The U.S. government is generally supportive of countries like Canada pursuing their own sovereign artificial intelligence strategies, so long as it’s not at the expense of “the best technology,” an American official said Friday.

Those comments and others were made during a briefing call held with reporters by the U.S. Embassy in Ottawa on how the U.S. and Canada are aligning and seeking further partnerships in AI.

“I see no issue with Canada promoting Canadian companies,” said the U.S. Embassy official, who spoke on background in order to freely discuss U.S. AI policy. “This is the same principle that we see in the American AI export plan.

“The challenge is when that push comes at the expense of the best technology. In niches where Canadian firms are the world-leading technology, they should be selling around the world. And where there are stronger U.S. firms or technologies, we think that we should be partnering up as trusted allies and selling our technology together around the world.”

The official highlighted areas where the U.S. and Canada are already working closely together, such as on export controls to ensure technology and intellectual property isn’t being stolen by other countries.

Countering Chinese influence, “particularly in international governance bodies” so that Beijing isn’t in a position to “dictate” AI use by sovereign nations, is another area of growing cooperation.

But the presentation also exposed key differences in the two countries’ approach, particularly on the issues of trust and safety.

While Canada is seeking to address concerns like privacy and online harms through new legislation — which include the potential for hefty fines on AI companies that don’t comply, including American firms that currently dominate the AI space — the U.S. official made clear that companies themselves are being left to address those concerns rather than government.

Get breaking National news

“The AI firms are best suited to identify the needs that consumers really care about, and then respond to those needs,” the official said. “That’s innovation at its basis.”

That includes companies working closely with communities as they pursue large data centre projects, the official added, and “understanding their concerns around water use, electricity prices, sound pollution and implementing policies as part of their projects that respond to the needs of the local community.”

“I want to be clear that pro-innovation regulation does not mean no regulation,” the official said. “It means flexible, risk-based approaches that can adapt to match the speeds that AI is moving at.”

The U.S. official would not comment on the Canadian online harms or privacy legislation, nor whether AI partnerships are being discussed during broader trade negotiations between Canada and the U.S.

Ipsos polling on AI views among dozens of countries has shown Canadians are among the most wary populations when it comes to the fast-evolving technology.

Yet the polling also shows Americans aren’t far behind: just 33 per cent said they are excited about products and services that use AI — the third-lowest score among surveyed countries — compared to 26 per cent of Canadians.

• Google hit with US$1B fine as EU says app store, search stifles competition
• OpenAI says its AI hacked another company in ‘unprecedented cyber incident’
• AI data centres have 4 in 5 Albertans concerned over environment, electricity costs: Poll
• Peterborough music community divided over AI use for concert posters

The Trump administration’s AI action plan from last July has sought to loosen regulations and speed up permitting for large-scale data centres in the U.S., with an overarching goal of fostering widespread AI innovation and adoption while securing American dominance over the technology.

“America leads the world in AI technologies, and American AI benefits the world by setting the gold standard for technology that is responsibly developed, rooted in transparency, and free from ideological bias and manipulation,” the U.S. Embassy official said.

The action plan also calls for the export of that technology to allies and partners, and the U.S. official said Canadian companies have joined American-led consortia groups to export their own components alongside U.S. firms.

Canada’s AI for All strategy, unveiled in June, does not mention the U.S. as a partner in its pillar on building global alliances. Instead, it focuses on agreements Canada has already signed with allies in the European Union, Middle East and Indo-Pacific.

Yet AI Minister Evan Solomon has acknowledged that, at least for now, Canada will have to rely in part on existing American technology as it builds up its sovereign capabilities, while also maintaining partnerships on data security and safety.

“Sovereignty is not solitude,” Solomon said during a discussion at last month’s U.S.-Canada Summit in Toronto. “It doesn’t mean that we’re building a moat. No country can do that in technology, no country can do that in trade.

“Our best partner, our biggest partner, is the United States. It is, it will be, it remains vitally important.”

The U.S. official said it “doesn’t make sense for Canada to try to build out every part of the AI stack,” a term used to encompass the hardware, models, software and other components needed for AI technology.

“There are specific technologies where the Canadian companies are very strong, and can partner well with U.S. companies or with other foreign companies, and building a trusted AI stack,” the official said.

“But just the actual cost of trying to build everything from zero all the way up and doing it entirely Canadian, it would be expensive, it would be slow, and it would result in the Canadian AI economy being far behind.”