TORONTO — Canada is seeking to position itself as a leader in the use of AI in the public service, launching a Group of Seven program to find ways that the technology can improve government operations.

Leaders at the G7 Summit in Kananaskis, Alta. on Tuesday issued a joint statement on “AI for prosperity” heavily focused on boosting adoption of the technology, and far lighter on regulatory talk than the communiques from previous meetings.

That includes encouraging more AI in government. Ottawa plans to host a challenge program and a series of “rapid solutions labs” to develop new applications of AI for the public sector. The G7 governments will collaborate to prepare the tools for large-scale use, and a new AI network will share open-source code between members. The group also promised to work with emerging economies to make AI more accessible elsewhere.

The new Canada-led programs come as the Liberal government under Prime Minister Mark Carney touts AI’s potential to make federal departments more productive and efficient. It’s also signalled Ottawa will use its buying partner to help Canadian AI firms grow. “Driving adoption means we’ve got to walk the walk,” AI Minister Evan Solomon said at an event last week.

Other G7 countries have already launched similar strategies. The Labour government in the United Kingdom estimates it can save £45 billion ($83.5 billion) a year by automating and digitizing more civil service work. And it’s promised to test and scale AI tools across departments with dedicated funding. British bureaucrats are already saving time using generative AI systems for text-based tasks like drafting correspondence and updating records. 

Both London and Ottawa have this week committed to work with Toronto-based Cohere on ways to use the firm’s technology in government operations. Carney and Solomon have both recently met with Cohere executives, as well as leaders from other tech startups like Ada.

Beyond their own workforces, the G7 countries are also trying to encourage small and medium-sized businesses to use AI. In Kananaskis, they agreed to create an “adoption roadmap” that includes plans to ensure firms can access the processing power to run applications and open-source models. The G7 countries will also back programs to develop “easy to implement” AI tools, and to train small companies on how to use them. 

The leaders’ statement also promises to encourage “responsible AI deployment.”  Unlike past G7 summits, the statement issued out of Kananaskis did not produce any major new governance or rule-making initiatives. 

By contrast, Japan’s 2023 presidency launched the Hiroshima AI Process, designed to promote “safe, secure, and trustworthy AI.” Running alongside the U.K.’s landmark AI Safety Summit, the G7 initiative came up with voluntary principles for organizations developing and using the technology. 

During Canada’s last G7 presidency in 2018, countries committed to promoting “human-centric AI,” including by encouraging industry-led work on safety work and backing research into the technology’s potential negative impacts like bias and privacy. That led to the Global Partnership on Artificial Intelligence (GPAI), which studies issues like data governance and the future of work.

Over the last two years, G7 countries have shifted their emphasis from AI rule-making to adoption. In the U.K., the Labour government has scaled back the focus on safety, while France used its recent AI summit to pitch for data centres and other investments. In the U.S., the Trump administration—which opposed GPAI in its first term—has warned against “excessive regulation” of AI and threatened to retaliate if other governments impose new requirements on American tech firms, many of which are at the top of the AI food chain.

Citing the shift in international mood, Solomon said last week that Canada must also move from “over-indexing” on AI risks to encouraging its adoption. “The United States and China have no desire to buy into any constraint or regulation,” he said, so Canada trying to “go it alone” would be “a waste of time.”

The Kananaskis summit did not produce the traditional joint communique, but the U.S. did sign on to the AI statement as well as other pre-negotiated declarations on critical minerals, quantum technologies and wildfire mitigation.