Price of the robots
Automation could spark a deeper migration crisis
Despite the pall of uncertainty Brexit is supposed to have cast over business confidence, and a self evidently slowing UK economy, Britain’s remarkable job creating machine continues to fire on all cylinders.
According to the latest Manpower survey of hiring intentions, employers remain generally optimistic, with more than three times as many firms saying they plan to take on additional staff in the final quarter of this year as those saying they will cut jobs.
Participation in the workforce has never been higher, and vacancies as a percentage of the unemployed are at an all-time record. Nor is this just a British phenomenon. Worldwide, there are more people in work today, even as a proportion of our ever growing population, than at any stage in history. Ironically, those often accused of destroying employment are also some of the biggest providers of the new jobs. New automated Amazon fulfilment centres at Warrington, Tilbury and Bristol are expected to create more than 3,500 jobs, even if it is fair to say that the displacement effects elsewhere in high street retailing are likely to be much bigger.
All the same, there is little evidence to suggest that technology in itself reduces demand for workers in the long term. Yes, innovation is proving very successful at displacing jobs, but thus far at least, even better at creating them.
Why then do we worry so much about the impact of automation and artificial intelligence (AI) on the world of employment? The optimists, among which with qualifications I count myself, argue that as with previous industrial revolutions, the problem will eventually sort itself out. More highly paid forms of alternative employment which today seem barely imaginable will in time come along to replace the jobs being rendered obsolete by today’s technological advances. This is what has invariably happened in the past.
But whereas it’s reasonable to think the long-term impact will be overwhelmingly positive, there is good cause for concern over the short to medium-term compositional consequences. What is more, today’s computer revolution displays very different characteristics from past periods of rapid technological and industrial change. As Adair Turner, former chairman of Britain’s Financial Services Authority, pointed out at a recent conference, once a piece of software has been created, it can be repeatedly duplicated at virtually no cost. Nearly all the innovations of the electro-mechanical age required significant physical investment, which in itself generated lots of new jobs. That may not be the case with much of today’s innovation.
More worrying still, the scope of what computers can do is expanding at lightning speed. Unlike previous generations of machine, modern computers can learn, giving rise to an entirely different pattern than hitherto seen of human substitution.
Everything from translation to medical diagnosis and now driving a truck or car is under threat. There are admittedly certain things the machines cannot do, and are most unlikely to be able to in any foreseeable future. This may seem to provide a natural constraint on their penetration of the jobs market. Carl Frey and Michael Osborne of the Oxford Martin School have identified three key bottlenecks. Any form of employment that requires creativity, social intelligence, or the powers of perception needed for complex manipulation can for the moment expect to remain immune to the robots. Unfortunately, this doesn’t give much comfort, for these skills show surprisingly little intensity in the jobs market as it stands. In any case, Frey and Osborne conclude that approximately 47pc of US jobs are vulnerable to automation.
To see what the economic, social and political consequences of today’s automation might be, it is useful to revisit Britain’s industrial revolution in the 19th century. For 40 years or more, it had virtually no positive impact on the condition of the ordinary working man or woman; if anything they became worse off. Older male workers with relatively well paid artisan skills lost out badly. The first factories were made simple enough to be tended by almost anyone, so that by 1830 around half those employed in textiles were children. Many of the displaced males ended up in lowly paid agricultural employment or other forms of subsistence work.
The parallels with today’s world are obvious. For the displaced weavers of Northamptonshire read Theresa May’s “left behinds” and “just about managing”. Eventually, everyone felt the benefits of the industrial revolution, but it took decades and there was a fierce political backlash in the meantime. Not for nothing, observes Frey, did political leaders prior to the “great escape” of the industrial revolution tend to ban labour saving technology.
The British government was one of the first to embrace the innovations of the time, taking sometimes brutal action against any attempt to halt the spread of the machines. This may partially explain why Britain’s industrial revolution was the first. It also helped that the great bulk of the population was disenfranchised, and therefore in little position to protest.
Yet it is not just the effect of automation on our own shores we need to worry about. Equally concerning, it may close the door on economic advancement in Africa, a continent with an exceptionally young, and substantially uneducated, demographic whose population is set to double over the next 30 years. Production in commodity manufacturing has to date tended to chase the cheapest labour markets, but when labour becomes costless, goods can be manufactured almost anywhere.
One of the great prizes of automation is that it should help make production more local once more. At the same time, however, it may deprive large parts of Africa of the opportunity, already grasped by much of Asia, to clamber aboard the global supply chain. The consequences are potentially chilling. “We do not know what the path to prosperity for Africa might be,” says the aforementioned Turner, “but we had better work it out, because if we don’t it will make today’s migration crisis look trivial by comparison”. As I say, the machines will one day make us all much richer; but the collateral damage along the way threatens to be considerable.
‘Machines will make us all richer but the damage along the way could be considerable’
Robots at work in one of Amazon’s automated fulfillment centres. The company says three new centres in Britain will create 3,500 jobs