You’ve probably heard of the “China shock,” the mortal blow China’s entry into the World Trade Organization delivered to American (and to some extent, Canadian) low-skill manufacturing. But what about the “clerical shock” from when computers and software wiped out a similarly large number of jobs? I hadn’t until Google’s chief economist told me about it.

“You haven’t heard of [it], nobody has heard of [it],” said Fabien Curto Millet. “The reason nobody frets about it is that it’s kind of a diffuse shock. It happens across the entire economy, and the people affected have easier ways to on-ramp to other opportunities. They don’t have to up and move across the country and leave their communities.”

Let’s back up a bit. Google offered an interview because Curto Millet was in Canada for Toronto Tech Week. I was in Canada for Tech Week too, but not Toronto. No problem, the Google people said, we’ll do a video call. I pointed out that we could do a video call anytime. True, the Google people said, but Curto Millet doesn’t do interviews just any old time. False scarcity. It’s a powerful force.

False scarcity also describes how Curto Millet thinks about the “jobs apocalypse.” He’s been engaged in a back-and-forth with Stanford University economist Erik Brynjolfsson over the latter’s contention that artificial intelligence is already taking work from early-career workers. I’ve wondered the same. Curto Millet insists there are more important variables at work, such as the effect of higher interest rates.  

“The ‘jobs apocalypse,’ it’s regrettable that that’s become part of the zeitgeist,” Curto Millet said. “It’s not well supported in evidence.”  

Statistics Canada published some new evidence on Friday. If the apocalypse is coming, it hasn’t yet begun. One harbinger will be youth unemployment, because AI is so good at entry-level tasks. The employment rate of Canadians aged 15 to 24 is down significantly from 2023, but has mostly been climbing this year. 

“We are deploying [AI] aggressively internally,” John Graham, chief executive of the Canada Pension Plan Investment Board, said in an interview last month. “We still think that we should have a lot of junior employees because they’re the ones that are doing things with it that we never thought you could do with it. It has not impacted how we think about the workforce at all.” 

The other harbinger—the demise of white-collar work—also hasn’t appeared. Employment in professional, scientific and technical services was 6.8 per cent higher in May than in March 2023, when OpenAI supercharged the AI race with the release of ChatGPT-4, compared with an increase of 4.5 per cent for overall employment. 

Employment in goods-producing industries was only 1.3 per cent higher, making you wonder if a wave of AI-driven automation might actually take place in fields, factories and mines. 

Maybe the best we can do at this stage is make sure we’re thinking clearly about how AI could affect employment. That’s why I started with the China shock. Technology disrupts differently than trade, especially if that technology is general-purpose. It will make strong performers stronger, but will also offer weaker performers a chance to level up. Technology caused clerical jobs to disappear, but overall employment still grew. It probably wasn’t always easy, but those workers found other jobs. 

Trade shocks hit harder because they make companies and even entire industries uncompetitive almost overnight. The social and political dimensions are amplified because those companies and industries often are the beating hearts of communities and regions. 

If a white-collar worker in Toronto loses their job because their employer decided to replace them with a chatbot, they can probably find work with an employer who places a higher value on human intelligence. But if a blue-collar worker in St. Thomas, Ont., loses their job because of U.S. automobile tariffs, they will have a harder time finding another one because it will take time for local manufacturers to figure out how to compete, assuming a way can be found.  

“Not enough people appreciate how [AI] is a very, very different kettle of fish from trade shocks,” Curto Millet said. 

Something else to consider is whether AI is in the same league as general-purpose technologies (GPTs) such as steam engines, electricity and computers. The Bank of Canada isn’t yet ready to declare we’ve begun the next technological revolution. External deputy governor Michelle Alexopoulos said in a speech last month that AI could top out as a standard technology, such as barcode scanners. 

“If AI proves to be more of a run-of-the-mill innovation, then it won’t lead to massive productivity growth and it won’t radically transform the economy,” Alexopoulos said. Graham was similarly resistant to the hype. He described adoption by staff at CPP Investments as “tremendous” and said that as a result, the fund is “probably” more productive. But the country’s most important investor said he hasn’t yet spotted the “killer app” that would convince him to go all-in on the technology. “It’s still a little bit of an open question,” he said. 

You won’t be surprised to learn that for Google’s chief economist, the question is settled. Curto Millet is convinced that AI is a general-purpose technology and that its impact will be the same as other GPTs: some jobs will be lost, but more will be created. The scope of the federal government’s AI strategy suggests Prime Minister Mark Carney agrees. Let’s hope this time isn’t different.

Kevin Carmichael is The Logic’s economics columnist and editor-at-large. He has spent more than two decades covering economics, business and finance for outlets including Bloomberg News, The Globe and Mail and the Financial Post, where he also served as editor-in-chief.