The Journal of Things We Like (Lots)
Select Page
Frank Pasquale, A Rule of Persons, Not Machines: The Limits of Legal Automation, George Wash. L. Rev. (forthcoming 2018), available at SSRN.

It’s funny that people who are so infinitely fallible consistently seek to eliminate that fallibility—to get rid of the vagaries, inconsistencies, and unpredictable nature of human decision-making. In this insightful article, Frank Pasquale exposes a recent incarnation of this effort and its effect on the future of the legal profession.  Legal futurists insist that software and new technology can edge out lawyers with a better, more efficient, and more consistent product. There is little that lawyers do that cannot be done better by artificial intelligence, smart contracts, and other block chain technologies. Governance itself will be more efficient, fair, and even-handed if we minimize the human element. Pasquale guides us through the flaws in the argument, the dangers and unintended consequences of the unbridled use of these tools. In doing so, he argues that legal futurists ignore the irreducibly human and discretionary nature of the law and he concludes with a more modest future for technology in the law.

Pasquale begins by exploring and debunking several myths about law and technology.  Promotors of the new legal technology suggest that the products can eliminate human discretion. Part of the appeal of legal automation is that it can replace bias with fairness and human error with mathematical precision. The more social scientists teach us about how bias works, the more skeptical we become about the ability of well-trained individuals to make good decisions. But this goal is illusory. Technology does not remove human choice. It merely shifts responsibility from lawyers, judges, and regulators to programmers.  In doing so it hides the human choices that are equally plagued by error and bias under the guise of neutrality and objectivity. It shifts decisions from those trained in law to those in a different discipline.

Legal futurists argue that technology solves a problem that has plagued the bar for over a century—this vaunted technology, they claim, will expand access to legal services by the poor and middle class. By routinizing and mass marketing legal information, LegalZoom and other innovators have been able to reduce costs and provide access to important services to those who couldn’t afford it before.

But, without disputing the value of products like LegalZoom, Pasquale notes that this software makes up a small fraction of the aspirations of legal futurists who seek to mechanize private legal arrangements among powerful businesses as well as governance itself. Many of the innovations will not begin to address unequal access to justice because they are designed for high-end consumers who do not serve lower-income clients at all. Further, even if legal technology is preferable to lawyers in some areas because of its power to reach underserved communities, it is no panacea: It is plagued by error and risks empowering new financial interests to influence normative policy areas. The tax software companies have, for instance, developed into a strong lobby, at times pushing for laws that create a market for their product but undermine the public interest in simplifying certain types of tax law. In addition, without the proper legal training and specialized knowledge, programmers can miss essential components of the law. The wills that LegalZoom drew up, for instance, neglected the effect of employer-sponsored retirement accounts, leaving many individuals with no provision for the distribution of those funds.

Pasquale also exposes a third myth—the promise of perfectly impartial technological governance—as both unrealistic and undesirable. True, robotic law enforcement seems unproblematic in some areas. Red light cameras, for example, automatically take a picture of cars speeding through stop lights and send notices of conviction and fines to the owners. But Pasquale explains that this simple tool also has downsides. Most notably, it can shift costs from the government to individuals. It does so by imposing a difficult appeal process on those who have a viable defense. In addition, it dispenses with due process and denies the defendant the right to confront his accuser. Proponents argue that the genius of innovations like this is that they eliminate unwanted bias replacing it with a fair and impartial machine, but studies have shown that algorithmic risk assessment and sentencing regimes are not immune to racial and other biases, but instead integrate them into the system. Robotic law enforcement spawns market reactions like donotpay.com, which serves a legitimate purpose in helping innocent individuals avoid paying their fines but also encourages bad actors to game the system, which leads inevitably to more technological innovations to monitor individual behavior.

Another unrealistic promise of legal futurists is that technology can displace discretionary value judgments. Block chain technologies purport to replace both intermediaries and regulatory regimes. This innovation, for instance, could allow an individual to transfer a car without waiting on the long lines at the DMV to register ownership. But as, James Grimmelmann and Arvind Narayanan have argued, block chain technology can eliminate banks, exchanges, and registries like the DMV, but without a regulatory system, run by humans, the technology itself cannot solve the problem of equity and ownership. The technology may, for instance, make it possible to transfer a car immediately by using digital signatures. It could guard against theft by making it such that the buyer’s key will not work in the car until the money has been debited from his account and credited to the seller’s. But if a hacker manages to obtain someone’s key, we still need laws, norms, and human judgment to resolve the dispute.

Ultimately, computers cannot capture the “messy complexity of discordant human meanings” that define our physical reality. (p.26.) Equity requires human judgment, a nuanced, culturally sensitive understanding of meaning. As Pasquale recounts, in the 1990s, the first computerized chess program beat a grandmaster. By the 2000s, no grandmaster could beat a computer in a game of chess, but a grandmaster and a computer together can still beat a computer alone. The addition of human understanding, judgment, and the nuanced play of meaning is essential.

Pasquale speculates that part of the enthusiasm for technology is driven by investors. In addition, legal futurists may be motivated by the age-old desire of one part of the profession to assert superiority over another. No one likes to be associated with the image of the old guard, resistant to change and insistent on the way things have always been done. Legal futurists tend to cast those who are skeptical about technology in this way. Old, out-of-touch, stodgy, and perhaps even self-interested, these lawyers will do anything to preserve their world which is quickly disappearing whether they like it or not.

But, ironically, it is the legal futurists who are clinging to an outdated view of the law. They assume that the law is a set of clear rules established by a recognized authority prior to their application. Underlying their worldview is an understanding of law comprised of clearly ascertainable edicts that determine appropriate conduct and specific legal outcomes. But this formalist conception of the law was debunked over a century ago. Even in seemingly simple arrangements, law is full of ambiguity and its application is indeterminate and inevitably dictated by the views of the decision-maker.

Embracing a more contemporary view of law as a set of fair processes, which ensure a degree of internal logic, a connection between law and reasonableness, Pasquale argues that it is less obvious how a machine alone could surpass a human. If reasoned elaboration is the hallmark of law, then it seems more nuanced, more integrally connected to human understanding. Outcomes are both culturally dependent and linguistic in nature. If law is essentially a social institution, then it is harder and perhaps even impossible to mimic perfectly in code.

Like chess, the law is not a mathematical equation, and human relations are even more complex than chess moves. The sometimes arcane legal rules crafted and applied by humans are a suitable if imperfect adaptation to the messiness of real social interactions. Pasquale concludes that legal technology can be an important tool for lawyers. Together the lawyer, technology, and artificial intelligence may surpass the lawyer alone. But in order to realize this potential, legal futurists need to develop automation that cultivates and helps develop attorneys’ services rather than seeks to replace it.

Download PDF
Cite as: Rebecca Roiphe, The Life of the Law Cannot Be Coded, JOTWELL (June 4, 2018) (reviewing Frank Pasquale, A Rule of Persons, Not Machines: The Limits of Legal Automation, George Wash. L. Rev. (forthcoming 2018), available at SSRN), https://legalpro.jotwell.com/the-life-of-the-law-cannot-be-coded/.