Almost never as a clean swap for the whole job — far more often one task inside it, while the rest of the role stays with the person doing it.
Short answer
Rarely a whole role outright. Canada’s own federal definition of AI is written around tasks, not jobs, and Statistics Canada’s newest employment analysis found that job growth continued through the generative-AI period even in occupations most exposed to it — while cautioning that a task being technologically replaceable is not the same as an employer actually replacing it.
A 2026 Statistics Canada study on AI and Canadian employment quotes the federal government’s own working definition of the technology, and it is written entirely at the task level: the study notes that the Directive on Automated Decision-Making “defines AI as information technology that performs tasks that would ordinarily require biological brainpower to accomplish, such as making sense of spoken language, learning behaviours or solving problems.”
That is a list of tasks, not a job title. It matters here because an actual role — underwriter, dispatcher, marketing coordinator — normally bundles dozens of distinct tasks, and only some of them resemble the kind of task that definition describes.
From November 2022, when generative AI tools went mainstream, to December 2025, the same StatCan analysis by Tahsin Mehdi and Marc Frenette found that “employment generally grew regardless of potential occupational exposure to and complementarity with AI.” Coding-intensive professions grew at a similar rate to other jobs overall, though the gains concentrated among workers aged 30 to 49 while the number of coding professionals under 30 stagnated.
The study is also direct about who is actually at risk in the jobs that are more exposed: they “are more likely to be higher-paying… full-time and permanent,” so “AI-driven layoffs could potentially involve the loss of high-quality jobs” — not primarily the lower-wage roles a first guess might assume. The authors add an honest caveat worth keeping: it is “unclear whether more recent trends reflect the advent of AI” or other forces, including post-pandemic labour adjustment and trade tensions.
The same study is explicit that its estimates rest on technological feasibility — whether a task could be done by AI — not on what employers have actually done about it. Employers, it notes, “may not immediately replace human labour with AI, even if it is technologically feasible to do so, because of financial, legal and institutional constraints.”
That gap between could and did is where most of a role survives even after several of its individual tasks become technically automatable. For the specific tasks changing first inside Canadian workplaces, see which work tasks AI changes first; for the wider jobs debate, see will AI replace Canadian jobs.
See how to sequence which tasks to change first, and in what order, before touching a whole role.
