When Kristalina Georgieva warned in Davos that artificial intelligence is about to hit the labour market like a “tsunami,” most headlines focused on the scale of the wave.
I think they missed the more uncomfortable detail: where it lands first.
Not on senior executives.
Not on highly specialised engineers.
But on the first rung of the career ladder — the part young workers are supposed to climb.
According to the IMF, around 60% of jobs in advanced economies will be affected by AI, with roughly 40% worldwide facing disruption, automation, or deep task reshaping. That’s the statistic everyone quotes.
The quieter truth is this: the tasks disappearing fastest are the ones entry-level workers do.
The Jobs That Taught You the Job Are Going Away
Entry-level roles were never glamorous. Junior paralegals summarising documents. Graduate analysts are cleaning spreadsheets. Junior copywriters are drafting first passes that no one ever sees.
Those jobs existed for one reason: to turn beginners into professionals.
Now, many of those tasks are being handled — faster and cheaper — by AI systems that don’t need training, mentoring, or career progression.
I’ve spoken to recent graduates who aren’t struggling because they lack skills. They’re struggling because the jobs they trained for no longer exist in recognisable form. The ladder isn’t broken — it’s being quietly removed.
That’s the real paradox of AI-led productivity gains: we’re optimising the middle and the top of organisations while hollowing out the bottom.
This Isn’t Job Loss. It’s Job Access Loss
Economists often frame automation as displacement. Someone loses a job; another job appears elsewhere.
That model breaks down when entry points vanish.
If companies stop hiring junior roles because AI can handle the “basic” work, how do workers ever reach the advanced roles AI supposedly complements? Experience becomes mandatory — and impossible to acquire.
This is how you end up with a generation that did everything right and still can’t get in.
Not unemployed.
Just locked out.
A Messy, Lopsided Fallout
The IMF is blunt about the risk: AI will likely widen inequality before it narrows it — if it ever does.
Workers whose roles are enhanced by AI will see productivity gains and wage growth. Workers whose jobs are automated or bypassed will face stagnation or disappearance. Entire career paths may compress upward, skipping the years where people used to learn by doing.
In emerging markets, the picture is even more uneven. While fewer jobs may be immediately automated, the lack of digital infrastructure and reskilling pipelines means workers have less protection when disruption arrives.
This isn’t a clean transition. It’s a messy, lopsided fallout.
Why the Speed Changes Everything
We’ve been here before — mechanisation, computers, the internet. But those transitions unfolded over decades.
AI is moving on a software timeline.
Universities update curricula every few years. Labour policy moves more slowly. Social safety nets are slower still. Meanwhile, AI capabilities are leaping forward every quarter.
That mismatch — not the technology itself — is what should worry policymakers.
At Davos, even tech leaders quietly acknowledged the risk: if AI’s gains concentrate too narrowly, public trust in the technology won’t just erode. It will snap.
The Uncomfortable Choice Ahead
The IMF isn’t arguing for slowing AI down. That ship has sailed.
The real choice is whether governments and institutions acknowledge what’s happening early enough to intervene — or wait until an entire cohort of workers discovers that the door closed behind them without warning.
Reskilling matters. Education reform matters. Labour protections matter.
But none of it works if we keep pretending this is just another automation cycle.
It isn’t.
AI isn’t just changing what work looks like.
It’s changing who gets to start.
And if we lose that first rung, rebuilding it later will be far harder than admitting the problem now.
Related: The AI Bubble Is a Bubble — And Everyone Pretending Otherwise Knows It
