NAFTA moved factories. AI will remove Canada’s competitiveness.

Mark Carney’s Davos message should unsettle anyone whose organization runs on U.S. technology: the rules-based order is fading and “compliance won’t buy safety.” Leverage is back.

Donald Trump illustrated that reality within days, publicly threatening a 100% tariff on Canadian goods tied to Canada’s dealings with China. Whether that specific threat lands is beside the point. The lesson is that dependence is not treated as partnership. It is treated as a pressure point.

Canada already lives with one form of dependence: roughly 70% of our exports still flow to the United States. Now we are building a second, deeper dependence, and doing it voluntarily. We are making Canadian businesses and institutions dependent on dominant platforms and AI systems, largely U.S.-based, to function.

This is not a bearish argument about AI. I’m an AI strategist, and I believe the technology can deliver exponential outcomes for productivity, education, innovation, and health research. This is a warning about dependence by default: a slide into functional subservience, where “using AI” becomes “needing foreign AI” to operate. It starts as convenience. It becomes an addiction once workflows, budgets, training, and expectations reorganize around it. Treat this as a call to arms to develop far more AI systems in Canada that represent Canadian interests and amplify Canadian capabilities.

If “NAFTA” feels like ancient history, here is the only lesson worth keeping. When a country sheds productive capability, it loses leverage, and rebuilding capacity is painfully slow. Industrial decline is rarely a single event. It is a long chain of decisions that feel rational in the moment and irreversible in hindsight.

AI is the same mechanism applied to knowledge work. It doesn’t move a plant. It moves work patterns: drafting, analysis, customer correspondence, recruitment screens, procurement language, lesson plans, triage scripts. When these “first passes” leave people and become platform defaults, a country doesn’t just lose jobs. It loses the pipeline that produces experts, managers, and founders.

Scott Galloway has a blunt way of describing the Canada–U.S. imbalance: Canada sells lower-margin inputs while the U.S. sells higher-margin finished products. His line is designed to stick: for every dollar Americans make from products shipped into Canada, they get about three dollars in shareholder value, while Canada gets about one. You can debate the math. The mechanism is the point.

In 2026, the highest-margin layer is increasingly digital: cloud, productivity suites, advertising platforms, and AI copilots embedded inside them. If Canada standardizes on foreign-owned AI as the “brain” of our institutions, we become buyers of intelligence rather than builders of it. We train those systems on Canadian workflows and import their defaults into how we communicate, decide, hire, teach, and serve the public.

Here’s the frustration: Canada is not short on talent or technical achievement. We are a hub. We just keep failing at retention and leverage.

Mozilla has long been a counter-weight to closed platforms, and it actively employs engineers across Canada working on core internet products and infrastructure, including Firefox-related roles “Remote Canada.”

Microsoft’s Vancouver presence is not a sales office. It is a major research and development hub. Microsoft itself describes “Microsoft Vancouver” as a research and development centre with thousands of engineers, and Microsoft Research also operates a Vancouver lab focused on bridging advanced AI and real-world applications.

Shopify is not just “a Canadian success story.” It is global commerce infrastructure. Shopify describes itself as “essential internet infrastructure for commerce,” used by millions of businesses in 175+ countries.

Toronto’s MaRS Discovery District is one of the country’s anchor innovation hubs, connecting startups and scale-ups to commercialization support, capital, customers, and talent.

And Canada’s universities have produced or trained world-class minds that shaped modern AI. Geoffrey Hinton is University Professor Emeritus at the University of Toronto and widely recognized as a foundational figure in neural networks. Ilya Sutskever studied at the University of Toronto (BSc/MSc/PhD) under Hinton and helped build seminal deep learning work. Even Elon Musk spent part of his post-secondary education at Queen’s University.

So the problem is not that Canada can’t build. The problem is that we keep building—and then letting the value chain, the defaults, and the long-term upside concentrate elsewhere.

The labour shock is bigger than youth unemployment. Entry-level roles get squeezed because “the AI can do the first pass.” When juniors stop getting real reps, you don’t just lose jobs. You lose your future seniors. Then mid-career work gets distorted as managers get pushed into supervising systems: vendor outputs, failures, exceptions, and compliance drift. Finally, brain drain accelerates as wealth and salary gravity concentrates where platforms and capital are deepest.

This is not CanCon 2.0. It is Main Street economics. Canada needs sovereign options so our entrepreneurs, schools, hospitals, municipalities, and SMEs are not permanent tenants in a foreign-owned digital economy. This is also an ecosystem story: if the default stack is foreign-owned, Canadian startups face a harder climb to become defaults at home, and Canadian talent trains on tools and workflows that pull opportunity elsewhere.

Leaders do not need a 200-page index to start. They need a simple dependence test:

1. List your top five core functions. How many would materially fail without AI tools today? (0–1 / 2–3 / 4–5)

2. Of the AI tools behind those functions, how many are foreign-owned? (0–1 / 2–3 / 4–5)

3. If access was restricted for 30 days due to a dispute, would you be unable to operate? (No / partially / materially)

4. Are you amplifying people or overwriting them? Amplify = humans draft/decide first. Overwrite = AI drafts/decides first. If 3+ of 5 are overwrite, you are replacing capability.

5. Are you reinvesting gains into Canadian capability, or just booking efficiency?

Canada has proven it can build globally relevant platforms. Shopify is one example. The Canadian AI ecosystem is another. But Canada also keeps exporting breakthroughs into foreign stacks. That is the retention problem in one sentence.

Action plan: fund Canadian alternatives to models and dominant platforms; expand institutional AI research on labour and platform impacts; educate Canadians on the stakes and the choices; create youth pathways to build Canadian AI solutions; incentivize switching, starting with public procurement; and require simple “platform nutrition labels” so leaders can see what they are funding and what defaults they are importing.

Canada should push hard on AI. The risk is not the technology. The risk is dependence by default. If we want to compete and retain our position globally, we need sovereign alternatives to dominant platforms and AI systems that are reshaping our labour market, our culture, and our leverage—and we need a serious plan to retain the talent that proves, again and again, that Canada is already a hub.