AI is transforming health care, but not everyone will share the benefits

Prime Minister Mark Carney recently launched AI for All — an ambitious five-year national strategy. It aims for $200 billion in economic growth, 250,000 new jobs and a surge in AI adoption across Canada, from 12 per cent to 60 per cent by 2034.

As someone who works daily at the intersection of medicine, health systems and equity — I’m a hospitalist physician and the inaugural senior associate dean for health equity and systems transformation at the Cumming School of Medicine, University of Calgary — I read the strategy with genuine enthusiasm and an urgent concern: has equitable implementation even been considered?

The strategy is largely silent on one problem: transformative technologies do not reach all communities equally, and the gap between intent and impact is widest where stakes are highest.

In health systems, those stakes are life and death.

AI is already transforming Canadian health care. Will that transformation be equitable and inclusive of all Canadians?

Black, Indigenous, racialized patients suffer

Canada’s AI adoption rate sits at just over 12 per cent — one of the lowest rates among peer nations. The AI for All strategy names this as a risk.

But there is a second gap demanding equal attention: within that 12 per cent, AI tools are not being adopted uniformly. They cluster in institutions with existing resources and technical infrastructure. They are designed, trained and validated on datasets that under-represent Black, Indigenous and racialized Canadians.

They embed the assumptions of the systems that built them. And when they fail — when the diagnostic algorithm performs worse for a Black or Indigenous patient, when the clinical decision tool reflects a history of exclusion — that failure is not random.

It is patterned and predictable. An equity failure, by definition.

The strategy’s flagship AI Missions Program could transform Canadian medicine. But ambition without equity infrastructure does not lift all boats. It raises the tide for those who already have one. Institutions with the fastest and deepest technical talent will move first. Those serving the most marginalized populations will not.

This is not pessimism. It’s a question of whether equitable implementation has been built into the strategy’s foundation — and Canada has the capacity to do exactly that, if the will exists.

Implementation failure is equity failure

At the Cumming School of Medicine, I established the Health Equity AND Systems Transformation (HEST) Innovation Lab. The lab hosts a team of implementation scientists, clinicians, equity scholars and community knowledge-holders under the scientific direction of Dr. Nonsikelelo Mathe, who teaches that implementation failure is equity failure.

The two are inseparable. With investigator funding exceeding $1.16 million across six active awards, our platforms include the Black Health Equity Network initiative — a bilingual digital infrastructure co-developed with the Association of Faculties of Medicine of Canada to dismantle anti-Black racism in Canadian medicine.

They also include SafeSpeak — an AI-enabled, trauma-informed chatbot interface for safer disclosure of racism and discrimination during the course of medical training.

Our lab is positioned to model exactly the equity and implementation layer Canada’s AI health mission requires.

Teach the history of algorithmic bias

Canada’s AI strategy has an innovation strategy, an adoption strategy and an economic strategy. What it lacks is an equitable implementation strategy.

I have three specific asks for those shaping the framework.

First, require equity impact assessments for all health AI deployments receiving federal funding. No diagnostic tool or clinical algorithm should reach scale without evaluating performance across race, income, language and geography. The methodology exists. The mandate does not.

Second, fund equitable implementation science alongside technical development. Adoption is not implementation. Implementation is the structured process of embedding solutions into complex human systems effectively, sustainably and equitably. It is slower, harder to measure and completely missing in this strategy.

Third, build equity infrastructure into the National AI Literacy Initiative. Training one million students in AI without including the history of algorithmic bias and risks for historically excluded communities will produce practitioners who are technically capable but unprepared for the equity dimensions of the systems they will build.

Who benefits? Who is left behind?

The cost of inaction is not stasis. AI systems that embed bias at the point of adoption become infrastructure — replicated, scaled and increasingly difficult to correct.

Every month without an equitable implementation framework is a month in which health AI tools are trained, validated and deployed without equity as a design requirement. Disparities will not simply persist. They will be encoded, accelerated and entrenched at a scale Canada has never seen before.

Canada’s AI reputation was built on technical excellence and a commitment to AI that is trustworthy and human-centred. AI for All inherits that legacy. The question is whether it will honour it in its architecture, not just its rhetoric.

The future of Canadian AI will not be determined by algorithms alone. It will be determined by who benefits, who is left behind and whether we build equity into implementation before inequity becomes embedded at scale. AI for All is an ambitious promise. Now Canada must decide whether it intends to keep it.