybe first drafts, but they will still need a lot of work to finish for the foreseeable future.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>