For Digital Health Week, we asked Canadians about AI in healthcare to understand where they’re comfortable seeing it used, and what outcomes they want if it’s used at all.
We found two signals strongly predict whether Canadians will say “yes” to AI in healthcare: how they feel about their own health and how they view the health system overall.
Why this matters: in a world of rapid change and eroding trust in governments and institutions, securing genuine public buy-in isn’t optional. It’s the only way to move beyond pilots and responsibly realize AI’s benefits at scale. Without that consent, deployments stall and skepticism hardens.
Where Canadians are comfortable with AI
We asked where Canadians are comfortable with AI helping make decisions or provide recommendations across financial services, government, workplaces and healthcare providers/healthcare services.
Three numbers frame the conversation:
AI in decision-making: Where do people trust it most?
Graph Insights Credit: Environics Research
This tells us that, overall, Canadians are quite cautious about the use of AI in healthcare. To better understand why some are more comfortable than others, we explored a range of variables and identified two particularly powerful drivers:
- How people feel about their own health (personal health status)
- How they feel about the system (confidence that the healthcare system is stable vs. in crisis)
How people feel about their own health
We see a similar dynamic in our PatientConnect segmentation: people who feel their health is in a good place tend to trust clinicians more, are more proactive, and are more open to technologies that help them stay healthy.
Comfort with providers using AI rises from 10.5% among Canadians who rate their health as poor to 20.9% among those who say very good.
How people feel about the system
Comfort with AI being used in healthcare settings also shifts with healthcare system confidence: it’s 14.3% when Canadians see the healthcare system as in crisis, versus 20.3% when they rate it excellent. And the “not comfortable anywhere” group drops sharply across the same views: 56% (in crisis) to 28% (excellent).
This shows us that comfort with AI in healthcare isn’t just about the technology itself or demonstrating clinical benefits. Comfort grows when people feel secure in their own health and have confidence in the healthcare system and it erodes when they feel their health is poor or outside of their control or believe the system is in crisis.
What do people want from AI in healthcare?
In addition to asking whether people are comfortable with AI, and where, they’ll accept it, we asked what they want AI to do if it’s part of the system:
We asked Canadians: “When AI is used in healthcare, which outcomes are most important to you personally?”
What we found: when people see the system as in crisis, they want AI to reduce wait times and improve access – meaning they want to see it used to clear the backlog first. When they see the system as excellent or stable, they want AI to improve diagnostic accuracy. For those in the middle, many emphasize keeping a human clinician involved while using AI to make care both faster and more precise.
PatientConnect – Understand the psychographics that drive decision making for healthcare patients
Moving from Caution to Confidence
Our findings show that comfort with AI in healthcare grows when people feel informed, in control, and able to trust themselves, their providers, and the system as a whole. So the path forward isn’t just about building better AI models but about building trust.
This means designing, communicating, and governing AI tools in ways that feel safe, human, and genuinely useful in people’s real lives. The considerations below outline some considerations on where focus if we want AI in healthcare to feel less risky, more relatable, and worth saying “yes” to.
Focus on building AI literacy
Raise baseline understanding of what AI does, what it doesn’t, and where and how humans stay in control. Plain-language explainers, real clinical examples, and transparent FAQs help people see how AI is integrated and why it adds value helping to reduce uncertainty and boosting confidence without the jargon.
Design with “oversight-first,” not “AI-first”
Make it clear where and why AI tools are used, how the clinician is involved, and how the patient is included in the final decision. In lower-trust or crisis contexts, lead with safeguards: privacy protections, bias testing and auditability before talking benefits. Oversight should be the headline.
Put trusted KOLs out front (not politicians)
Ask clinicians, pharmacists, nurses, and patient advocates who are actually using AI to explain how it works in practice and what protections exist. Finding ways to convene these credible voices for town halls, short videos, and case vignettes. Peer and professional trust beats political messaging every time.
Speak to people’s values, mindsets, and motivations
Patients are people first. This means they’re going to see the value of AI in different ways. If we only look at them through demographics or condition type, we miss the nuance. Two people of the same age, gender, and diagnosis can have totally different reactions to AI depending on their values, mindset, and how they experience the system.
That’s where a values-based lens like our PatientConnect segmentation can help you understand people at a deeper level. Tools like this remind us that trust in AI isn’t built with one generic message; but built by speaking to what actually matters to different kinds of patients.
Speaking to people’s values, mindsets, and motivations helps ensure the message matches what matters most to them. For example:
Control & autonomy: For those who are strong on control and autonomy and are people who value independence and self-direction, there is an opportunity to highlight opt-in/opt-out choices, data control, and clear human sign-off on all decisions.
Safety & fairness: For those who are driven by safety and fairness and are people who worry about risk, harm, or being overlooked, there is an opportunity to show how systems check for bias, what quality review steps exist, and simple, visible escalation paths if something doesn’t feel right.
Access & dignity: For those who prioritize access and dignity and people who feel the system is hard to navigate or disrespectful of their time there is an opportunity to emphasize shorter waits, fewer repeat visits, and easier, more respectful navigation through care.
Confidence & competence: For those who value confidence and competence and are people who place strong trust in clinical expertise, there is an opportunity to frame AI as support, not replacement, helping skilled clinicians catch more, earlier, and make more confident decisions.
When we design and communicate AI through this more nuanced, human lens, it’s much more likely to feel relevant, respectful, and worth saying “yes” to.
The path forward is clear: if we want wider comfort with AI in healthcare, we must build trust intentionally. That means designing tools that are transparent, governed responsibly, and aligned with what matters most to real people and not assumptions about them. In order to see AI succeed and have meaningful impact on the system, we need to ground our decisions in values and create a system where patients feel informed, seen, and supported.
