Canada lacks the leverage to push back or the ability to control
In Ottawa on Thursday, Canadian Prime Minister Mark Carney placed his country at a crossroads familiar to smaller nations throughout history: the moment when economic integration shades into dependency, and dependency into subordination. His warning concerned artificial intelligence — who builds it, who owns the data it consumes, and whose values it ultimately reflects. Canada's answer, he proposed, was not withdrawal but coalition: a middle-power strategy of building sovereign capacity and gathering allied democracies around a shared alternative to the platforms already reshaping the world.
- Canada's AI dependency is nearly total — researchers train on foreign clouds, businesses store sensitive data abroad, and the government itself runs on infrastructure it does not own.
- Only 12 percent of Canadian businesses use AI at all, leaving the country near the bottom of global rankings in adoption, literacy, and public trust in the technology.
- Carney framed the threat in geopolitical terms: AI has become a game of scale dominated by hegemons and hyperscalers, and nations without sovereign capabilities risk becoming subordinate to them.
- The government's response combines a national AI supercomputer, new data protection legislation, and free literacy programs in schools and community centers to close the gap from the ground up.
- The strategic centerpiece is coalition — pooling research, computing power, and purchasing decisions with allied democracies to build a credible alternative to dominant market actors.
Mark Carney stood before Canada on Thursday with a warning rooted in a pattern he had observed before: the way dominant powers use integration — trade, supply chains, and now artificial intelligence — as a quiet instrument of coercion. AI, he argued, was different in scale and speed. Canadian data crosses borders constantly. Researchers train models on foreign platforms. Sensitive government information sits in foreign jurisdictions. The dependency, he said, was nearly total, and it carried real risk: foreign entities shaping Canadian lives, accessing Canadian data, and tilting economic competition against Canadian firms while Ottawa lacked the leverage to respond.
The strategy Carney unveiled was ambitious without being isolationist. Canada would build a world-leading public AI supercomputer and introduce legislation to strengthen data protection. But the heart of the plan was coalition — the idea that a group of aligned democracies, pooling talent, computing power, and purchasing decisions, could together offer a credible alternative to the platforms already dominating the global AI landscape. It was a middle-power calculation: too small to compete alone, but large enough to lead if others followed.
What gave the moment its urgency was how far Canada had already fallen behind. Only 12 percent of Canadian businesses were using AI, with small and medium enterprises even further back. Carney announced a national literacy initiative — free AI learning kits for schools and community centers, courses on spotting bias and misinformation, tools to prepare workers for the economy ahead. The acknowledgment was implicit but clear: sovereignty is not only a matter of infrastructure. It depends on whether ordinary Canadians can understand and navigate the technology now reshaping their world.
Mark Carney stood before Canada with a warning that sounded almost quaint in its directness: the country's future hinges on who controls its artificial intelligence. Speaking Thursday as he unveiled his government's strategy on AI, the Canadian Prime Minister laid out a stark picture of technological subordination—one where foreign platforms could shape Canadian lives, access Canadian data, and tilt economic competition against Canadian firms, all while Ottawa lacked the leverage to object.
The concern wasn't abstract. Carney pointed to a pattern he'd observed before: the way dominant economic powers use integration—supply chains, trade agreements, now AI—as a tool of coercion against smaller nations. He'd warned about this at Davos earlier in the year. But AI, he suggested, was different in scale and speed. Most of the data feeding these systems crosses the border. Canadian researchers train models on foreign cloud platforms. Canadian companies store sensitive information in foreign jurisdictions. The government itself relies on infrastructure it doesn't own. The dependency was nearly total.
"That creates real risks that foreign entities could access Canadian data, deploy AI products that shape Canadian lives without reflecting our values," Carney said, "and tilt the playing field against Canadian firms—while Canada lacks the leverage to push back or the ability to control." The government's strategy document framed the problem in geopolitical terms: AI had become a game of scale, dominated by what it called hegemons and hyperscalers. Countries that didn't build their own capabilities risked becoming subordinate to them.
The response Carney outlined was ambitious but not isolationist. Canada would build a world-leading public AI supercomputer. It would introduce legislation to strengthen data protection and privacy. But the centerpiece was coalition-building—the idea that a group of aligned democracies, pooling research talent, computing power, and purchasing decisions, could create a credible alternative to the dominant players reshaping the global AI landscape. It was a middle-power strategy: too small to compete alone, but large enough to lead if others followed.
What made the moment urgent, though, was how far behind Canada had fallen. Only 12 percent of Canadian businesses were using AI at all. Adoption among small and medium-sized enterprises was even lower. Globally, Canada ranked near the bottom in AI training, literacy, and public trust in the technology. Carney announced a literacy initiative—free AI learning kits in schools and community centers, courses designed to help Canadians spot bias and misinformation, tools to build skills for the job market ahead. It was an acknowledgment that sovereignty wasn't just about infrastructure. It was about whether ordinary Canadians could understand and use the technology reshaping their economy.
The strategy represented a shift in how Canada was thinking about technology policy. Rather than simply adopting whatever the market offered, the government was asserting that certain capabilities—certain data, certain computing power, certain decisions about how AI would be deployed—needed to remain under Canadian control or at least Canadian influence. It wasn't a rejection of foreign technology. It was a claim that dependency itself was a vulnerability, and that building alternatives, even imperfect ones, was worth the cost.
Citas Notables
Foreign entities could access Canadian data, deploy AI products that shape Canadian lives without reflecting our values, and tilt the playing field against Canadian firms—while Canada lacks the leverage to push back or the ability to control.— Prime Minister Mark Carney
La Conversación del Hearth Otra perspectiva de la historia
Why does Carney frame this as a sovereignty issue rather than just a competitiveness problem?
Because he's seen what happens when one country controls the infrastructure another depends on. It's not just about losing market share. It's about losing the ability to say no—to refuse a platform, to protect your data, to make decisions that reflect your own values.
But Canada could just buy better AI tools from the US or elsewhere. Why build its own?
Because buying means you're always a customer, never a peer. You can't set the terms. You can't decide how your data gets used. And if there's a conflict—political, economic, security—you have no alternative.
The 12 percent adoption rate seems low. Is that the real problem?
It's a symptom of the real problem. If Canadian businesses aren't using AI, they're falling behind. But they're also not developing the expertise to build or modify AI for Canadian needs. You end up importing both the technology and the dependency.
What does a coalition of democracies actually do that Canada can't do alone?
Pool resources. A single country can't build a supercomputer competitive with what the US or China can build. But ten countries together? They can. They can also set standards, share research, and create a market large enough that companies have incentive to serve it.
Is Carney worried about the US specifically, or about AI concentration in general?
Both. He's careful not to name the US directly, but yes—he's concerned about American dominance. But the deeper worry is that whoever controls AI controls the future. He wants Canada to have a say in that future, not just be subject to it.