Tony Frost and Christian Dippel are associate professors of business, economics and public policy at Western University’s Ivey Business School.
An old maxim in development economics holds that if a country discovers oil after it has built strong social and political institutions, oil can be a blessing. However, if a country discovers oil before those institutions are developed, it often becomes a curse.
Artificial intelligence is producing a similar divide in Canada today – not between countries, but between generations. For workers who already have experience and expertise, AI is turning into a powerful productivity booster. For young Canadians trying to begin their careers, it threatens to wipe out the entry-level rungs they need to build those very skills.
We see this divide firsthand. In our work with companies and mid-career professionals looking to leverage the possibilities of AI, we see how quickly the technology can extend the reach of people who already have judgment and subject-area expertise.
In our classrooms with university students, we see the opposite: young Canadians watching AI take over the very tasks – first drafts, initial analysis, basic coding and spreadsheet work – that once helped them build competence. The same technology that lifts seasoned professionals is making it harder for new workers to gain a foothold.
The reason is simple. AI is a skill amplifier, not a skill creator.
Experienced professionals have three powerful advantages when it comes to using AI. First, they know how to ask the right questions to generate high-quality output – be it code, research or insights from data. Effective prompting is mostly a matter of applied judgment: the ability to visualize what “good” looks like, to steer the model toward a desired outcome and to specify the constraints that matter or the pitfalls to avoid. Seasoned experts tend to have the internal capacities that make well-designed prompting intuitive.
Second, workers with a deep base of knowledge are able to detect when AI-produced output is wrong or off-track. Spotting hallucinations requires mastery: having a sense of what is plausible, what violates known facts and what simply doesn’t smell right. When you have years of education and experience to draw on, it becomes much easier to filter out nonsense and to use AI effectively at high velocity.
Third, experts know how to assemble pieces into a coherent whole. AI is extraordinarily good at generating components – for example, code snippets, text blocks, tables, outlines and conceptual options. But transforming these components into integrated, meaningful wholes is a fundamentally different skill, one that is learned mostly through practice and accumulated understanding.
For early-career workers, those same three capabilities are precisely what they lack – and the tasks that once helped them build these capabilities are disappearing. A decade ago, new graduates learned by doing the work that senior colleagues did not have time for: drafting the first version of the report, building the initial model, coding the basic functions, conducting the background research. Those tasks were imperfect, messy and slow – but they were essential.
The solution is not to resist AI, nor to pretend that junior roles can be frozen in time. The solution is to rebuild early-career pathways so they cultivate the skills that AI cannot replicate. That means reshaping university programs, internships, apprenticeships and early-career roles to focus more on judgment, integration, applied problem solving and contextual reasoning – not just task execution.
Firms also have a critical role. Employers need to create new on-ramps that deliberately expose young workers to higher-judgment tasks earlier, supported by structured feedback and supervision. The old apprenticeship model – learning by watching and doing alongside more experienced colleagues – needs to be rediscovered and modernized for an AI-intensive world.
One practical approach is to revive “studio-model” rituals borrowed from fields such as architecture and design: regular pin-up sessions where junior staff present work in progress and more experienced colleagues walk through their reasoning, critique assumptions and narrate how they would attack the same problem.
Another promising development uses AI itself as a tool for apprenticeship. Some firms now ask junior employees who rely on AI for first drafts to “show their prompt work” – explaining how they framed the task, why they chose certain prompts and what they kept or discarded as they iterated. This simple practice forces meta-cognition, reveals gaps in judgment and gives mentors a clear window into how early-career workers are learning to think.
With one of the highest post-secondary education rates in the world, Canada has an opportunity to turn AI into a blessing for workers at all career stages, not just those with established credentials and a deep practical base of experience. AI, like oil, is a resource. Whether it becomes a blessing or a curse will depend on what we build around it.
This column is part of Globe Careers’ Leadership Lab series, where executives and experts share their views and advice about the world of work. Find all Leadership Lab stories at tgam.ca/leadershiplab and guidelines for how to contribute to the column here.