Complicating the Narrative of Legal Automation
Richard Susskind has been predicting “the end of lawyers” for years, and has doubled down in a recent book coauthored with his son (The Future of the Professions). That book is so sweeping in its claims—that all professions are on a path to near-complete automation–that it should actually come as a bit of a relief for lawyers. If everyone’s doomed to redundancy, law can’t be a particularly bad career choice. To paraphrase Monty Python: nobody expects the singularity.
On the other hand, experts on the professions are offering some cautions about the Susskinds’ approach. Howard Gardner led off an excellent issue of Daedalus on the professions about ten years ago. He offers this verdict on the Susskinds’ perfunctory response to objections to their position:
In a section of their book called “Objections,” they list the principal reasons why others might take issue with their analyses, predictions, and celebratory mood. This list of counter-arguments to their critique includes the trustworthiness of professionals; the moral limits of unregulated markets; the value of craft; the importance of empathy and personal interactions; and the pleasure and pride derived from carrying out what they term ‘good work.’ With respect to each objection, the Susskinds give a crisp response.
I was disappointed with this list of objections, each followed by refutation. For example, countering the claim that one needs extensive training to become an expert, the Susskinds call for the reinstatement of apprentices, who can learn ‘on the job.’ But from multiple studies in cognitive science, we know that it takes approximately a decade to become an expert in any domain—and presumably that decade includes plenty of field expertise. Apprentices cannot magically replace well-trained experts. In another section, countering the claim that we need to work with human beings whom we can trust, they cite the example of the teaching done online via Khan Academy. But Khan Academy is the brainchild of a very gifted educator who in fact has earned the trust of many students and indeed of many teachers; it remains to be seen whether online learning à la Khan suffices to help individuals—either professionals or their clients—make ‘complex technical and ethical decisions under conditions of uncertainty.’ The Susskinds recognize that the makers and purveyors of apps may have selfish or even illegal goals in mind. But as they state, “We recognize that there are many online resources that promote and enable a wide range of offenses. We do not underestimate their impact of threat, but they stand beyond the reach of this book” (p. 233).
Whether or not one goes along with specific objections and refutations, another feature of the Susskinds’ presentation should give one pause. The future that they limn seems almost entirely an exercise in rational deduction and accordingly devoid of historical and cultural considerations.
Experts with a bit more historical perspective differ on the real likelihood of pervasive legal automation. Some put the risk to lawyers at under 4%. Even the highly cited study by Carl Frey and Michael Osborne (The Future of Employment: How Susceptible Are Jobs to Automation) placed attorneys in the “low risk” category when it comes to replacement by software and robots. They suggest paralegals are in much more danger.
But empirical research by economist James Bessen has complicated even that assumption:“Since the late 1990s, electronic document discovery software for legal proceedings has grown into a billion dollar business doing work done by paralegals, but the number of paralegals has grown robustly.” Like MIT’s David Autor, Bessen calls automation a job creator, not a job destroyer. “The idea that automation kills jobs isn’t true historically,” Steve Lohr reports, and is still dubious. The real question is whether we reinforce policies designed to promote software and robotization that complements current workers’ skills, or slip into a regime of deskilling and substitution.