Listen to daily commentary from program director Murray Wood
Murray Wood says the federal NDP seems to be heading for a political wipe out, as the left-leaning vote seems to be goin...
TORONTO — Medical researchers and lawyers say our rocky relationship with the U.S. creates an urgent need to protect a critical Canadian resource: patient health information that can be used to train artificial intelligence.
“Our health data is the most valuable health data set in the world,” said Natalie Raffoul, an intellectual property lawyer in Ottawa.
“You can’t go to any other jurisdiction and be able to pool a data set like this because no one else has a public health system like this with the kind of ethnic diversity that we do.”
Many Canadian institutions use cloud servers run by American companies to store health data, experts say. That, combined with President Donald Trump’s stated objective to make the U.S. a world leader in AI and his desire to make Canada a 51st state, means it’s possible that his administration could come after our data — perhaps citing national security concerns as he has with tariff executive orders, experts say.
Dr. Amol Verma, a professor of AI research and education in medicine at the University of Toronto, said as artificial intelligence is increasingly used in health care, algorithms need to be trained on the most representative data possible to deliver accurate and useful results.
The U.S. doesn’t have that level of inclusivity in its own health data because its private health-care system means that many people without health insurance might not be accessing care and therefore their health information wouldn’t be captured, he said.
That means that an AI model trained on U.S. data could be biased or not work well “in certain racial populations or linguistic populations” said Verma, who is also an internal medicine specialist at St. Michael’s Hospital in Toronto.
Dr. Kumanan Wilson, research chair in digital health innovation at the University of Ottawa and a physician at The Ottawa Hospital, said health information, largely from electronic medical records, “could be of significant economic benefit to the U.S. and having access to our data would be very valuable.”
“Our major cloud providers are all American. They’re AWS — Amazon Web Services — Microsoft Azure, and there’s Google Cloud. And all of these could potentially be vulnerable to U.S. legislation if the Trump administration wanted to access that data,” he said, noting that many Canadian hospitals have data on those cloud servers.
Michael Geist, Canada Research Chair in internet and e-commerce law at the University of Ottawa, said this wasn’t a concern on his radar prior to Trump’s inauguration — but things have changed.
“The recent political events in the relationship … between Canada and the United States requires at a minimum a willingness to re-examine or rethink just about everything,” he said.
The Canadian Press reached out to Amazon Web Services, Microsoft and Google for comment.
Both Microsoft and Google said they were not going to comment or speculate on hypothetical scenarios, but noted that if any government wanted access to data, it would have to serve them with a lawful court order or warrant.
Microsoft also said its “legal compliance team reviews all requests to ensure they are valid, rejects those that are not valid, and only provides the data specified.”
Google said that if it received a court order that it believed should be directed to the owner of the data, it would point the requester to the client.
A spokesperson for Amazon Web Services referred The Canadian Press to its website, which says the company would not disclose data unless it was required to do so by a “legally valid and binding order.” It also says that the company defends customer information and has previously “challenged government demands for customer information that we believed were overbroad.”
AWS also issued a statement saying it has two data storage regions within Canada.
“There have been no data requests to AWS that resulted in disclosure of enterprise or government data stored outside the U.S. to the U.S. government since we started reporting this statistic,” the statement also said.
But both Raffoul and Geist said storing Canadian health data within Canada may give institutions a false sense of security if the company is American.
“Putting valuable data and intellectual property in the hands of a Canadian subsidiary to a foreign multinational is essentially like handing it to a foreign multinational. And so this is where there’s a lot of confusion,” Raffoul said.
Geist said he wasn’t confident the Trump administration would respect the companies’ privacy measures, saying “their effectiveness may not be ironclad.”
In addition, “even if the data is retained in Canada that does not provide a full guardrail against a U.S. court demand for disclosure,” he said.
“Canadian law on its own — at least as privacy law currently stands — isn’t really going to be enough because there really isn’t much by way of penalties. That’s been one of the long-standing criticisms of Canadian privacy law,” Geist said.
One way to strengthen privacy laws he said, is to add a “blocking statute.”
A blocking statute would state that a U.S. company would face severe penalties in Canada if it disclosed Canadian data, which would allow an American judge to rule that there’s a strong reason why the company can’t abide by a U.S. order to release it.
The best solution, experts say, is to eliminate the risk altogether by building more Canadian-owned technology companies that provide health data storage.
In addition to having enviable health data, Canada is arguably a world leader in AI, experts say. The Nobel-prize winning “Godfather of AI,” Geoffrey Hinton, has done much of his work at the University of Toronto and as chief scientific adviser at the Vector Institute in Toronto. The Canadian government has also established AI “clusters,” including SCALE AI and DIGITAL, to advance research in the field.
Verma says the current wave of patriotism across Canada sparked by Trump’s aggression is an opportunity to pool health data across provincial and territorial boundaries and use it to build even stronger AI algorithms in health care.
“That’s where I think there’s a really exciting opportunity in this moment because our political leaders are talking about the need for our provinces to work together,” he said.
“I think there’s a window of opportunity where Canada can capitalize on its strengths and lead. But if we are slow or complacent, then we will fall behind,” Verma said.
This report by The Canadian Press was first published April 24, 2025.
Canadian Press health coverage receives support through a partnership with the Canadian Medical Association. CP is solely responsible for this content.
Nicole Ireland, The Canadian Press
