The juniors we don’t train today are seniors who’ll fail us tomorrow

From Dr Helmy Haja Mydin

An investor friend recently shared how happy he is because he is spending less on junior staff and interns. To quote him: “AI is cheaper and can do more for me than a bunch of inexperienced kids”.

He is not the only one who thinks that way.

In offices, hospitals and trading floors everywhere, senior personnel are choosing to use artificial intelligence (AI) over young graduates.

The turnaround time is quick, and the output is usually good enough. In the end, the young analyst may even be fired.

In the short term, this is efficient. Output is faster and cheaper.

But while we get efficiency, it raises one obvious question: if the junior never does the work, how does he even learn to do it?

That aside, even experienced practitioners are affected. Researchers in Poland compared how doctors performed unaided colonoscopies in the three months before an AI detection tool was introduced against the three months after.

The doctors who had become accustomed to the AI were measurably worse without it. Their adenoma (a pre-cancerous growth) detection rate fell from 28.4% to 22.4%.

An independent commentator called it the first real-world clinical evidence of AI-driven deskilling affecting patient outcomes. It is an unsettling finding for experts: the rot is not confined to the untrained.

In medicine especially, the risk is high.

A finance junior who cannot check a model loses money. A doctor who cannot appraise a clinical trial, and who defers to a fluent, confident, wrong machine, harms a patient.

Understanding statistics and methodology makes the difference between knowing that a drug works and merely believing an influencer saying so.

Critical appraisal is what separates evidence from the endless stream of plausible nonsense that social media generates daily. Doctors who cannot do that unaided are not doctors. They are a transcription service with a stethoscope.

Beyond medicine, the debate goes on.

While calculators, spreadsheets, and search engines only led to the death of mental arithmetic and the library card index, AI offloads judgement itself, not merely computation.

While this is still an assertion at best, there is no longer any doubt that the market is already restructuring around the assumption.

Britain’s Big Four accountancy firms have cut graduate intakes sharply, PwC by around 6% and KPMG by close to a third, while graduate adverts in UK accountancy have fallen 44% year on year.

PwC’s own analysis of more than a billion job adverts found that AI-exposed junior roles are now seven times more likely to demand traditionally senior skills, leadership, strategic thinking, than the least-exposed ones.

We are, in effect, asking school-leavers to arrive “pre-seniorised” – possessed of judgment we have simultaneously removed every opportunity for them to acquire.

Some firms, McKinsey among them, insist on the opposite and are hiring more graduates on the theory that AI makes fresh minds more valuable.

We are willing, as a civilisation, to divert staggering quantities of capital, electricity and fresh water into data centres on the promise of these tools. That may prove a fair trade.

But we are spending real, finite resources to automate the very tasks that people used to become competent in, while investing almost nothing in the humans who must eventually supervise, correct and perhaps even outlive the machines.

None of this is an argument against technology. It is to keep training juniors to do the hard thinking even when the machine could do it faster, mentorship that reviews AI output rather than replaces the reviewer, the discipline to be slower on purpose where slowness is how learning happens.

The head of Amazon’s cloud division called replacing junior staff with AI one of the dumbest things he had ever heard, warning that a company with no one coming up through the ranks eventually explodes on itself.

On the talent pipeline, he is right.

The trouble is that none of this is what unsupervised incentives will produce. Training the young is a cost borne now for a benefit reaped years later, by which time investors would have exited or the executive who cut the scheme would have moved on.

So it falls to those of us who chair boards and run institutions to make the unfashionable case: that the worth of an organisation is not only its next quarter but its next generation.

It is not something the market will correct in the short-term, it is something that needs to be driven by long-term planning and government policies that will help fill the gap.

If we do not invest in the young today, we are not saving money. We are borrowing it from a future that will arrive staffed by seniors we declined to train.

Dr Helmy Haja Mydin is a consultant respiratory physician and chairman of the Social & Economic Research Initiative.

The views expressed are those of the writer and do not necessarily reflect those of FMT.