How Much Should You Trust AI in Your Health Care?

Canadians have apparently made peace with a computer reading their scan. What they have not made peace with is where the computer lives.

That is the shape of a national survey released on 1 October by Santis Health and Nanos Research, and the split it exposes is sharper than the usual technology-anxiety story. Strong majorities back AI in health care when it helps a clinician work faster. Those same Canadians go cold the moment the question turns to which country the tool runs in and where the data ends up.

More than two-thirds, 68 percent, say they are uncomfortable to some extent with the Canadian health system using AI tools based in the United States. Four in five, 80 percent, agree or somewhat agree that Canadian health data should stay in Canada and not be used to train models elsewhere. Neither of those numbers, though, amounts to a rejection of the technology itself.

The Bottom Line

Canadians support artificial intelligence as an assistant to their clinicians and oppose it as a replacement for one, and they draw a hard line at their health data crossing the border. Seventy-one per cent are open to AI reviewing their tests or scans if it shortened the wait to see a specialist, 86 percent want Ottawa to do more to regulate how AI companies use health data, and across every use tested, Canadians were more likely to accept AI in an assistive role than an autonomous one. For patients, nothing changes tomorrow, though the tools are already turning up in Canadian care whether or not anyone announces it.

Key Takeaways

Sixty-eight per cent of Canadians are uncomfortable to some extent with the health system using AI tools based in the United States.

Eighty per cent agree or somewhat agree that Canadian health data should stay in Canada and not train models outside it.

Seventy-one per cent are open or somewhat open to AI reviewing their medical tests or scans if it cuts their wait to see a specialist, while 23 percent are not.

Eighty-six per cent say the federal government needs to do more to regulate how AI companies use health data, and 42 percent are unsure whether those companies currently protect it well.

Sixty per cent trust health care providers most to manage their health data. Canadian technology companies draw five per cent and foreign technology companies less than one per cent.

The survey reached 1,037 Canadians aged 18 and over between 30 August and 2 September 2026, with a margin of error of plus or minus 3.1 percentage points, 19 times out of 20.

Why Should Canadian Health Data Stay in Canada?

The findings suggest people are separating the technology from the companies and countries behind it. Canadian health data has become a sovereignty question rather than a privacy question, and the distinction matters, because the same respondents who want their records kept onshore are perfectly relaxed about researchers using them.

Three in four Canadians, 76 percent, are open or somewhat open to more of their anonymized health data being used by researchers to train AI models. Comfort runs high for specific research uses: 82 percent are comfortable to some extent with researchers using AI to discover potential new drugs, 75 percent with evaluating whether those drugs work as intended, and 73 percent with monitoring drug safety.

Trust clusters tightly around the people already holding the file. Sixty per cent name health care providers, meaning doctors' offices, clinics and hospitals, as the party they most trust to manage their health data safely, followed by health researchers at 14 percent and provincial health authorities at 11 percent. Technology companies finish last by a distance, Canadian ones at five per cent and foreign ones at under one per cent.

What Worries Canadians About US-Based AI Health Tools?

The discomfort with US-based AI health tools sits at 68 percent, and the survey does not read as a verdict on American engineering. It reads as a verdict on jurisdiction. Once health data leaves the country, the rules that govern it do too, and 80 percent of respondents would rather it did not leave at all.

Asked unprompted about the drawbacks of AI in the health system, Canadians named the potential for a misdiagnosis first at 23 percent, with data privacy and security risks close behind at 21 percent. Those two worries pull in different directions for policy, since the first is a clinical safety problem and the second is a legal one, and the survey suggests the public is holding both at once.

Accountability is the other half of the concern. Fifty-seven per cent believe that if an AI tool makes a mistake or misdiagnoses a patient, responsibility should be shared across the doctors and clinics who used it, the health authorities who approved it, and the developers who built it, rather than any one of the three.

Could AI Cut the Wait to See a Specialist?

That is the trade Canadians appear most willing to make. On AI and specialist wait times, 71 percent said they are open or somewhat open to an AI tool reviewing the results of their medical tests or scans if it reduced how long they waited, while 23 percent were somewhat not open or not open.

The appetite is not difficult to explain in a country where specialist waits have become a story in their own right and emergency departments are losing the physicians who staff them. When the status quo is a nine-month referral, a faster read of a scan starts to look less like a gamble and more like an exit.

There is a limit, and the survey found it in the same place every time. Across every use tested, including renewing prescriptions, issuing specialist referrals and advising on ongoing conditions, Canadians were consistently more likely to accept AI assisting a health care provider than operating on its own. Patrick Nelson, managing partner at Santis Health, framed the result as a mandate, saying Canadians have told them the health system needs to change and that AI can help deliver it.

Who Is Asking for More AI Regulation in Canada?

Almost everyone, as it turns out. Support for AI regulation in Canada reached 86 percent, combining the 62 percent who agree and the 24 percent who somewhat agree that the federal government needs to do more about how AI companies use health data. A further 42 percent said they are unsure whether those companies currently do a good job of protecting it, which is less a criticism than an admission that nobody can tell.

Nik Nanos, chief data scientist at Nanos Research, tied the whole picture to that uncertainty, saying Canadians' optimism about what AI can bring depends on whether decision makers can deliver on clearer AI accountability and privacy. The optimism is real and measurable. Forty-seven per cent agree to some extent that the benefits of AI in health care outweigh the downsides, against 24 percent who disagree.

Canadians also want public money behind the technology rather than private money alone. Eighty-two per cent support taxpayer funding for AI tools that help researchers develop new drugs or devices, 81 percent for tools that improve how governments manage the health system, and 75 percent for tools that assist providers delivering care.

Should You Be Using AI for Health Questions Yourself?

Plenty of people already are. Thirty-six per cent of Canadians report using AI tools to better understand their health, 24 percent to make health care decisions for themselves, and 11 percent to make decisions for someone else. Using AI for health questions has become a mainstream habit without anyone formally endorsing it.

The reassuring finding is about restraint. Among those who have used these tools, 60 percent say their AI use has never led them to skip professional care in a situation where they would otherwise have sought it. People appear to be treating the chatbot as a first pass rather than a final answer, which is roughly the right instinct.

Expectations are also forward-looking. Fifty-one per cent think future AI tools will do a good or very good job of improving their personal health, while only 28 percent say current tools manage that today. If you are using one now, check where the tool is based and what its privacy policy says happens to what you type into it. A symptom you cannot interpret, such as dizziness that keeps coming back, is a good reason to get a human opinion rather than a confident paragraph.

Your Next Step

Check where any AI health tool you use is based, and read what its privacy policy says it does with what you type in.

Ask your clinician directly whether AI is involved in reading your scans or drafting your notes, and where that processing happens. You are entitled to an answer. If you do not have one, start with a family practice accepting new patients.

Treat an AI answer as a starting point and never as a diagnosis, particularly for a new or worsening symptom.

Get a human opinion when something concerns you. Compare real-time wait times at walk-in clinics near you, or search clinics and providers by symptom and city.

Keep reading. The Medimap Health Hub tracks how Canadian care is actually changing, from wait times to staffing.

Frequently Asked Questions

What did the Santis-Nanos survey find about AI in health care?

It found Canadians support AI in health care as an assistant to clinicians rather than a replacement. Seventy-one per cent would accept AI reviewing their tests or scans to shorten specialist waits, while 68 percent are uncomfortable with the system using tools based in the United States and 80 percent want Canadian health data to stay in Canada.

Do Canadians want AI to diagnose patients without a doctor?

No. Across every use tested in the survey, including prescription renewals, specialist referrals and advice on ongoing conditions, Canadians were consistently more likely to accept AI being used to assist a health care provider than to operate autonomously. Support for AI acting on its own was lower in every single category measured.

How many Canadians already use AI tools for health questions?

Thirty-six per cent report using AI tools to better understand their health, 24 percent to make health care decisions for themselves, and 11 percent to make decisions for someone else. Among users, 60 percent say their AI use has never caused them to skip professional care they would otherwise have sought.

Who do Canadians trust most with their health data?

Health care providers, by a wide margin. Sixty per cent name doctors' offices, clinics and hospitals as the party they most trust to safely manage health data, followed by health researchers at 14 percent and provincial health authorities at 11 percent. Canadian technology companies draw five per cent, foreign ones under one per cent.

How was the survey conducted and how reliable is it?

Nanos Research conducted a random-digit-dialling dual-frame hybrid telephone and online survey of 1,037 Canadians aged 18 and over from 30 August to 2 September 2026, as part of an omnibus survey. Results were weighted by age and gender using census data and stratified geographically. The margin of error is plus or minus 3.1 percentage points, 19 times out of 20.

Disclaimer

This article is general health information and is not medical advice. Medimap is not a doctor, and nothing here replaces an assessment by a qualified health professional. If you are experiencing a medical emergency, call 911 immediately.