When Prime Minister Mark Carney unveiled Canada’s new AI for All strategy, he framed artificial intelligence (AI) as a driver of economic growth, innovation and national prosperity. Yet AI is also increasingly shaping geopolitical power. While the strategy, released June 4, recognizes Canada’s dependence on foreign technology infrastructure, it largely overlooks the international and security dynamics that will shape whether its ambitions succeed.
Former US National Security Advisor Jake Sullivan recently told an Ottawa audience that Canada and the United States are bound by three destinies: geography, democracy and technology. He was describing a reality, but also a vulnerability. Canada’s AI strategy seeks to reduce dependence on American companies, computing infrastructure and technology ecosystems at a moment when the unpredictability of the current US administration makes that dependence harder to manage. Despite that, the strategy misses the broader geopolitical picture.
AI for All reflects genuine priorities. Its focus on economic competitiveness, Canadian sovereignty over data and infrastructure, and building public trust in AI systems all matter. However, it largely treats AI as a domestic economic challenge at a moment when governments around the world increasingly treat it as a geopolitical one.
The United States, China and other major powers are not building AI policy primarily around productivity; they are building it around strategic advantage. The Trump administration’s recent executive order on advanced AI reflects the tension between commercial competitiveness and national security. China continues to integrate AI into military modernization and broader efforts to expand technological influence. Canada does not need to match the scale of either power — as a middle power, it cannot. But it does need to understand how that competition will shape the conditions under which its own AI ambitions are pursued.
In a race to the bottom on AI development, the risks extend well beyond economics. Strategic competition among the great powers and larger states can accelerate unsafe deployment, deepen technological dependencies and contribute to geopolitical instability, and yet Canada’s AI strategy appears to gloss over this reality.
AI for All’s most promising geopolitical idea is the expansion of the newly formed Sovereign Technology Alliance, which strengthens partnerships with Germany and like-minded countries in Europe, the Indo-Pacific and the Gulf to share AI research, standards and infrastructure. The instinct is right. No middle power achieves technological sovereignty alone.
But the strategy gestures toward this cooperation without fully operationalizing it. Canada should use these partnerships to pursue shared computing infrastructure, joint research initiatives, common standards for trustworthy AI and coordinated approaches to technology governance. It should also identify the international fora where it intends to not merely participate but lead.
Of course, sovereignty and interoperability are often in tension. As countries pursue AI sovereignty and develop distinct models, standards and technical ecosystems, there is a growing risk of fragmentation. For allies that increasingly rely on AI-enabled systems for defence, intelligence sharing and critical infrastructure, the ability to operate together cannot be taken for granted. Canada should work with partners to ensure that efforts to strengthen sovereignty do not succeed at the expense of interoperability and collective security.
What Canada needs, then, is not a security annex bolted onto an economic strategy. It needs an AI foreign policy that treats this technology as what it already is: a source of geopolitical influence as much as a driver of economic growth.
Such a policy would provide a framework for navigating alliances, international standards, export controls, access to advanced computing infrastructure, and the risks posed by military and dual-use AI systems. It would recognize that Canada’s AI future will be shaped not only by domestic investment and regulation, but also by the decisions made in capitals around the world.
Any AI foreign policy should also clarify what the government means by “Canadian values” in practice. If those values are meant to distinguish Canada’s approach from competing models, they should be clearly articulated and accompanied by a plan for how they will be upheld when economic opportunities or geopolitical pressures create incentives to compromise them. After all, values matter most when they are tested.
That framework also needs to extend trust beyond domestic policy. Canadians rightly ask whether AI systems are safe and aligned with democratic values. They should also be asking whether the infrastructure they depend on is resilient, whether strategic dependencies are being managed, and whether Canada has a plan when geopolitical tensions disrupt technological supply chains, as they inevitably will.
Managing the highest-risk applications of military AI and autonomous systems will also require engagement beyond Canada’s closest allies. As the Cold War demonstrated, stability often depends as much on dialogue among competitors as on cooperation among friends. If Canada hopes to contribute meaningfully to global AI governance, it must be willing to engage not only with like-minded states, but also with those whose values differ sharply from its own.
Ultimately, Canada is unlikely to lead through scale, but it can lead through diplomacy, coalition-building and the development of practical governance frameworks that help reduce risk while preserving the benefits of technological innovation.
The Carney government has positioned Canada as a trusted, values-driven actor in the AI era. That is a credible ambition. But credibility requires more than an economic strategy with a sovereignty dimension.
Decisions made in Washington and Beijing will shape the future of AI regardless of what Ottawa does. But as the prime minister alluded in his Davos speech earlier this year, middle powers are not condemned to be spectators. The question is whether Canada helps shape the rules, partnerships and institutions that govern AI, or simply adapts to choices made by others.
