The fight over how fast AI labs should move is loud. The Canadian number that matters most sits at the other end of the pipe, and nobody is arguing about it.
- 52.7% of Canadian businesses told Statistics Canada they have no plans to use AI in the next 12 months.
- Among them, 79.1% say AI is not relevant to what they produce or sell.
- Slowing the frontier changes what arrives next. It does nothing for a firm that hasn’t used what already shipped.
More than half of Canadian businesses, 52.7%, have no AI plans for the next 12 months, according to Statistics Canada’s Canadian Survey on Business Conditions released 31 August 2026. Among those firms, 79.1% say AI isn’t relevant to their goods or services. That is the Canadian AI problem, and the loudest AI debate of the month doesn’t touch it.
TL;DR
- Statistics Canada reported on 31 August 2026 that 52.7% of Canadian businesses have no AI plans for the next 12 months, down from 66.7% a year earlier.
- 79.1% of those firms say AI is not relevant to them. Only 10.8% cite privacy or security, and 9.9% cite a lack of knowledge.
- Anthropic CEO Dario Amodei’s September 2026 essay asks frontier labs to pace capability gains. Pacing acts on the top of the curve.
- The Canadian gap sits at the bottom. Tools already on the market clear the bar for most small-firm tasks.
- The fix is one relevant task, proven inside a month, with the software you already pay for.
What happened with Canadian businesses and AI plans?
Statistics Canada’s third-quarter survey found 25.2% of businesses plan to use AI in the next 12 months, up from 14.5% a year earlier. The share with no plans fell to 52.7% from 66.7%. Most non-planners, 79.1%, say AI is irrelevant to them. It was also the survey’s last edition.
The trend line is real progress. In the third quarter of 2024, 71.8% of businesses had no AI plans. Two years later that share is 19 points lower.
The reasons table is where it gets interesting. Statistics Canada asked the non-planners why. Relevance swamps everything. Privacy and security concerns come in at 10.8%. A lack of knowledge about what AI can do sits at 9.9%.
Look at that table as a business owner. Fear barely registers, and cost doesn’t make the list. The dominant answer is “this has nothing to do with my shop.”
One more detail sits at the bottom of the release: this was the final iteration of the Canadian Survey on Business Conditions, which has run quarterly since 2020. The longest-running quarterly read on Canadian business AI plans just ended.
Why does the AI slowdown debate miss Canadian businesses?
The slowdown debate is about how fast the most capable models improve. Amodei’s essay proposes embedded evaluators, coordinated pacing among labs and a narrow antitrust waiver. None of that changes what a 40-person Canadian firm can already buy today, which is where the 52.7% are stuck.
In We Must Pace the Frontier, published in September 2026, Amodei argues for slowing capability gains while safety science catches up. He suggests the US government enable lab-to-lab safety talks with a narrow waiver for certain conversations. We covered the Toronto objection to that idea in our column on the AI slowdown cartel argument.
Here is the contrarian part. Both sides of that fight assume the frontier is the constraint. For most Canadian businesses it isn’t.
Statistics Canada’s second-quarter AI analysis put AI use at 19.2% of businesses overall. In construction it was 9.2%. Wholesale trade, 7.9%. Agriculture, forestry, fishing and hunting, 4.5%. Those firms are not waiting on next year’s model. They haven’t found a use for last year’s.
What does this cost a business that has no AI plans?
Nothing on the invoice, which is the trap. The cost shows up as hours. The Bank of Canada’s August 2026 Sparks article found only 8% of businesses use AI significantly in core operations, while 28% neither use it nor plan to this year.
Think about the operators inside that 79.1%. A roofing company. A seed dealer. A family trucking firm. Each one has a weekly task that eats two hours: quotes, dispatch notes, supplier emails, safety paperwork.
That task is relevant. The label “AI” is what feels irrelevant.
An estimate, clearly labelled as ours: two hours a week recovered is roughly 100 hours a year. At a loaded cost of $45 an hour, that’s about $4,500 per person. It won’t show up in any survey, which is exactly why no one fights about it.
What should leaders do next?
Pick one recurring task, not an AI strategy. Run it through a tool you already pay for, measure the hours for four weeks, and decide from your own number. Skip the slowdown debate entirely. Your constraint is relevance, and only your own work can prove it.
- List the five tasks that repeat every week and take longer than an hour. Pick the most boring one.
- Check what you already pay for. Microsoft 365 and Google Workspace plans may already include AI features. Ask your IT provider before buying anything new.
- Run the task through the tool for four weeks. Track minutes saved in a simple sheet.
- Stop reading frontier model news for this decision. A 2025 model does quote drafts and email summaries fine.
- Ask any vendor one question: show me this working on a firm my size, in my industry, in Canada.
- Keep a human sign-off on anything that goes to a customer.
The skeptic’s view
The 79.1% might be right. A lot of small firms sell physical work that software barely touches, and a four-week trial costs attention a 12-person business doesn’t have. Pushing AI onto a shop where it saves 20 minutes a month is waste. That’s a fair position, and for some firms it’s the correct one.
The weaker version of the argument is “not relevant to what we sell.” Almost every firm writes, schedules, quotes and chases invoices. Those tasks are where the hours hide. The skeptic is right about the product and usually wrong about the paperwork.
What to watch
- 28 October 2026. The Bank of Canada’s next Monetary Policy Report. Watch for any change in how it describes AI’s effect on Canadian productivity.
- Fall 2026. Statistics Canada has said the business conditions survey is ending. Watch for which program picks up the AI adoption questions, and if the “not relevant” reason survives into the new questionnaire.
- Next 90 days. Any formal commitment from a frontier lab to Amodei’s pacing proposal. If one comes, check the next Canadian adoption numbers. Our view is they won’t budge.
Our prediction: the next Statistics Canada release measuring business AI use will still show fewer than 30% of Canadian firms using AI. Frontier pacing will not move it in either direction.
FAQ
What percentage of Canadian businesses have no AI plans?
52.7% of businesses have no plans to use AI in the next 12 months, according to Statistics Canada’s third-quarter 2026 business conditions survey, released 31 August 2026.
Why do Canadian businesses say they won’t use AI?
Among businesses with no AI plans, 79.1% say AI is not relevant to their goods or services. 10.8% cite privacy or security concerns and 9.9% cite a lack of knowledge.
How many Canadian businesses use AI now?
Statistics Canada’s second-quarter 2026 analysis found 19.2% of businesses used AI to produce goods or deliver services in the prior 12 months, up from 6.1% two years earlier.
Would slowing AI development affect small businesses in Canada?
Very little in the short term. Pacing proposals target the most capable frontier models. Most small-firm tasks, such as drafting, summarizing and scheduling, are already handled by tools on the market today.
What is the easiest first AI project for a small Canadian business?
Pick one weekly task that takes more than an hour, such as quotes or supplier emails, and run it through AI features in software you already pay for. Measure the time saved for four weeks.
Closing analysis
The frontier will keep moving, slower or faster, and it will keep getting the headlines. The Canadian number that deserves the attention is 79.1%. That’s a relevance problem, and relevance gets solved one task at a time, in your own shop, by someone who knows the work.
If you run an Alberta business and want a practical starting point, the Alberta AI Advantage blueprint is built for exactly this stage. Operators talk through first projects every week on the AI Podcast.
Sources
- Statistics Canada, The Daily, Canadian Survey on Business Conditions, third quarter 2026, 31 August 2026. www150.statcan.gc.ca
- Statistics Canada, Analysis on artificial intelligence use by businesses in Canada, second quarter of 2026, 11-621-M2026010. www150.statcan.gc.ca
- Bank of Canada, Canadian businesses’ use of AI, what the evidence shows, Sparks, August 2026. www.bankofcanada.ca
- Dario Amodei, We Must Pace the Frontier, September 2026. darioamodei.com
Related reading
- Cohere calls the AI slowdown a cartel
- AI adoption in Canada, what StatCan’s Q2 2026 numbers show
- AI for business in Canada, our reference hub
Disclosure
The author has no relevant financial, advisory, or board relationships with any party named in this column.
Zak Hussein writes on AI for Canadian business owners and operators. He is the founder of AI Magazine Canada and CEO of ORKA AI.
