Is a robot going to steal your job? This economist can tell you
This time around it’s artificial intelligence vs the middle classes, economist Carl Frey tells Harry de Quetteville
In 2013, Carl Frey, a willowy young economist who grew up in the southern Swedish university town of Lund, sat down with Michael Osborne, a colleague of his at Oxford University. Frey wanted to understand how automation might affect current careers: will we all be enhanced and supercharged by technology, or demoralised and made redundant?
It is the $64,000, twist or bust question of our age. For while today’s advances – artificial intelligence, robotics, autonomous vehicles – are widely accepted as transformational, no one can say for sure whether that transformation will be for good or ill.
Frey tried to find out. And, ironically, the reason he recruited Osborne, a professor of machine learning, was to develop an algorithm to partially automate the task. The results were startling. In a list of 702 occupations, 47 per cent were deemed at high risk of disappearing in the next few years.
The paper was a sensation. And now Frey, 35, is back with a new book, The Technolog y Trap, which aims to diagnose the consequences of our own industrial revolution still further. Will we be richer, poorer, happier, sadder, more fulfilled or less, politically more or less stable? And if we don’t like it, can we do anything?
Predictions from his 2013 paper, outlandish at the time, have begun to come true. “We found that fashion models were exposed to automation,” he says, at a table in the Oxford Martin School, sheltered from the summer tourist hubbub.
“That seemed slightly ridiculous.” But the mockers then had never met Imma or Shudu, beautiful yet uncanny computergenerated mannequins who
now have huge followings on Instagram and are used in mainstream advertising campaigns. “It’s happening at a rapid pace,” he shrugs.
The sweeping effect of today’s innovative industrial processes certainly merit the term “revolution”. What Frey thinks is more important is to determine what kind of revolution. Will it be the archetype – the Industrial Revolution that began in 1769, when Richard Arkwright, a Preston-born inventor, patented his spinning frame, which automated the process of cotton spinning and destroyed the lock that middle-income artisans had on weaving?
Or will it be the consumer-driven, wage-accelerating, emancipating revolution of 20th-century America, where men in the workplace were paid better wages as automation helped them become immensely more
For 70 years ‘those who lost out to automation didn’t live to see enrichment’
productive, and women at home found time spent on domestic chores was halved by the likes of dishwashers, fridges and washing machines?
Not that the original Industrial Revolution did not immeasurably improve lives. It did. Just not, Frey points out, at the beginning. Or for a long time after. Indeed, Frey says, “three generations of Englishmen were made worse off ” while a tiny number became impossibly rich. Only from the 1840s did the wider workforce feel the rich glow of greater health and wealth. For 70 years until then, “those who lost out [to automation] did not live to see the day of the great enrichment”.
So, Frey asks, is today’s technological disruption more like the 70-year whirlwind of 1770-1840, or the largely benevolent turbulence of the 20th century? And, as a result, can we expect popular resistance to automation, as with the former, or almost none at all, as with the latter? And if popular resistance occurs, will governments make concessions to it, or crack down?
All of us stand at a fork in the road, he says, and the question we must all ask is whether modern technology is “enabling” or “replacing”?
Ultimately, he says, the first industrial revolution had unhappy beginnings because, on the whole, it replaced workers. But the 20th century’s leap forward enabled them. Automation allowed them to move from “sweat and drudgery to well-paid jobs that were less physically demanding”. Wages rose in line with productivity, and the poorest benefited most. Inequality flattened, “laying the foundations for a growing and increasingly prosperous middle class”.
Frey’s news for today is not good. Post-eighties computerisation began to erode the mass middle-class jobs created by the postwar manufacturing boom, he says. And, he notes: “After reviewing [more] recent technological developments – machine learning, machine vision, sensors, various subfields of AI, and mobile robotics – my conclusion is that they are predominantly replacing… and will worsen the employment prospects for the already shattered middle class.”
His dismal assessment is that today’s industrial revolution risks looking a lot like that of 1769: careers ended, wealth falling predominantly to a few, all softened with the promise that things will be all right in the end – though not if you are one of America’s three million truck drivers or the similar number of people in the UK working in retail sales.
And yet, he is no doom-monger. “I’m an optimist about AI,” he says. He thinks today’s tech will ultimately “give a boost to productivity and make more jobs interesting”. It is how we manage the decades of disruption until that happens – our own 17701840 – that matters. It is possible, he thinks, yet represents a challenge that is already dictating and monopolising politics: whether to engage with global competition, or to retreat and protect national industries; whether to side with disgruntled, displaced workers or to crush them.
Frey does offer some sticking-plaster solutions: dramatic tax breaks for low incomes; wage insurance for those moving to lower-paying jobs; reduced planning restrictions to make housing cheaper in booming cities; better transport to facilitate access to those cities and credit to help relocation. He hopes they can mitigate what he calls the “tragedy of technology” – that it makes us healthier and better off on average but not for everybody or, at least, not in the short run. “We need to use policy to smooth the transition.” It’s a transition, he thinks, that we’ve only just embarked upon. Unfortunately, policy is the one area where Frey feels disruption is too slow. “Tony Blair was at a recent conference,” he recalls, laughing ruefully. “He said if Clement Attlee came back to the UK he would walk around and marvel at the extraordinary, unimaginable innovation and invention. And then he would walk into Whitehall and feel completely at home, like nothing had changed. It’s probably true.”
‘AI will give a boost to productivity and make more jobs interesting’