Canada Pushes G7 to Create New AI Safety Body

Canada's Artificial Intelligence Minister Evan Solomon says he has been in direct talks with his international counterparts about establishing a new organization specifically dedicated to tackling AI safety, framing Canada's approach as "advocating alongside our partners" rather than acting unilaterally, according to CTV News's reporting on Solomon's comments.

Solomon's remarks land directly in the wake of the UN's own independent scientific panel warning, an episode we covered in detail in our earlier reporting on that panel's finding that traditional AI safeguarding is "unravelling". That panel, co-chaired by Montreal-based AI pioneer Yoshua Bengio, drew its warning directly from the July hack of Hugging Face, and Solomon's push for a formal G7-level safety body represents Canada's concrete institutional response to exactly that kind of finding.

This isn't Canada's first move into formal AI safety institution-building. The Canadian AI Safety Institute, or CAISI, launched with an initial $50 million investment over five years announced in Budget 2024, operating through the Canadian Institute for Advanced Research specifically to bolster the country's capacity to address AI safety risks. CAISI already participates in the International Network of AI Safety Institutes, a coalition that first convened in San Francisco in November 2024, connecting directly to our earlier reporting on Canada and Germany's combined $300 million investment in Yoshua Bengio's LawZero safety nonprofit.

Solomon's push for a formalized G7 body builds on a genuinely active pattern of Canada signing individual bilateral safety agreements over the past year. Canada and Australia formalized a memorandum of understanding deepening cooperation between their respective AI Safety Institutes, focused on joint research, risk mitigation, technical information sharing, and talent exchange. Canada separately signed a joint statement with Norway exploring collaboration on frontier models that establish "safety and responsibility as a foundational design principle," and contributed $1 million toward the UK AI Security Institute's Alignment Project, a broader $29 million Canadian initiative backed by an international coalition including Schmidt Sciences and Amazon Web Services.

What distinguishes Solomon's current push from these earlier bilateral agreements is the scale and formality being proposed. Rather than a series of individual country-to-country partnerships, a dedicated G7-level organization would represent a genuinely more institutionalized, multilateral approach to AI safety coordination among the world's major economies, connecting directly to the broader wave of global institutional AI safety attention we've tracked closely, including King Charles III personally convening the CEOs of OpenAI, Anthropic, Google DeepMind, and Nvidia just days before Solomon's comments.

Solomon's push for formal international coordination lands in genuinely stark contrast to the position articulated by the United States at the UN just one day earlier. We covered in detail President Trump's own UN General Assembly remarks rejecting global AI regulation entirely, declaring the US "totally rejects any attempt to construct a globality scheme to control" AI. That direct juxtaposition, Canada actively organizing multilateral safety cooperation while the US explicitly rejects it at the same international forum, illustrates the genuinely fractured state of global AI governance heading into the final months of 2026.

This story is worth understanding for any business planning multi-jurisdictional AI compliance strategy, since Canada's push for a formalized G7 safety body, if it materializes, could eventually establish coordinated safety standards across multiple major economies simultaneously, a meaningfully different regulatory landscape than the current patchwork of individual national approaches.

For businesses evaluating where genuine multilateral AI governance momentum is actually building versus stalling, this announcement is worth watching directly against the US's contrasting UN position, since the eventual alignment or divergence between G7 partners and the United States will likely shape how fragmented international AI compliance requirements become over the coming year.

What is Canada's AI minister proposing?
Evan Solomon says he is in talks with G7 counterparts about establishing a new international organization specifically dedicated to AI safety, building on Canada's existing bilateral safety agreements and its own Canadian AI Safety Institute.

What AI safety infrastructure does Canada already have in place?
Canada operates the Canadian AI Safety Institute, launched with $50 million in Budget 2024, and participates in the International Network of AI Safety Institutes alongside bilateral safety agreements with the UK, Australia, and Norway.

How does Canada's position differ from the United States on AI regulation?
Canada is actively pushing for formalized multilateral AI safety cooperation, while President Trump told the UN General Assembly the US "totally rejects" any global scheme to control AI, representing a genuinely stark divergence between the two countries' approaches.

Canadian AI Minister Evan Solomon says he is working with G7 counterparts to establish a new international organization dedicated to AI safety, building on Canada's existing $50 million Canadian AI Safety Institute and a pattern of bilateral safety agreements with the UK, Australia, and Norway. The push lands directly amid the UN's own independent scientific panel warning that AI safeguards are "unravelling," and stands in stark contrast to President Trump's UN General Assembly remarks rejecting any global scheme to regulate AI. The contrasting positions illustrate the genuinely fractured state of international AI governance, with Canada pursuing formalized multilateral coordination while the United States explicitly rejects that approach at the same global forum.