AI arrives in almost every conversation about Canada's productivity problem, usually pulling in one of two directions: either as a long-overdue fix for a genuinely persistent problem, or as another overhyped technology unlikely to move a stubborn national number. Statistics Canada's own data — adoption figures, application patterns and a direct caveat about causation — supports neither the confident version nor the dismissive one.
Key takeaways
Statistics Canada's own framing of the stakes is unambiguous. In a study of AI adoption and productivity in Canadian firms, the agency states that “for decades, productivity growth has been sluggish, hampered by weak business investment… declining business research and development (R&D) expenditures, and stalling growth in patent applications,” and frames understanding AI's relationship to firm performance as important precisely “given Canada's persistent productivity challenges.” (Statistics Canada, Artificial intelligence adoption and productivity in Canadian firms) This is the backdrop against which every AI-and-productivity claim in the public conversation should be read: a real, long-standing, well-documented Canadian problem that predates generative AI entirely, now being asked to explain or be fixed by a technology that has been widely accessible for only a few years.
The honest starting point, covered in more depth in the earlier article on where AI fits in everyday work, is that adoption is real but far from universal. 19.2% of Canadian businesses reported using AI in the 12 months to Q2 2026, a figure StatCan says “has tripled since the second quarter of 2024 (6.1%).” (Statistics Canada, Analysis on artificial intelligence use by businesses in Canada, Q2 2026) Looking forward rather than back, a separate release found that in Q3 2025 only 14.5% of businesses planned to adopt AI over the following year, while a full two-thirds — 66.7% — reported no plans to, and the leading reason given, at 78.1%, was that AI “was not relevant to the goods or services they currently provide,” well ahead of cost, security concerns, or the view that AI is “not yet a mature enough technology.” (Statistics Canada, Analysis on expected use of artificial intelligence, Q3 2025) A national productivity story built on the assumption of fast, near-universal adoption does not match either of these releases.
StatCan's research on employment trends during this period is careful to avoid overclaiming causation, and the caveat is worth reproducing rather than summarizing away: “It is unclear whether more recent trends reflect the advent of AI, other economic factors such as labour market adjustments after the COVID-19 pandemic, rapid demographic shifts, recent trade tensions with the United States or a combination of factors that are shaping the Canadian economic landscape.” (Statistics Canada, Canadian employment trends in the era of generative AI) The same study did find that employment “generally grew regardless of potential occupational exposure to and complementarity with AI” between November 2022 and December 2025, but flags that jobs more exposed to AI “are more likely to be higher-paying, associated with workplace pension plans, full-time and permanent,” so that any future AI-driven layoffs “could potentially involve the loss of high-quality jobs.” That is a nuanced, hedged finding, not a verdict either way on the productivity question.
The clearest structural finding in StatCan's own adoption research is that AI adoption is not randomly distributed across firms with different existing capabilities. Firms with strong “complementary capabilities—including R&D, cloud computing, data analytics, advanced robotics, and information and communications technology (ICT) training for employees—are significantly more likely to adopt AI,” with firms already using data analytics measured at “15.0 percentage points more likely to adopt AI than firms that do not.” (Statistics Canada, Artificial intelligence adoption and productivity in Canadian firms) If AI does eventually move Canada's productivity numbers, this finding suggests the effect will show up concentrated in firms that already had the surrounding digital capability to use it well — not spread evenly across the economy as a uniform lift.
The same StatCan productivity article cites several external economists' estimates of AI's potential effect: Acemoglu (2024) at a rise of 0.5% to 0.7% in total factor productivity over a decade; Goldman Sachs (2023) at up to 1.5 percentage points of annual labour-productivity growth over a 10-year period in the United States; and, specifically for Canada, Filippucci et al. (2025) at an increase of 0.4 to 1.1 percentage points in annual labour productivity growth over the next decade. These are third-party economic estimates that StatCan itself is reporting, not a Government of Canada forecast, and two of the three describe the United States, not Canada. They are worth knowing exist; they are not a number to quote as settled fact.
A news story reporting “AI use among Canadian businesses has tripled since 2024” is accurately citing StatCan's own figure. Read as evidence that Canada's productivity problem is being solved, it overreaches — the same release shows 40.0% of businesses still say AI is not relevant to them at all, and StatCan's own employment researchers explicitly decline to attribute recent labour-market patterns to AI over other plausible causes. Read instead as evidence that a real, accelerating adoption trend exists, is concentrated in specific industries and firm types, and has not yet been shown to solve or fail to solve the underlying productivity problem, the same headline is simply accurate. The debate worth having is the second one, not the first.
Related: where AI fits in everyday work, why AI improved so fast after 2022, and, for a structured way to decide where AI belongs in a specific plan, the AI strategy & roadmapping hub.
No. StatCan reports adoption figures and cites third-party productivity estimates, but its own employment researchers explicitly state it is unclear whether recent labour-market trends reflect AI at all, as opposed to other economic factors. Adoption growing is a documented fact; a realized Canadian productivity gain from it is not, in the sources this series relies on.
StatCan's own framing names weak business investment, declining R&D spending and stalling patent growth as the documented drivers of the long-standing problem — all of which predate generative AI by decades. Nothing in the cited sources supports treating AI adoption as the cause of, or sole potential cure for, a problem with that history.
StatCan's adoption research points toward firms that already had complementary digital capabilities — data analytics, cloud computing, R&D, ICT training — before adopting AI, since those are the firms measurably more likely to adopt it in the first place and the most likely to be studied in future StatCan productivity releases.
This is one page in a plain-English series on how AI actually works and where it fits in a Canadian business.
