We increasingly live in a world defined the United States and China. Certainly these two superpowers are at the technological fronter in many areas, creating the innovations which will reshape our daily life.
That doesn’t mean other countries need to be passive or irrelevant. In a world in which Russia, China, and the U.S. are unpredictable, middle powers have a shared interest in maintaining a predictable, open, and rules-based international order. These include major R&D players such as the nations of the European Union, Great Britain, Canada, the Gulf states, India, Japan, South Korea, Singapore, Australia…
The challenge is that of creating structures which such diverse nations can work with to develop common frameworks and capacities that can protect but which, unlike superpower capabilities, cannot be weaponized against each other.
To that end, I have co-authored with Omer Bilgin of The Future Society, an AI governance group, a blog for the Martens Centre on how the EU and Canada can deepen their cooperation against AI threats. As EU and Canadian leaders meet in Montreal for a summit today and tomorrow, we propose practical measures for an EU–Canada AI Crisis Preparedness and Early Warning Partnership:
Establish trusted EU-Canada channels for sharing information on AI incidents and near-misses.
Aligning Canadian and EU AI incident reporting formats to reduce duplication.
Develop joint AI incident investigation, analysis, and crisis-response capabilities.
Conduct joint AI crisis-preparedness tabletop exercises.
Fund joint AI safety research through the Horizon Europe programme.
Explore Canadian participation in EU cyber-response networks.
Expand cooperation to AI evaluations and trustworthy applications in areas such as healthcare, industry, and public services.
Ultimately, develop the partnership into a model for wider international cooperation on AI safety, notably among middle powers.