In the absence of regulatory controls, artificial intelligence is predicted to increase inequality across nations, cities, corporations and individuals by concentrating productivity gains in the hands of the few.
Currently, a handful of U.S. companies almost entirely control how AI technology is developed and used. As a result, democratically-elected governments can do little except try to limit AI’s potential adverse effects after the event.
For this to change, Canada needs to pool its resources with other nations to create an independent research organization with the long-term and large-scale funding needed to develop leading-edge AI resources that provide alternatives to commercial AI.
Canada is the undisputed birthplace of modern AI. It is the home of many internationally-celebrated AI pioneers, including Yoshua and Samy Bengio, Geoffrey Hinton, Richard Sutton and Ilya Sutskever.
But as Google’s acquisition of British startup DeepMind demonstrated a decade ago, AI innovations and talent will ultimately be absorbed into U.S. behemoths absent trillion-dollar levels of home-country funding.
Funding on such a scale is unlikely. In 2026, American AI expenditure will exceed one-quarter of Canada’s total GDP. While Canada recently announced an AI for All strategy, it does not commit anywhere close to the level of funding needed to restore Canada’s historic leadership in AI.
Meanwhile, AI’s capabilities are growing exponentially, as AI tools are used to facilitate their own research and development. We can expect this trend to further widen the capability gap between the leading AI companies and their nearest competitors over time. In the long term, being second best in AI will not be a viable option. Whether we like it or not, commercial AI development is destined to become a winner-takes-all race.
Only the U.S. and China have the funds needed for the continued single-nation development of AI. But it poses an existential risk to all nations for AI to be dominated by just two superpowers — particularly ones that regularly ignore international laws and norms. American hegemony in AI will be especially costly to Canada because of its close ties to the United States.
Canada simply doesn’t have the heft to fight off the American AI-behemoths alone. The same is true for EU member states, Britain, Japan, South Korea and the rest of the world.
That is why scientists around the world have been calling for a CERN-like international AI-research agency for nearly a decade. To date, several half-measures in this direction have been launched. But none come close to establishing a well-funded, long-term, independent and physically-embodied organization like CERN, Europe’s organization for nuclear research.
Because of the winner-takes-all nature of the AI race, half-measures are doomed from the start. What is needed is a body with the scale, ambition, funding and expertise that can compete with Google, Microsoft, Amazon, Meta, OpenAI, Anthropic and so on.
It is not too late. The AI revolution is still in its infancy. The current generation of AI models have fundamental limitations that will require major breakthroughs to fulfill AI’s true potential. The next several years will deliver next-generation AI models with verifiable behaviour, conceptual understanding, strategic reasoning and goal-directed incremental learning.
The only way Canadians can ensure the safety and equity of this next-generation of AI is by pooling their resources with other nations to create an open, shared, public and inclusive AI research institute.
Some Canadian AI researchers fear that such action would move the centre of gravity of AI research away from Canada. But modern AI R&D requires energy and financial resources far beyond what most nations can provide individually.
To continue to have a say in how AI develops, Canada must leverage its historic leadership of AI research to create a multinational, world-leading and world-scale AI research centre capable of providing an alternative to the dominant American and Chinese AI giants.
