Canada’s AI health revolution is here. Is our health-care system ready?

The message from a panel discussion of health-care leaders, researchers and innovators was clear: The long-promised future of artificial intelligence in health care has arrived. 

Sensors built into orthopedic joint implants are currently tracking patient recovery at home. AI scribes are being used in primary care clinics throughout Canada. A mortality prediction algorithm introduced at a Canadian hospital five years ago has contributed to a 26 per cent reduction in unexpected patient deaths. But for transformation to yield better and more efficient health care, it requires stakeholders to build new systems and, policies, and a highly skilled workforce able to keep pace.

A picture of what is already possible

The conversation was part of an ongoing series organized by the Munk School of Global Affairs and Public Policy and the U of T’s Government Relations Office that brought researchers and government stakeholders together to address policy challenges in Ontario and Canada.

Attendees included representatives from the Ministry of Health, Ministry of Economic Development, Job Creation and Trade, Invest Ontario, the City of Toronto, the Toronto Academic Health Sciences Network and researchers at member hospitals, as well as industry giants AstraZeneca Canada and Roche Canada.

Muhammad Mamdani, director of the U of T’s Temerty Centre for AI Research and Education in Medicine (T-CAIREM) and Ontario Health’s clinical lead for AI; Duska Kennedy, Chief Digital Officer at North York General Hospital; and Noah Crampton, clinician-scientist at Toronto Western Hospital and CEO of Mutuo Health Systems led the discussion with presentations from each of their perspectives.

Clinicians are leading the change

Perhaps the most striking observation is that AI adoption is being driven by clinicians. North York General Hospital knows this firsthand, Duska Kennedy said. After implementing AI-assisted mammography screening, administrators found that a machine-generated second reading improved detection. 

The hospital has also launched emergency triage algorithms, an AI clinical scribe, and an internal policy chatbot that provides staff with instant access to hundreds of frequently updated policies and guidelines. Across these cases, the most successful implementations shared one important characteristic. They all had clinical champions who led the work from the start.

Another panellist pointed out that scaling AI is a group effort. It requires a multidisciplinary approach with clinical teams who deeply understand the problem, data scientists who can develop AI solutions, governments who enable AI adoption through policies and directives, and the private sector that can commercialize solutions for sustainability. Consequently, industry, government, and clinical teams should ideally co-design solutions from the outset.

The structural gaps need closing

To scale AI tools and applications at the provincial and national levels, investments in data infrastructure are needed. AI systems, for example, require high-quality, interoperable data, and secure computing environments capable of continuous learning.

Procurement also remains a challenge. Hospital budgets are frequently built around one-time capital purchases. However, continuous learning systems require ongoing investment, and current funding structures are simply not designed for it.

Perhaps the biggest challenge to sustained structural change is AI literacy, the skills needed to use and apply the technology. A 2025 KPMG survey ranked Canada 44th out of 47 countries in AI literacy with a range of educational initiatives, including paid summer research placements for trainees, workshops, professional development courses for clinicians and residents, and the development of an AI educational curriculum for medical schools around the world.

Another way to counter these challenges is through increased coordination among practitioners. T-CAIREM, for example, recently formed a community of practice for AI researchers and clinicians at hospitals in the Toronto Academic Health Sciences Network (TAHSN). The goal of the community is to share knowledge, discuss research developments, and explore best practices in the field.

The opportunity in front of us

To leverage the expertise and experience of so many significant stakeholders in Ontario’s health-care system, the participants proposed greater formal coordination among governments that hold data, universities that develop research techniques and tools, and the private sector that supports public-sector innovation at scale. 

One proposed starting point might involve the launch of provincial-level sandboxes. These shared environments would enable hospitals, researchers, and the private sector to innovate together by leveraging pooled data and skills. The potential payoff, one panellist noted, would far exceed what any organization could achieve on its own.

Despite the challenges ahead, the panel closed on a note of informed optimism. Ontario has the researchers. Ontario has the clinicians. Ontario has the data. What comes next depends on the courage to act differently and the conviction to follow through.