Beyond the AI Arms Race: Canada’s Democratic Alternative to U.S. Dominance

In the modern age of democracy, public trust is the driver of governments and their ability to impact change. Despite this, according to the 2026 OECD Survey on Drivers of Trust in Public Institutions, 41% of citizens across 33 countries do not trust their government to use AI transparently. In a study on the public perceptions of AI, researchers found that 4 in 5 Canadians (79%) have concerns with the possible outcomes of AI. Further, the Pew Research Center found that the majority of Americans (50%) are more concerned than excited about the growth of AI. 

Despite high levels of mistrust in these technologies, both Canada and the United States have pushed ahead with AI as a key strategic priority. Whereas Canada’s AI for All strategy is centred on adoption, sovereignty, and trust, America’s AI Action Plan emphasizes deregulation, speed, and global dominance. This difference in approaches matters because it shapes not only who leads in AI but also who AI benefits, who is protected from its harms, and whether democratic institutions are equipped and capable of governing these systems. 

Originally launched in summer 2025, the U.S. government’s plan for AI is an ambitious one, aimed at “winning the AI race”, asserting America’s global dominance, outcompeting rival countries by rapidly building AI infrastructure (e.g. data centres), and projecting American power internationally. To do this, the action plan pushes for broad deregulation and major infrastructure investments, while treating AI more as an instrument of hard power than an economic asset. Under this perspective, federal AI policy is meant to eliminate barriers to AI deployment rather than shape public outcomes. 

Canada’s newly released strategy takes a different approach. AI for All prioritizes adoption as its main goal, but argues that it will succeed only if Canadians trust, understand, and benefit from AI. Its six pillars form a more democratic picture of AI governance, particularly one focusing on protecting Canadians and democratic institutions. Ultimately, this strategy treats literacy, sovereignty, and legitimacy as drivers of innovation, rather than obstacles. 

These differences reflect two very different theories of national strength. America’s model assumes that the quickest means to AI leadership is to slash regulatory barriers, ramp up private sector investments, and embed AI into national security objectives. By contrast, Canada’s model claims that national strength is rooted in strong public trust, broad adoption, and democratic values. 

This difference is absolutely fundamental to our discussion on AI. Canada does not need to copy America’s “hard power” approach to remain competitive. Instead, Canada’s advantage may be in avoiding the false choice of “innovation or accountability”. Recent court cases and investigations have shown the very real harms caused by AI companies, which in turn have severely diminished public trust in AI. When citizens do not trust AI systems, they refuse to adopt them – and the productivity gains and increased efficiency promised by AI strategies fail to materialize. Public trust is not a soft value; it is the foundation for any level of meaningful AI adoption. 

The stakes right now could not be higher, especially when mis/disinformation, political deepfakes, and algorithms driving online harms weaken our public discourse and undermine our democratic institutions. Encouragingly, the AI for All strategy identifies the real harms these threats pose to Canadian democracy and connects its policy response with online safety, election protection, and privacy reform. By contrast, America’s AI Action Plan aims to limit political and regulatory constraints by de-emphasizing misinformation as a significant risk, instead focusing on free expression and ideological neutrality as core principles for AI systems, in a time when its own political polarization grows by the day. 

But Canada should not take this as a reason to stay complacent. The current government has given little in the way of clear, tangible policy steps included in the strategy, and Canada has historically struggled with slow technology adoption, particularly among workers, small and medium enterprises (SMEs), and public institutions. Meaningful public engagement, digital media education and AI literacy, and procurement and legislative reform will help determine whether AI for AIl becomes more than just a slogan. If the Canadian government wants citizens to adopt AI with confidence, it must make these AI technologies usable, trustworthy, and beneficial to Canadians and their everyday lives. 

When looking at both countries’ approaches, there is also a lesson to be had in sovereignty. While America’s plan views AI infrastructure (e.g. data centres, semiconductors) as the engine driving strategic and global technological dominance, Canada’s strategy recognizes that sovereign compute, talent, and governance capacity are essential if Canadians are to control and own the systems they rely on rather than getting caught up in the American juggernaut. As a middle power, Canada’s sovereignty means the capacity to choose, regulate, and make adjustments on our own terms.

The question as to whether Canada can outpace the United States in an AI arms race is a non-starter: Canada cannot (and should not) try. Instead, the better question is whether Canada can build AI systems and technologies that are more inclusive, trustworthy, and aligned with democratic values that drive trust and accountability.  

Canada should double down on what makes AI adoption safe and ethical: quality digital media literacy education, transparency measures, and accountability safeguards. The implications of AI extend far beyond the innovation file, meaning that a “whole-of-society” approach is the only way to make AI trustworthy and usable, without undermining Canada’s democratic institutions, and safe for citizens to use. 

Ultimately, the lesson here is that trust matters. Ambition without trust leads only to policies that exclude citizens’ voices, and AI systems that continue to cause harm and fuel political instability. If we want AI to truly serve the public interest, Canada needs to measure success by whether citizens can use, trust, and benefit from these technologies, and whether these systems can be held accountable when they cause harm.