Could lawyers and doctors soon be replaced by robots?by Giles Wilkes / December 10, 2015 / Leave a comment
Published in January 2016 issue of Prospect Magazine
The iCub humanoid robot at IDSIA’s robotics lab in Switzerland ©Juxi The Future of the Professions: How Technology will Transform the Work of Human Experts, by Richard Susskind and Daniel Susskind, Oxford University Press, £18.99 Four years ago, a poem called “For the Bristle- cone Snag” found its way into the Archive, a student magazine published at Duke University in the United States. Containing the less-than-immortal lines, “They attacked it with mechanical horns because they love you, love, in fire and wind,” the nine-line stanza appears to muse on the destruction of a tree. There is a reason for the poem’s clumsiness: it was written by a computer algorithm, created by an undergraduate curious to see how artificial intelligence could create something as inherently human as poetry. On the evidence of “For the Bristlecone Snag” it is too soon to predict when the TS Eliot Prize will be won by a robot. But elsewhere, few occupations can rest easy. Resistance is not so much futile as economically bad form. Ever since the Luddite machine-breaking rebellion 200 years ago, a key part of every economist’s training has been to learn to scoff at technophobes. Machine efficiency allows resources to go further so what does it matter if workers are displaced? An evolving economy generates new uses for their labour, usually tasks that are less grimy and repetitive than before. Of course it is better to be the person who invents or owns the machine than a worker who suddenly finds he or she is surplus to requirements. But these distributional trembles should be dealt with by the political class, not economists. As miners have been replaced by coal-cutting machines and bank clerks by computerised ledgers, this complacency has held firm. But the latest, information-centred, phase of industrial revolution has a more unnerving character. The replacement of human labour by machines was supposed to hit a limit. At some point, routine tasks such as scanning warehouse inventory evolve into the less routine, such as deciding a marketing strategy or designing an undergraduate course. Beyond this limit, there is meant to be a space where mankind’s skills remain irreplaceable, securely remunerated and with the humans still firmly in charge. But perhaps not. That is the most dramatic argument one might take from The Future of the Professions by Richard and Daniel Susskind, a father and son team. Their view is that what once hit cloth weavers and miners, and is now happening to cashiers and travel agents, will eventually sweep up the advanced professions. The traditional ways of working currently enjoyed by lawyers and doctors will be broken and revolutionised. This will happen through the encroachment of open IT systems and ever more capable artificial intelligence, able to make the sort of fine judgements—including those with a moral angle—that currently rely on experienced human beings. Read more: Is technology set to steal your job? Capital punishment: why new technologies are hurting us (for now) This is a bold book, not least because of its target. Despite 30 years of the information revolution, trades such as law, medicine and accountancy have remained well beyond the grasp of ordinary people, and protected from the market forces wearing away at many other industries. Contrast this with a more thoroughly disrupted industry, the media, which has seen explosive growth in consumer choice. Why should any vocation have a protective wall thrown around it? More than an audacious stab at futurology, this makes the authors’s argument an act of delicious iconoclasm. Many will find this hard to swallow. Law, medicine, architecture and education are built on human expertise. They demand years of training and require a level of trust that can only exist between human beings. For a medical diagnosis or legal opinion one needs not just data processing but judgement, empathy and imagination. Those who reach the peak of their profession correspond to Arthur Schopenhauer’s definition of genius: not merely hitting the mark that others cannot hit, but hitting one that no one else can see. “Eighteen years have passed since Deep Blue, an IBM computer, beat Garry Kasparov at chess. Now Kasparov would struggle to beat your iPhone” Law, accountancy and medicine also enjoy a level of social influence that cannot be explained merely by an appreciation of their skill. Most have evolved as closed shops, demanding specific learning and credentials. The economist would describe this in terms of monopoly power: leading lawyers, consultants and accountants are richly rewarded for their scarcity. The Susskinds put it more grandly: “The professions are at the heart of our social and working lives… Their practitioners save our lives and keep us in good health, they educate our children, counsel and enlighten us spiritually, advise us on our legal entitlements, manage our money, assist us in running our businesses, help us complete our tax returns, design our homes, and much more.” I find two strong claims of the authors easy to believe, and a third that still awaits proof. The first is that the relentless onward march of information technology is bound to result in computer systems with abilities that rival or even exceed those of advanced professionals—although the authors wisely do not put a date on this. The second is that this is clearly desirable and good for welfare. But whether this should happen just because it can is another matter altogether. Technological improvement is theoretically limitless. Eighteen years have passed since Deep Blue, an IBM computer, beat Garry Kasparov at chess. Since then, as Moore’s law has predicted, computer processing power has doubled every 18 months; now Kasparov would struggle to beat your iPhone. Four years ago, IBM’s Watson beat two of the greatest champions ever to play Jeopardy!, an American television game show demanding erudition and lateral thinking. Watson’s method was brutal: load up terabytes of unstructured data, apply probabilistic algorithms and spit out an answer. If this is what could be achieved in 2011, it is easy to foresee a time when, as the Susskinds put it, “future systems could articulate and balance moral arguments, identify consistencies and illogicalities, point out assumptions and presuppositions… and identify conclusions.” The human brain is still the most complicated structure in the universe, each of its 80bn neurons connected to 10,000 others. A typical objection is that zeros and ones can never realise human-like thought. But computerising professional work does not demand a silicon version of a human professional. In the words of one journalist, Watson won its game without “knowing” that it won. Finding new solutions to problems is a neat summary of what technological progress means. That such progress benefits us is the reason to welcome this next stage of disruption. Superficially, professionals appear willing to buy into the information revolution and the tools it has brought. Lawyers accept emailed documents and court appearances by video link. Higher education is being revolutionised by online learning through massive open courses. But it is too early to extrapolate such improvements into a future in which the machines take over the decision-making role of humans. As a feat of futurology, this book could have benefited from a more sceptical take on how monopolies evolve to shield themselves from the forces that threaten them. The Susskinds set out a model for the evolution of professional work in which the last stage is “externalisation,” where the deep practical knowledge of the professions is released to the commons. Knowledge wants to be free, and market forces will drive it that way. Yet, as they acknowledge, professions also show a profound reluctance to share. This is understandable. A legal firm does not spend years accumulating an understanding of specialist law only to give it away. As the authors themselves observe, professions do not only compute and evaluate, but also take moral responsibility: “We want another human being to have reflected, and perhaps agonised, over decisions and advice that matter to us.” But this responsibility brings the possibility of power, which can be used to resist certain ways of working. Consulting a gigantic database of legal documentation may come cheap. Generating an opinion will not. Also, by attempting to sweep up all the professions into one view the authors lose the ability to distinguish between how each of them might respond to the challenge and opportunity. Accountancy, law and medicine are not only technologically different, but enjoy differential access to market and state power. The weavers and spinners displaced 200 years ago had little choice in the matter. Other products resist this, and some forms of professional service will pull off this trick better than others. Lawyers are to some extent responsible for the rules they follow, whereas medics and architects need to master the laws of physics and biology, which can presumably be grasped by super-advanced machines. The Future of the Professions is not an easy read. For a work produced by two real humans, it is strangely bloodless in style, enumerating positions and counter-positions like medieval scholastics debating the divine. Sometimes I wished there was a machine to read the book for me. But this dry argumentation does nothing to detract from the audacity of their thesis. The Future of the Professions helps us to recognise the professions’ current methods as convoluted, self-serving rituals designed to wrap simple tasks in mystique. Their practitioners do this to enjoy the same power over the rest of us that the first scribes exercised over the wondering illiterates around them. Better, surely, for everyone to learn how to read.