When it comes to robots usurping human jobs, most libertarians will point to history to prove that they are no threat to overall human employment. However, in NewScientist (4 March 2017, paywall) disagrees:
Those who back the robot revolution often point out that previous upheavals have always created new kinds of jobs to replace the ones that have gone extinct. However, one important pressure valve might not work this time. Previously, when automation hit one sector, employees could decamp to other industries. But the sweep of machine learning means that many sectors are automating simultaneously. So maybe it’s not about how many jobs are left after the machines are done taking their pick, but which ones.
The evidence so far suggests they might not be very satisfying. For example, we have seen the rise of the “gig economy”, in which algorithms direct low-skilled human workers. While this is an employer’s dream, it is frequently an insecure, unfulfilling and sometimes exploitative grind for workers.
Sumit goes on to discuss the taxation of robots and the possible use of UBI to cushion the plight of displaced workers; perhaps robot taxation could be used to fund UBI. But his suggestion that the robot onslaught will hit all sectors simultaneously is both interesting and somewhat suspect. Why? Because robots don’t come equipped ready to take over a job. They require specialized training, even more so than the more advanced humans they are replacing, because they have no instincts and no general base of knowledge to fall back on. However, once a training result is established, it’s far more easily replicated than it is among humans; and, of course, they’re cheaper, on average, than humans, and if an accident occurs, there’s little chance of being sued just because a robot is wrecked.
At least not until the robots are artificially intelligent and can value their own existence.
