Commenting the Canadian AI Strategy

I participated on two interviews on Radio-Canada’s Première Chaîne. First I joined Madeleine Blais-Morin on Ça nous regarde on June 4, then Patrick Masbourian on Tout un matin on June 5. We dissected the newly unveiled federal framework, “AI for All,” presented by Prime Minister Mark Carney and AI Minister Evan Solomon. While the strategy offers a welcome, human-centric alternative to Silicon Valley’s hyper-speculative and unregulated approach, a closer look at the official data reveals significant friction between Ottawa’s policy ambitions and Canada’s physical and economic realities.

Here is a summary of the core themes I broke down during the segment:

The Canadian AI Paradox: Canada proudly ranks 1st in the G7 for AI research publications per capita and houses a world-class talent pool. Yet, we lag behind at 20th in the OECD for business AI adoption (only 12% of Canadian companies use it), and rank a shocking 44th globally in public digital literacy. We are brilliant at inventing the future in academic labs, but struggle to adopt it in our daily economy.

The 70% Brain Drain: A sobering admission in the official report highlights that nearly 70% of Canadian-founded AI startups eventually relocate their headquarters abroad—mostly to the US. While new initiatives like the Sovereign Compute Access Fund aim to provide affordable processing power to keep local talent here, stopping the tech exodus remains an uphill battle.

The Power Grid Bottleneck: The strategy states that Canada will need 5.5 gigawatts of sovereign computing power by 2030 to remain commercially viable. However, our current flagship infrastructure projects only cover about 2.5% of that target. We want a sovereign AI future, but we literally lack the electrical outlets to plug it into—a physical reality that recently forced New York to halt data center expansions.

The Trust vs. Privacy Contradiction: Ottawa repeatedly claims that public trust is the ultimate catalyst for AI adoption. Yet, this narrative clashes directly with the government’s concurrent push for Bill C-22 (Lawful Access), which expands police surveillance over citizens’ wireless metadata. The state cannot easily market digital trust while simultaneously advancing mass surveillance frameworks.

Championing Practical, Human-Centric AI: On a positive note, the strategy excels when it highlights AI as a human collaborator rather than an automation tool meant to replace workers. Real-world Canadian examples—like AI-driven pediatric cardiac screening in Halifax or sustainable soil-mapping technology in Saskatoon—prove that our ecosystem thrives when solving tangible social and environmental problems.

Links