As the United States and China race to dominate artificial intelligence (AI), countries like Canada face a critical choice: remain a passive consumer of systems built elsewhere or forge a third path through strategic international collaboration.

Canada cannot afford to be a downstream user of opaque, black-box models developed abroad. Yet much of the current debate frames this as a binary between dependence and technological nationalism or even isolation. As Canadian institutions increasingly rely on AI systems developed and hosted by firms such as Microsoft, much of this infrastructure remains subject to U.S. legal frameworks like the CLOUD Act, leaving core data, model governance, and system oversight outside Canadian jurisdiction.

However, a credible alternative already exists. The Public AI network’s inference layer, built on top of Switzerland’s Apertus large language model, demonstrates what a middle power coalition could build. It offers a practical example of how AI infrastructure can be developed collectively rather than controlled by a handful of firms or states.

No single government can compete against the massive AI investments of US Big Tech and China. This challenge is not new as it is the logic that gave rise to Airbus and the European Organization for Nuclear Research (CERN). Instead of a purely domestic “made-in-Canada” model, Canada is well positioned to convene an international AI consortium like Airbus or CERN. By pooling resources across compute, software, and data with a group of strong democracies such as Japan, Germany, Taiwan, South Korea, the United Kingdom, and Australia, Canada can achieve the scale necessary to present a meaningful alternative to existing models.

We call this solution digital sovereignty through collaboration: pooling resources with like-minded governments, companies, academia, civil society, and open-source projects to create collective capacity. It is already taking shape in initiatives such as Germany’s Sovereign Tech Fund, which provides sustained support for critical open-source technologies, pointing to how shared investment in foundational systems can be coordinated across borders.An “Airbus for AI” approach brings together democratic nations with shared values to build public AI infrastructure at sufficient scale to serve as a meaningful third pole. In other words, Public AI is the new multilateralism.

As Mark Carney argued at the World Economic Forum, “Middle powers must act together, because if you are not at the table, you are on the menu.”

Partners should include countries across Asia and the global majority, and non-state actors including AI corporations like Cohere and Hugging Face, experts in decentralisation like the Ethereum Foundation, and foundations like the Future of Life Institute and Mozilla. No single actor controls the full AI stack, so building a credible alternative requires combining state capacity, private-sector innovation, and civil society oversight. By prioritising AI investments in broader societal functions, including democratic participation and access to trusted information, Canada and its allies can ensure that the future of AI serves the public good as opposed to just market interests.

Moving from Scraping to Data Commons

While generative AI is overhyped, data remains the most undervalued technological asset. Canada can better leverage the power of data by establishing a world-leading data strategy built on ethical, privacy-aware data commons.

Canada can fund institutions to produce, evaluate, connect, archive, and curate data trusts and cooperatives in areas of Canadian strength, such as finance, culture, healthcare, and clean energy. This includes a commitment to Indigenous and other cultural data sovereignty, ensuring communities maintain control over their own knowledge systems. By leading in the establishment of cross-border data-sharing agreements for public AI development, Canada could become the sandbox for global leadership in data sovereignty through collaboration.

Open Source AI as a Public Utility

Public AI means building systems that are made available as public goods, similar to roads or utilities. A fully open stack that includes all training code and documentation alongside data for pretraining, post-training, and evaluation, will provide the freedom for users to fork any element for their own use. It also encourages transparency necessary to scrutinise any model for cultural or language bias, security threats, or model behaviour. As seen with Linux, the economic potential of open-source artefacts funded by public dollars is massive. Canada should champion an internationally focused open-source strategy. Recent proposals for a Canadian “CanGPT,” modelled in part on public-service institutions, reflect growing interest in public-interest AI, even if questions around openness, governance, and international coordination remain unresolved.

Canada’s Strategic Advantages

Canada possesses a rare combination of attributes that position it to lead this coalition. Canada’s trusted diplomatic soft power and tradition of multilateralism position it perfectly to build international consensus around public AI infrastructure. Abundant clean hydroelectric power and fresh water, particularly in Quebec and British Columbia, provide a sustainable foundation for AI infrastructure. World-class AI talent, developed through the Pan-Canadian AI Strategy, the world’s first national AI strategy, launched in 2017. It has helped establish leading institutes like Mila, led by founding scientific director Yoshua Bengio, one of the most cited researchers in artificial intelligence, alongside the Vector Institute and the Alberta Machine Intelligence Institute. And with amongst the world’s most diverse populations, Canada’s stable and democratic governance can support the development of inclusive, multilingual AI.

Canada’s Moment in Public AI

For Canada, this represents both a return to our strengths in multilateralism and an innovative approach to technological development. Rather than attempting to win a game designed for larger players, we can help change the game itself. Canada needs to continue our history as a middle power that punches above its weight by leading a coalition of governments, companies, non-profits, and open-source projects to build the most open, ethical, and trustworthy AI.

The window for establishing alternatives to the current trajectory of AI development is rapidly closing. As compute resources concentrate, data sources are exhausted or locked up, and technical talent flows to a handful of companies led by a small number of billionaires with decreasing democratic checks on their power, the potential future benefits of AI for the average person, relative to its risks, diminish.

With clear vision, strategic investment, and international leadership, Canada can ensure the future of AI serves not just market interests, but the public good. In doing so, Canada can secure our place in the next technological era while staying true to our values of pluralism, inclusion, and shared prosperity.

This conversation is already moving into policy and practice, including at the DemocracyXChange (April 16–18), where Ana Serrano and Jake Hirsch-Allen are scheduled to speak on public AI. Open Canada and the Canadian International Council (CIC) are partners in this year’s summit, with main stage sessions available virtually.

© OpenCanada