Canada, a global hub for artificial intelligence research and home to some pioneers of the technology, on Thursday announced plans to position itself as a leader among middle powers vying to build sovereign A.I. capability.
These plans, unveiled as part of the country’s national A.I. strategy, commit to injecting millions of dollars into research facilities; to introducing relevant privacy and consumer protection legislation; to building a public A.I. supercomputer; and to creating free programs to learn A.I.
When it comes to tech policy, Canada has tended to ride the tailwinds of the United States. But amid continuing political and trade tensions, the United States, its former ally, is notably absent from Canada’s vision for its A.I. future.
“Prosperity and sovereignty in the age of A.I. belong to nations that can build, adopt and govern A.I. on their own terms,” Prime Minister Mark Carney told reporters on Thursday from a downtown Toronto hospital that is leading health technology innovations.
Canada has reoriented itself to align with like-minded middle powers such as Australia, France and Germany in its A.I. ambitions, just as it has done with its military, trade and energy infrastructure projects.
“Canadian A.I. adoption will be prudent, pragmatic and pro-worker,” Mr. Carney said.
Still, he emphasized that the conversation with the United States was not adversarial, and that companies based there would continue to play an important role in Canada’s technology ecosystem. Mr. Carney pointed to the example of Anthropic, which has granted the Canadian government access to its latest A.I. model, Mythos. The company has said the model is so powerful that it is too dangerous to release publicly.
Public-safety threats spurred by the use of A.I. chatbots have become a worry for Canadians after a mass shooting in February at a school in Tumbler Ridge, British Columbia, that killed eight people.
Eight months before the attack, the shooter’s account with OpenAI had been flagged internally for content that violated the company’s policy and was suspended. Though there was discussion within the company about raising these concerns with Canadian law enforcement, this did not happen. Open AI has since apologized and revisited its policies for elevating safety concerns.
The government plans to advance legislation focused on protecting children’s information and on safeguarding personal privacy rights against “deepfakes” and surveillance pricing, the practice of adjusting the prices people see based on personal data about them.
Canada also wants to mirror the approach it has taken with defense sovereignty — procuring locally and building domestic capacity before shopping elsewhere — with its digital sovereignty.
Mr. Carney emphasized that other core Canadian values, like the French language, its Indigenous heritage and its history, particularly as the training ground of three “godfathers” of A.I. — Geoffrey Hinton, Yoshua Bengio and Richard Sutton — will be reinforced in the government’s shaping of the technology’s uses. The computer scientists are known for their breakthroughs in laying the mathematical and theoretical groundwork for modern A.I.
The government has set an ambitious target to create 250,000 jobs in the A.I. space in the next five years. The goal is likely to face major hurdles because of Canada’s perennial “brain drain” problem, in which the country trains highly skilled workers who then depart, usually for the United States, in search of lower taxes, higher wages and more opportunity.
“Canada helped make modern A.I. possible, and Canadians should be proud of that,” Aidan Gomez, chief executive of Cohere, a Canadian A.I. company, said in a statement.
“Canada has seen too many big ideas grow elsewhere,” Mr. Gomez added. “A.I. should be where that changes.”
