We continue to overestimate the threats automation and technology pose in replacing human-filled jobs, writes Philip Cross.
By Philip Cross, March 21, 2017
Dominic Barton, head of Finance Minister Bill Morneau’s economic advisers, recently repeated the doomsday scenario that automation will wipe out 40 per cent of the jobs that exist today, raising the spectre of mass unemployment. People are attracted to such gloomy forecasts because of our insecure relationship with machines, going back to the Luddites of the 19th century. Polls show a majority of Canadians feel automation will destroy jobs.
However, technology experts are evenly divided about whether automation will eliminate more jobs than it creates. Optimists argue that it’s the nature of work and the required skills that is changing, from the “technical, classroom-taught, left-brain skills” needed over the last two centuries to skills based on human social relationships, as Geoff Colvin writes in his book Humans Are Underrated. The economic historian Deirdre McCloskey estimates that one-quarter of GDP already is earned from “sweet talk in markets and management.”
This shift actually favours what humans are good at. The neuroscientist Michael Gazzaniga concluded that “Our big brains are there primarily to deal with social matters, not to … cogitate about the second law of thermodynamics.” Instead of the emphasis on logical or repetitive processes demanded by the industrial revolution, we will use our brains for what they are made for — creativity, empathy, reciprocity, art, storytelling, all skills involving social interaction and not logic.
Employment is falling in transactional occupations (bank tellers, checkout clerks) and in production jobs, but is growing in areas based on human interaction such as health and elderly care, teaching, daycare, counselling, and recreation services.
Machines can only replace the logic/math aspect of human intelligence. But there are many other types of intelligence that machines cannot replace, including spatial, interpersonal, naturalistic, linguistic, musical and kinesthetic. Machines lack creativity, empathy, humour, taste, courage or integrity. There will always be a market for these skills and values.
A change already is underway in the occupational mix of skills used in the workplace. Employment is falling in transactional occupations (bank tellers, checkout clerks) and in production jobs, but is growing in areas based on human interaction such as health and elderly care, teaching, daycare, counselling, and recreation services. This shift requires less formal education of people, as reflected in the recent decline in the skill premium for university graduates, after steadily rising for decades. The threat to university educators might be why they generate so many pessimistic portrayals of the future economy.
Some jobs won’t be missed; the French composer Claude Debussy’s brother was a “cesspool scavenger.” Boring, repetitive jobs are best left to robots; as Henry Ford complained, “Why is it every time I ask for a pair of hands, they come with a brain attached?” Some jobs don’t require brains, which instead are best applied in the workplace to interpersonal skills, not imitating a machine. The history of automation is that the jobs it eliminates are the ones humans do not do well — heavy grunt work in the 19th century, dull, repetitive work in the 20th century, making consistently good decisions in the 21st century (such as driving).
Economists have been forecasting technological unemployment so long it’s a wonder anyone listens to such predictions anymore.
Nor will automation displace all routine jobs. Overdependence on automated production processes without human input makes them brittle and inflexible; ask anyone after they talk to an automated answering service. Over-reliance on robotic methods of decision-making led to the epic failure of judgment that caused a near meltdown of the global financial system in 2008. Ben Bernanke was refused a mortgage by a bank’s computer algorithm after leaving his job as head of the Federal Reserve Board. Machines work better when augmented by humans, not when left on their own. Computers can play chess, but they play better with guidance from an expert human player.
Forecasters also exaggerate the rate at which technological change impacts the workplace. Nine out of 10 workers today are in occupations that already existed a century ago. Some changes may seem inevitable, but still take time to arrive. The history of machine automation is that it does replace jobs, but slowly, allowing time for people to find new ways to employ their innate skills. The pace of technological change is especially glacial in government, where a bureaucratic environment of form-filling and routine regulation without creativity or empathy make it readily amenable to replacement by machine.
Economists have been forecasting technological unemployment so long it’s a wonder anyone listens to such predictions anymore. David Ricardo in the 19th century worried that machines would be “injurious to the labouring class.” Keynes misdiagnosed the source of rising unemployment in 1930, claiming “We are afflicted with a new disease of which some readers may not have heard the name, but which they will hear a great deal in the years to come — namely, technological unemployment.” President Kennedy’s advisers told him that the major domestic challenge was “to maintain full employment at a time when automation … is replacing men.” Walter Leontief in 1983 speculated workers would be permanently sidelined by technology, like horses were by vehicles. There is a deep irony that at the very moment economists worry about the secular stagnation of economic growth, we worry about jobs being threatened by secular innovation in technology. Profound innovation will create the income growth needed to boost the demand for labour.
Philip Cross is a Munk Senior Fellow at the Macdonald-Laurier Institute.