Getty Images/iStockphoto
The Terminator isn't likely to come from the future to wipe out humanity, but rather to eliminate human jobs – or so the steady stream of forecasts and headlines suggests.
Most recently, the United Nations last month warned of robots and artificial intelligence destabilizing the world through the mass automation of work. In response, or rather in preparation, the UN is opening the Centre for Artificial Intelligence and Robotics think tank in The Hague to study the potential threat.
The warning follows a host of reports in the same vein, such as a widely cited 2013 study from Oxford University that found 47 per cent of jobs in the United States could be replaced in the next 20 years. Closer to home, Ryerson University's Brookfield Institute for Innovation + Entrepreneurship last year estimated that automation could affect 42 per cent of Canadian jobs in the same time frame.
It's no surprise that people, no matter what their job is, are getting worried.
But experts – even some of the report authors themselves – feel the growing pessimism is missing much of the nuance around the issue. The media, they say, may very well be blowing the job apocalypse out of proportion.
"It's easy sometimes for people to grab at the sensational headline and make it more one-sided than it really is," says Sean Mullin, executive director of the Brookfield Institute. "There is an inherent bias toward looking at the downsides of technological change."
Brookfield's report last year highlighted a number of job categories – notably retail salesperson, administrative assistant, food counter attendant, cashier and transport truck driver – that were at high risk of being affected by automation.
But, as Mr. Mullin points out, that doesn't mean those jobs will be completely eliminated. More likely, people working in those occupations will give up parts of their jobs to automation in exchange for new tasks and responsibilities.
"Some of it means jobs will be destroyed, but part of it is also that the nature of jobs will change and evolve," he says.
In a more recent report in June of this year, Brookfield looked more closely at the issue using data from a similar study by the McKinsey Global Institute that broke occupations down into 18 capabilities. Among them were sensory perception, cognitive processing and natural language understanding; the idea was to predict which skills could ultimately be performed by machines.
McKinsey found that about 50 per cent of work activities qualified, but only 5 per cent of occupations could be fully automated. It also concluded that just because machines could perform those skills, they aren't guaranteed to.
Employers will have to consider technical feasibility, costs, economic benefits and social acceptance, among other factors, when deciding whether to automate a particular task. In many cases, it won't make sense to have a machine perform an activity, even if it technically can do so, the report concluded.
In Canada, a number of industries run a high likelihood of being largely automated, Brookfield has found, including accommodation and food services, transportation and warehousing, and manufacturing. Together, these categories account for about 24 per cent of Canadian jobs.
Low-risk or relatively insulated industries, including education, professional services and cultural production, however, account for more overall jobs – about 28 per cent. The rest of the nation's jobs fall between the two extremes and might therefore see partial automation.
As Mr. Mullin points out, the majority of the population is likely to see some of their work tasks automated, which can be a good thing because it will enable them to perform other, higher-level activities.
"As it makes certain parts of the economy more efficient, it'll free us up to do the things only humans can do. That's where the job growth will come from."
Looking solely at how robots and artificial intelligence will take over existing jobs also ignores the fact that other new technologies will create entirely new occupations that were previously unimaginable.
Among them are social media managers and marketers, says Phil Westcott, managing director of Filament, a Britain-based AI developer. Such vocations were impossible to conceive of 15 years ago, before Facebook and Twitter, but they are now among the fastest-growing jobs. Social media, for its part, is a blend of several technologies – smartphones, portable cameras and fast wireless networks, among others.
The merging of AI, the "internet of things" and augmented reality will similarly result in a wave of new jobs that aren't yet conceivable.
"The playing field of creativity is between those three technologies," Mr. Westcott says. "You're going to see the leap forward when you mix those three."
Experts say the real debate should instead focus on whether the expected gains from automation will be properly distributed to all members of society.
"In theory, artificial intelligence should unleash more economic activity, and it probably will at a macro level, but I don't necessarily think this will be a great thing for many people in the workplace," says Sunil Johal, policy director at the Mowat Centre School of Public Policy and Governance at the University of Toronto.
"The broader trend of the past 20 or 30 years indicates that inequality and wage stagnation are likely to be exacerbated rather than get any better as technology continues to progress," Mr. Johal says.
The conversation should therefore focus less on which jobs will be lost and more on necessary changes to social programs, from minimum wage to employment insurance to skills training, he says.
"There's a whole range of programs that need to be updated for the new reality of work."