By Randy Gordon of Crowe & Dunlevy
William Butler Yeats believed that history operated according to 2000-year cycles, with one era succeeded by its antithesis. In his famous poem “The Second Coming” he perceives violent disruption at the heart of the shift from one era to another: “twenty centuries of stony sleep/Were vexed to nightmare….”
Since Yeats wrote those words, the pace of historical change has seemingly accelerated, with some alarmists predicting that humans are becoming all but obsolete. It is perhaps no coincidence, then, that gloomy prognosticators have taken the 200th anniversary of Karl Marx’s birth to summon up the dreariest aspects of what Thomas Carlyle called the “dismal science.”
Economic thinkers of a certain bent have – for at least the last two centuries – posited that capitalist societies would increasingly bear two related characteristics: first, great wealth inequality and concentration and second, the “immiseration” of larger and larger swaths of the population.
David Ricardo believed that technology was one of the culprits in that machinery substituted for labor and thereby injured laborers as a class. Marx picked up the point and extended it by observing that technology not only displaced workers but brought new workers (e.g., women and children) into the labor pool, thereby increasing the size of the “reserve army” of labor, and therefore “the lot of the laborer, be his payment high or low, must grow worse.” Thomas Piketty has recently carried the arc of history forward, holding that technological progress allows for the substitution of capital for labor and an attendant concentration of income and wealth.
“Marxism” and “communism” are loaded terms, and for that reason I’ll eschew the second entirely and set aside questions concerning the social and political utility of the first. Nor do I intend to defend the full run of Marxist economic tenets – even many ardent supporters concede that he was wrong on multiple counts.
Rather, my purpose is to look more closely at the indisputable fact that for generations technology has eliminated all manner of work and ask whether that process – now that technology is capable of performing “intellectual” tasks – will displace the work of lawyers.
The answer is, of course, that it already has in some ways. So the question presented is really one aimed at discerning degree, not kind. But first a bit of history about history.
Brian Leiter has noted that, under one view of Marx’s theory of historical change (which he ultimately concludes is “false” as a general law), all socioeconomic orders share three features.
First, there is the stage of development of the forces of production, the means by which people produce life’s needs and wants. Human labor is the most primitive (and relatively static) component of output production, one increasingly supplemented by technology, which ranges from rudimentary tools to supercomputers.
Second, there are the relations of production, i.e., any given socioeconomic orders’ distribution of property rights in the forces of production. Leiter explains that “In the classic Marxist theory, the ‘proletariat’ own only their own labor power, while the ‘bourgeoisie’ own all the other forces of production and purchase the labor power of the proletariat for a survival wage.” Nowadays, some who sell their labor power often get far more than survival wages – doctors, lawyers and business consultants come to mind. To this category of workers we will shortly return.
Third, and this one need not detain us long, there is the ideological superstructure of society: e.g., the basket of beliefs and ideas that contains things like religion and politics.
Under what Leiter calls the “orthodox functionalist” version of the theory of history, the three components interact in ways that support a given system (e.g., the ideological superstructure can be explained by its efficacy in legitimizing the relations of production).
Change happens when the relations of production hinder growth in the forces of production: “for example, at some point feudal relations of production were incompatible with exploiting forms of production made possible through steam and water power (as well as mechanical tools), thus giving the nascent bourgeoisie an incentive to overthrow the feudal relations of production in order to allow them to effectively exploit these new productive forces.” Again, this may be wrong as an explanatory scheme, but technological innovation plainly disrupts the labor status quo at any particular time.
At least since Ricardo’s time, economists and policymakers have worried about “technological unemployment.” Process innovations inevitably, immediately and directly lead to job loss. But there are indirect effects that can compensate for technological unemployment. Classical economists theorized (what Marx called the “compensation theory”) that market forces can offset losses in one sector by increasing employment in another.
For example, the introduction of motorized vehicles (a “product” innovation) put mule skinners out of work but vastly increased the need for factory labor. Ford’s introduction of the assembly line (a “process” innovation) reduced the number of man-hours required to make a car, but reduced prices such that consumer demand spiked and required more manufacturing labor (not to mention increased downstream demand for things like car salesmen, mechanics, and replacement part manufacturers).
This classical view has not gone unchallenged, and it seems to me that economic theory has no clear answer to the question of the ultimate impact of process innovation on employment. I raise the issue because the technological displacement of workers is often identified as one of the causes of the immiseration of the proletariat and the collateral polarization of society.
Marx’s thesis was widely discredited in the 20th century, as workers enjoyed an improved lot and the middle class greatly expanded. Now, of course, the thesis is back in the news as wealth has once again begun to concentrate and pundits announce a second Gilded Age (but at a time when unemployment in the U.S. is at or near an all-time low).
It remains unclear to me where all the contradictory theories and supporting facts and arguments will lead, but we probably don’t need Ricardo-Marx-Piketty to tell us that unconstrained capital follows truly equivalent low wages and technology replaces some kinds of work.
And the legal field is not immune. Some traditional legal functions have been exported (e.g., certain types of document review and research to India) or replaced with technology (certain types of document review and research by software). The open question is whether legal jobs are really in wholesale danger at a time when law-school applications are once again trending upwards. To that question we now turn.
Students of technology make a distinction between labor saving and labor displacing innovations. But, of course, the distinction is not a sharp one: Innovation can save labor, create labor, or displace labor.
Take the current dash to develop autonomous vehicles. Right now, the companies developing the technology are creating good jobs for engineers and those supporting them and collateral benefits to people like intellectual property lawyers. Once the technology is perfected and widely adopted, though, the demand for professional drivers will be sharply reduced – and one can easily imagine ripple effects.
As Camilla Hrdy has indicated, innovations come in several stripes, some of which aim to improve procedures but have nothing to do with labor reduction. She points to anesthesia, which aimed to relieve pain during surgery, not to reduce the amount of labor involved in surgery. By the same token, some innovations are labor saving without being labor displacing. Here, she points to the chair lift, which saves labor by making it easier for skiers to ascend mountains but doesn’t displace workers. So the real concern is technology that “saves” desirable labor (i.e., labor that is not itself immiserating) that would otherwise be performed by humans.
For most of the period after the Industrial Revolution, automation entailed the substitution of manual labor with machine labor. Again, much of that displacement was socially desirable in that it rid the workscape of many dangerous and dehumanizing tasks.
But now that machines are increasingly able to perform “cognitive” tasks, displacement is just as likely to impact white-collar work as blue-collar work. As Hrdy suggests (citing Oxford University and McKinsey Global studies), nearly half of U.S. occupations are “potentially automatable” over the next couple of decades. And there is a mounting fear that – unlike the past – innovation will destroy more jobs than it creates.
These fears may or may not be justified (Oren Cass, for example, argues and marshals a good deal of evidence suggesting that it is not automation – but rather industrial policy – that has eliminated jobs). But this is beside the point for our purposes. The important takeaway is that the pace of change is accelerating and all professions – including law – are likely to feel the impact.
In the last few years, lawyers have begun to fret that the legal profession might fall to artificial intelligence. The use of natural language processing and predictive software in the review of documents appears to have provoked the anxiety. A 2011 New York Times article is often cited as a bellwether of impending doom.
It is true that technology has made it much easier now than in years past to handle the massive amounts of documents produced in large-scale litigation and complicated transactions. What’s often missing in these discussions is that the massive amounts of documents (millions of emails, text messages, drafts, spreadsheets, etc.) are themselves creations of technology that did not exist a generation ago.
One might argue, then, that law firms initially reacted to the avalanche of documents by staffing up the ranks of junior lawyers, but technology and outsourcing soon enough provided more cost-effective means of shoveling the snow. So, the question becomes, is the profession really at risk from technology and, if so, where and to what extent?
Richard and Daniel Susskind posit that it’s hard to avoid the conclusion that there will be a steady decline in the need for all traditional professional workers, including lawyers. In one sense, technology has been changing law practice for years. Bartleby the Scrivener long ago lost his job to the typewriter (whether he “preferred not to” remains lost in the mists of time). Just over the course of my own practice career, law firms went from having libraries of 15,000 square feet or more to having no libraries at all – the Bloomberg, Lexis and Westlaw databases killed the need.
Dana Remus and Frank Levy recently concluded a comprehensive study of what lawyers actually do with their time and the extent to which existing technology could substitute for that work. What they found was that lawyers spend very little time on document review and a great deal of time on analysis and strategy, writing, advising clients, court appearances and fact investigation. Document review is plainly “routine” and thus subject to sophisticated computer processing. But are these other activities capable of being modelled?
The short answer is “not yet,” at least not across the board. Certainly, document review will continue to be dominated by software, but even in the due-diligence context the impact will remain moderate because software is not yet capable of drawing the sort of subtle inferences (e.g., problems in interpretation and ambiguity) that quickly spring to a lawyer’s mind.
Other tasks, like legal writing (as opposed to fill-in-the-blank form work), legal research, and legal analysis and strategy will certainly be aided – though not replaced – by recent advances in computational reasoning, especially with respect to questions reducible to statistical models.
But many core lawyer functions like advising clients, investigating facts, negotiating and appearing in court, are unlikely to feel the full bite of technology anytime soon. The reasons are many, as Remus and Levy point out, but they essentially reduce to the fact that these activities are unstructured and to some degree require emotional engagement.
At day’s end, we can safely say that technology will continue to reshape lawyers’ work. But will it replace it? I have no crystal ball, so I’ll just leave it in one wag’s words, which are truer now than when he said them many years ago: “Lawyers work hard and, like us, they’re human, many of them.”
Randy Gordon is managing director of the Dallas office of Crowe & Dunlevy, a regional law firm based in Oklahoma. He is also special adviser to the vice president of the School of Innovation and executive professor of law and history at Texas A&M University.
The foregoing should not be understood as, or considered a substitute for, legal advice. For specific inquiries, please contact Gordon, or another licensed attorney.