When Canada’s Parliament was prorogued in January 2025, it took the proposed Artificial Intelligence and Data Act (AIDA) down with it. AIDA, which spent nearly three years winding through Parliament, died without producing a single enforceable rule. Although an instinctive response may be to treat this as a governance failure that should be addressed quickly, it is worth taking a step back to question not only why AIDA failed, but also whether it was the right approach for Canada.
When it comes to AI governance, the major economies Canada trades with have moved decisively to enact regulation. Canada’s recently launched “AI for All” strategy mentions privacy and online safety laws as a major pillar, and more than 200 AI-related laws have been passed around the world in the last decade. However, with each state vying for a competitive edge, only a few have managed to enact policies that guide a borderless technology. The EU AI Act is now fully in force, imposing binding obligations on safety, transparency and human oversight for high-risk AI systems; the United States has signed a handful of presidential orders targeting government and sector-specific frameworks; and China has taken a centralized approach to align AI with national values and promote public safety.
But while these superpowers continue to develop competing regulatory frameworks, the World Economic Forum is flagging that these divergent rules are already leaving their trading partners behind and hindering international cooperation in the process. Each jurisdiction has built its own documentation requirements, its own conformity assessment processes and its own audit infrastructure — none of which are designed to recognize other nations. This was acknowledged by the Government of Canada’s own AIDA companion document, which stated that Canada needs a framework to “remain interoperable with international markets,” or else be sidelined in the competitive landscape. While these concerns illustrate the difficult reality that states face in navigating global AI governance, it also reveals why a made-in-Canada regulatory regime may be a hindrance.
With hundreds of millions of domestic consumers, global superpowers have the capability and capacity to dictate and weaponize standards. If any company wants to sell to 450 million Europeans, 330 million Americans or 1.4 billion Chinese, they have little choice but to comply.
When it comes to Canada, the market is simply not large enough to compel foreign alignment at the same level. Adding another layer on top of existing requirements, whether through “AI for All” or a renewed AIDA, may not protect Canadian companies; instead, it could price them out of the markets they are already trying to reach. More than 98 percent of Canada’s technology firms are small or micro-sized, with 85 percent having fewer than 10 employees and lacking dedicated compliance teams to absorb the burden. The CEO of Canvass AI, a Toronto-based industrial AI company, emphasized this issue when testifying before the World Trade Organization’s Committee on Technical Barriers to Trade, stressing that cross-jurisdiction compliance carried one of the highest costs and go-to-market barriers the company faced.
As a result, the consequences for talent are exacerbated. When a Canadian AI start-up cannot absorb the cost of entering international markets, decisions about where to incorporate become decisions about where to build careers. A study of nearly 3,000 venture-backed start-ups founded by Canadians between 2015 and 2024 revealed that among high-potential start-ups launched in 2024, only one-third remained headquartered in Canada. An environment where scaling internationally is harder than it needs to be presents a case where compliance costs are not justifiable compared to larger markets. The alternative, however, is not deregulation: it is coordination. The biggest friction in global AI governance is not that major economies disagree on values but rather that each has built its framework in isolation, despite applying it to a technology that is inherently global. A Canadian company that already complies with partners such as Europe or South Korea has, in practice, already demonstrated those values.
Coordination across nations means building the infrastructure that connects these processes. Verifying international regimes provides a vital assurance that different companies are meeting the same standards in the same ways across jurisdictions. Establishing joint sandbox agreements where companies test AI systems against multiple frameworks simultaneously, mutual recognition pathways that allow a conformity assessment conducted in one market to count toward another, or shared documentation templates acceptable across jurisdictions are all prospective initiatives that would simplify compliance for Canadian firms while upholding safety standards across borders.
More importantly, Canada is exceptionally well placed to make it happen. Through its 2025 G7 presidency, Canada opened the discussion to advancing AI cooperation across member states and built directly on the multilateral framework that Japan had established through the Hiroshima AI Process. The Canadian government followed this with the first meeting of the Canada-EU Digital Partnership Council, which focused directly on AI coordination and reducing barriers for small and medium-sized enterprises on both sides of the Atlantic. On the regulatory side, Japan’s AI Promotion Act explicitly lists international cooperation as a core principle and includes provisions for regulatory sandboxes, while the United Kingdom’s Digital Regulation Cooperation Forum has demonstrated in practice that cross-regulatory support reduces compliance costs, saves time and gets companies to market faster. The relationships are there, the frameworks are there and the precedents are there — all that is left is the political will to act on them.
However, this does not mean that Canada shouldn’t have its own AI governance, and it certainly does not mean that the AI strategy isn’t necessary. Sovereignty over how AI is developed and deployed within Canadian borders is legitimate and important. There are areas where Canada’s specific context genuinely demands domestic rules: online harms, Indigenous data rights, French-language obligations, federally regulated sectors such as banking, telecommunications and government use of AI, to name a few. These are genuine reasons for a Canadian framework, but they are arguments for a targeted intervention rather than a comprehensive regime that competes with the rest of the world. Anchoring Canada’s domestic rules to international standards and actively working to make them interoperable will protect sovereignty more effectively than acting alone.
Although the prorogation that killed AIDA was unintentional, it created something useful: a pause. The choice before the Government now is not between regulating AI and ignoring it; rather, it is between writing another rulebook that isolates Canada and increases the fragmentation Canadian companies are already paying for, or doing something harder and more valuable — becoming the country that makes the existing rulebooks work together.
