Most Canadians look at their own work and see relative stability, then look outward and assume everyone else feels under siege.Luca Bruno/The Associated Press
Paul Glavin is an associate professor of sociology at McMaster University. Scott Schieman is professor of sociology at the University of Toronto and author of the forthcoming book I Want M.O.R.E.: Why Your Job Still Matters. Alexander Wilson is a doctoral student in sociology at U of T.
A growing share of Canadian workers think artificial intelligence is coming for their jobs. In our latest survey of 2,000 workers, 26 per cent said machines or computers would likely be doing much of their job within a few years. That’s up from 17 per cent when we first asked the question in 2023.
A meaningful increase, no doubt. But nearly three-quarters of Canadian workers still look at the tasks they actually perform each day and don’t see machines taking over any time soon.
Until you ask them about everyone else. In our 2026 survey, fielded between March 6 and April 2 with the help of the Angus Reid Group, we asked the question: How do you think most other Canadian workers would respond? The picture flipped. While about a quarter of our participants said that automation is likely to affect their own jobs, more than three-quarters believed most others would say their jobs were at risk. So most Canadians look at their own work and see relative stability, then look outward and assume everyone else feels under siege.
This could reflect the “better-than-average effect,” the tendency for people to view themselves as more capable than their peers. Canadian workers could be overestimating how unique their own work is, seeing their skills as more complex or irreplaceable than other workers. It’s a plausible explanation, and some of our other studies document this pattern, but our data suggest an incomplete one.
When we asked workers to explain how they felt about technology replacing aspects of their work, the responses weren’t exercises in denial. A helicopter pilot told us that autonomous flight is around the corner, but that people aren’t ready to fly without human supervision. A correctional officer explained that a computer can’t relate to inmates the way a person can. A legal professional offered an account of why courtroom advocacy and professional accountability can’t be delegated to a machine.
Union members from SAG-AFTRA picket outside a Paramount Pictures studio in Los Angeles in November, 2023. Protection against AI encroachment on their work was of the key issues the workers identified during their strike.Chris Pizzello/The Associated Press
Far from burying their heads, these workers are reasoning carefully about the judgment calls, interpersonal demands, and moral weight of the tasks they actually perform. These represent tasks that, for the most part, leading economists suggest are not easily automated, at least for now.
Even among the quarter of workers in our study who said automation was likely to affect their jobs, only about one in five expected job loss. Instead, the most common response was that their job would be adapted to work alongside the technology. At this time, we see no sign that workers en masse are viewing AI as a personal existential threat.
McGugan: When will the AI boom end? Nobody knows, and the market doesn’t want to think about it
So, if workers aren’t deluded about their own situations, why do they tend to assume everyone else thinks differently? The answer may have less to do with cognitive bias than with the information environment. We experience our own work in high definition. We know exactly which tasks we do, which ones require discretion, which ones a machine couldn’t handle. But we see everyone else’s work through the blur of headlines. And right now, those headlines tell a single, relentless story: AI is transforming everything.
The result is that most workers share a similar private assessment – cautious but not panicked – yet assume they’re the exception rather than the rule. Social psychologists call this pluralistic ignorance: a situation in which people share similar private beliefs but mistakenly assume most others feel differently. It’s not a new phenomenon, and it’s one we’ve studied before. In our research on job satisfaction, most American and Canadian workers report being personally satisfied with their jobs but believe most others are deeply unhappy, likely because media narratives of widespread discontent shape how they imagine everyone else’s experience.
What makes pluralistic ignorance consequential – and not just an interesting quirk of perception – is that people often act on what they think others believe. When most workers assume that everyone else sees displacement as inevitable, employers may not have to work very hard to make it seem that way. And as personal concern grows, this dynamic may get worse. Workers who are starting to feel uneasy about their own jobs are doing so in an environment where they believe everyone else is already bracing for the worst – which makes resistance feel not just futile, but naive.
A number of major tech firms have already laid off sizable portions of their work force under the banner of AI-driven efficiency. The former CEO of Twitter, Jack Dorsey, attributed major cuts at Block to AI, and Meta has signaled similar moves, with automation likely to feature prominently in how those decisions are framed.
Some observers, including Sam Altman, the CEO of OpenAI, have begun calling this “AI-washing”: invoking artificial intelligence to justify cuts more plausibly driven by over-hiring or financial pressure. In public, companies point to AI. But when the state of New York actually required them to explain why they were cutting jobs – including the option to cite technological change – not one of the 160 firms filing in 2025 cited technology as a reason.
Leadership Lab: Jobs survive, pay and purpose don’t: The quiet risk of workplace AI
AI will undoubtedly reshape work, and the rising share of workers who sense this suggests it is starting to hit closer to home. But the stories about what AI can do have moved well ahead of what most workers actually experience, and that gap may be consequential for labour relations. An “AI disruption” narrative gives employers room to justify cuts and restructure the terms of work in the name of technological progress. And it may also give workers fewer reasons to push back, especially when the narrative treats disruption as a settled fact.
Whether widespread AI displacement will actually materialize remains an open question. But the belief that it will may be all the cover employers need to act as though it’s already begun.