Canadian health care does not need another technology that simply changes the shape of administrative burden. It needs technology that gives clinicians and health-system professionals time back for professional development and clinical work.
Healthy Debate has documented the problem clearly. Primary care faces a digital burden in which technology can add tasks, clicks, screens and cognitive load. Other contributors have argued that thoughtfully implemented, Artificial Intelligence (AI) can reduce administrative burden while strengthening clinical decision-making and patient care.
But what of the next generation of health-care workers?
The latest Stanford Digital Economy Lab employment update found that the employment shortfall for workers ages 22 to 25 in highly AI-exposed occupations widened from 15 per cent in the July 2025 data vintage to 19 per cent by June 2026, primarily because employers are hiring fewer young workers.
Health care should not allow administrative automation to produce the same effect inside its professional and operational pipelines.
An AI scribe can reduce documentation time. Good. The saved time should increase exposure to the patient, the reasoning, the uncertainty and the explanation behind the plan.
AI can summarize a chart or referral history. Use it. Then require the learner or junior analyst to identify what the summary misses, which information is unreliable and what fact would change the decision.
AI can prepare an operational report. Let it. Then give the newer employee responsibility for understanding why access, scheduling, staffing or workflow performance differs across clinics and patients.
AI can draft project requirements for a health-information system. Automate the draft. Then move junior health-IT staff into observing frontline work, testing edge cases and learning why a technically correct workflow can be clinically wrong.
These are better developmental experiences than clerical preparation.
A Healthy Debate article recently argued for AI governance that keeps physicians actively involved in shaping how technology is integrated. The same principle should apply to workforce design. The people closest to care should help decide what work automation removes and what higher-value learning replaces it.
Health-care organizations should measure the time it takes for employees to develop independent competence. How quickly can a junior analyst identify a misleading AI summary? When can a new health-IT employee recognize a workflow design that will fail for clinicians? How soon can a learner use an automated note without losing the habit of constructing an independent clinical picture?
If AI shortens those timelines, it creates capacity that compounds.
On the other hand, if it simply increases throughput or allows organizations to hire fewer junior people, it may reduce burden today while weakening the future bench.
That is particularly dangerous in a system already struggling with workforce sustainability. Healthy Debate has described home and community care as being at a workforce breaking point. Canadian health care does not have excess human capability to waste.
The strongest use of AI is therefore not just documentation reduction. It is learning-time recovery.
Managers should explicitly reinvest some of the time automation saves into case review, mentorship, patient conversations, cross-functional problem solving and supervised handling of exceptions. When an experienced clinician or administrator overrides an AI recommendation, capture why and use the case to teach others.
The widening early-career employment gap seen in other professions should push health-care leaders to protect the developmental function of work before automation changes staffing patterns by default.
The best AI adoption at work should make health-care professionals capable faster while giving them more time for patients and meaningful decisions.
AI can buy back time. Canadian health care should spend some of it on the people who will carry the system next.
