Canada is facing growing pressure to integrate Artificial Intelligence into its health system – often framed as a solution to physician shortages, long wait times and rising patient demand.
But before accelerating adoption, we need to ask a more fundamental question: What problem are we actually trying to solve?
If the goal is to improve care, AI has real potential. Evidence shows that AI-assisted tools can support clinical decision-making, reduce diagnostic errors and identify patterns across large datasets that may not be visible to clinicians alone, as highlighted in recent analyses by the Organization of Economic Co-operation and Development (OECD).
Used this way, AI strengthens health care.
But if the goal is to enable fewer health-care professionals to see more patients in less time, then AI becomes something else entirely: a workaround for a structural problem we have chosen not to fix.
Canada does not lack capable future physicians – it lacks capacity to train them.
According to the Association of Faculties of Medicine of Canada, medical school admissions have not kept pace with population growth. Since 2010, the number of seats has increased by about 6 per cent, while Canada’s population has grown by roughly 12 per cent.
This is not a pipeline problem. It is a capacity constraint.
At the same time, Canada’s health-care system continues to face persistent access challenges, including long wait times for care.
And yet, instead of expanding training capacity, the conversation increasingly turns to AI as a way to extend the reach of a constrained workforce.
This is not a solution. It is a substitution for the investment in the human infrastructure required to deliver care.
There is a subtle but important shift happening in how AI is being positioned.
The language is often about efficiency: reducing workload, increasing throughput and optimizing clinical workflows. These goals are not inherently problematic. But they risk reframing health care as a function of speed.
Patients are not transactions.
Care in the Canadian context has always included something less measurable: time, attention and human connection. The ability of a clinician to listen, interpret nuance and respond with empathy is not an inefficiency, it is a core component of quality of care.
AI can support clinical judgment. It can flag risks, reduce cognitive burden and improve consistency. But it cannot replace the relationship between a patient and a provider – nor should it be used to justify a system in which that relationship becomes secondary to volume.
There is often an assumption that faster, more technologically advanced health-care systems represent progress. The United States is frequently cited as an example – quicker access to specialists, shorter wait times, rapid adoption of new technologies.
But outcomes tell a different story.
According to data from the OECD, the U.S. spends far more per person on health care than Canada, yet has lower life expectancy and higher rates of preventable and treatable mortality.
Similarly, the Commonwealth Fund’s most recent international comparison ranked the U.S. last overall among high-income health systems – despite performing relatively well on measures of timeliness – while Canada performed better on several key outcome measures.
The lesson is not that Canada should accept inefficiencies. It is that speed, volume and technological adoption alone do not define a high-performing health-care system – and should not become the benchmark we aspire to.
Public perception reflects this tension.
Recent polling from the Canadian Medical Association shows that while many Canadians are beginning to engage with AI for health-related information, only about a quarter trust it to provide accurate guidance.
This is not simply a knowledge gap. It reflects a broader understanding of what care means.
AI may improve access to information. But it doesn’t replace care.
This is not an argument against AI. In fact, AI may be one of the most powerful tools available to strengthen Canada’s health-care system – if used correctly.
The most promising applications are not those that attempt to replace clinicians, but those that support the system around them.
In Canada, emerging evidence shows that AI can help analyze large health datasets to identify population-level patterns and improve system planning, while also reducing administrative burden on clinicians. A recent report from the Canadian Association of Drug and Technologies in Health highlights that AI tools are already being used to support diagnostic accuracy, automate documentation and improve access to care through scheduling and triage systems.
Administrative burden is a major driver of burnout among healthcare providers, and AI-enabled tools – such as clinical documentation assistants – have been shown to significantly reduce time spent on paperwork.
These are system-level improvements – not shortcuts.
They align with the structure of Canadian health care, which prioritizes equitable access, evidence-informed decision-making and patient-centred care.
Our health-care systems face real challenges: an aging population, increasing demand and limited system capacity. But we also have a choice. We can invest in expanding medical education, increasing training capacity and strengthening the workforce needed to meet future demand.
Or we can rely on technology to stretch a system that is already under strain.
The future of Canadian health care should not be defined by how efficiently we move patients through the system. It should be defined by how well we care for them. And that requires more than technology. It requires people.
