Robots probably won't take our jobs—for nowby Daniel Susskind / March 17, 2017 / Leave a comment
Can humans benefit from robots in the workplace? This question is commonly asked, but we often make two mistakes in answering it.
The first mistake is that, in everyday conversation about the future, we tend to talk about the different “jobs” that people do. We speak about “doctors” and “lawyers,” “teachers” and “accountants.” And we ask whether, one day in the future, they might wake up to find a robot in their place.
But talking about the future in terms of “jobs” is misleading. It encourages us to think of the work that people do as monolithic, indivisible lumps of endeavour. Yet it is clear that when you look under the bonnet of any particular job, people perform lots of different tasks and activities.
This is why statements like “doctors are safe but accountants are in trouble” are unhelpful. It is why claims like “X per cent of lawyers are at risk of automation” are misleading. Robots don’t entirely displace people from their jobs; instead, they change the sorts of tasks and activities that people might do in any job.
Take, for instance, a nurse. The sort of thing that a nurse might have done 30 years ago is very different from what a nurse might do today. Then, it might have been bedpans and bedside conversation. Today, nurses can prescribe certain types of medication and perform minor procedures. The same job, “nurse,” but made up of very different “tasks.”
The second mistake is that, in moments of optimism, we tend to talk about these machines as if they directly benefit workers, as if they necessarily make certain types of people more valuable and more important. We say, for instance, that a robot “complements” or “augments” a particular person, perhaps with a certain set of skills, in carrying out the tasks and activities in his or her job.
But this way of thinking gets things the wrong way around. Robots do not directly benefit workers; instead, robots make certain tasks and activities more valuable and important, and this might be beneficial for workers—but only if they remain best placed to perform these tasks and activities.
Think of the task of driving a car. The early “sat-nav” systems, which made it easier to navigate on unfamiliar roads, have up until now benefited human drivers. But this is only because human beings are still best placed to be at the wheel. In the coming years, this will cease to be the case. Driverless cars look set to displace people from the steering wheel. And “sat-navs” will, instead, simply make these driverless cars more efficient.
The general lesson, then, is that we need to think about the future in terms of “tasks.” We should not talk about jobs, because that is too high-level. It masks the deeper churn and change in the tasks that people might do in their work. Nor should we presume that technology directly benefits human beings. People will only gain if they remain best placed to perform tasks that become more valuable and important in the future.
In light of this, will robots benefit human beings in the workplace? Of course, it is possible—but it is far from inevitable. It requires that, as our economy grows, we not only create plenty of new tasks to be done, but that these new tasks are of a type that people, rather than robots, are best placed to perform.
For the next 15 years, it is likely that there will be enough work to be done by human beings. Today, many types of task are out of reach of automation—those that require certain types of creativity and interpersonal skill, for instance. The challenge will not be unemployment, but redeployment. We must make sure that people have the skills and capabilities that are required to perform these new activities.
But in the very long term, things look far murkier. As these machines become increasingly capable, they will continue to erode any residual advantage that human beings have in performing certain types of tasks. New tasks will no doubt arise as our economies grow, and we demand new products and services. But machines will take on swathes of this new activity, too. In the long run, there is no compelling reason to think that what remains will be sufficient to keep armies of traditional workers in well-paid employment.