AI Isn't Killing Junior Jobs. It's Rewriting Them.
AI-adopting firms are growing senior roles faster than junior ones. IBM's hiring U-turn shows why the entry-level job needs redesigning, not cutting.
In 2023, IBM paused hiring for around 7,800 roles it expected AI to replace.
This year it’s tripling its US entry-level hiring.
Every few months a study lands and LinkedIn declares the graduate job dead. The latest came from Stanford last month.
At firms adopting AI, senior employment rose 6.7%. Junior employment fell about 2.5%.
The study comes from Bharat Chandar at Stanford and Bouke Klein Teeselink at King’s College London. They went through 1.25 billion job postings and 154 million employment records across 41 countries.
Read quickly, it looks like AI is killing the entry-level job.
The detail tells a different story.
More seniors, slightly fewer juniors
The researchers spotted AI adopters from their job ads - anyone asking for ChatGPT, Claude, copilots and the like. Then they compared those companies’ overseas offices with similar ones that hadn’t adopted.
- Total employment grew about 3.3%
- Senior employment rose 6.7%
- Junior employment fell about 2.5%
- The junior share of the workforce dropped 2 percentage points, in 23 of the 31 countries with reliable estimates
The authors say the change “primarily comes from growth in senior employment rather than from a fall in junior employment.”
So most of the shift is companies hiring more experienced people into AI-heavy roles, and fewer juniors alongside them.
Same AI, different outcome
In August, Erik Brynjolfsson, Bharat Chandar and Ruyu Chen updated their Canaries in the Coal Mine research on US workers aged 22 to 25.
In the jobs most exposed to AI, employment for that age group is now about 19% below where it would otherwise be. A year earlier the gap was 15%.
It’s mostly happening through fewer hires, not more people being let go.
Then they split those jobs by how AI is used in them.
Where AI mostly automates the work, employment fell. Where it mostly helps people do the work, employment was “flat or rising, especially for experienced workers.”
The technology was the same. What changed was what the company decided the AI was for.
That matches what I’m seeing. It’s not across the board yet, but at some tech companies there’s an uptick in junior hiring.
I think the difference is the job description.
IBM changed its mind
That 2023 pause came from CEO Arvind Krishna, who told Bloomberg he could see those back-office roles being replaced by AI over five years.
As U-turns go, that’s a big one.
Its chief HR officer, Nickle LaMoreaux, didn’t pretend AI had gone away. “The entry-level jobs that you had two to three years ago, AI can do most of them,” she said in February.
Her answer: “You have to rewrite every job.”
Junior developers now spend less time coding and more time with customers.
Entry-level HR staff step in when the HR chatbot gets it wrong, fix what it said, and deal with managers directly.
Someone has to tell the chatbot it’s wrong. At IBM, that’s now a junior’s job.
Nobody is born a senior hire
Most companies cutting junior roles have been doing so for efficiency reasons. That is step one - hand the work over to AI.
That works in the immediate short term. The tasks get done, the headcount line shrinks, and the finance director is happy for a quarter.
The trouble is every competitor can do the same thing. Cost savings from AI converge quickly, and once everyone has them they stop being an advantage. (More on that in The companies winning with AI are not the ones cutting costs.)
And the seniors these firms are busy hiring had to learn the job somewhere. Nobody is born a senior hire.
LaMoreaux put it more bluntly. The most successful companies in three to five years, she said, “are those companies that doubled down on entry-level hiring in this environment.”
Fiverr’s small jobs are disappearing
Fiverr is seeing the same thing from the buying side.
It has almost 750,000 fewer active buyers than a year ago - down from 3.4 million to 2.7 million. But the ones still buying spend more, $368 each against $318, and clients doing $1,000+ projects grew 13%.
Its CEO, Micha Kaufman, says AI is absorbing the “high-volume, low-value, transactional tasks,” while bigger projects still need human expertise and accountability.
Many of those small Fiverr jobs were basically just a prompt - the same kind of work juniors used to learn on. But someone still has to decide which prompts are worth running.
I’m seeing a similar shift. One company I know cut the number of landing pages it outsources. Most are now built from a template using AI, and only the important ones get a human’s full attention.
Before you decide not to backfill
Most small and mid-sized companies aren’t running graduate schemes. But the same decision comes up every time a junior leaves and someone asks, “do we still need that role?”
Before saying no, I’d ask three questions:
- Which tasks in this role can AI do today? Task by task, written down. “Most of it” doesn’t count.
- What would this person do with the time? Customers, checking AI output, and the work nobody has had time for are the obvious candidates.
- Who checks the AI’s work now? At IBM it became part of the junior job. In a lot of businesses it lands on the most senior person by default, which is the most expensive checking you can buy.
If the answers point to a different job, write the different job.
(On which skills to hire for once the role changes, see What to hire for when AI is doing more of the work.)
Final thought
AI-adopting firms want more experienced people. They’ll need somewhere to get them from.
I’d rather be IBM in 2026 than IBM in 2023.
Frequently asked questions
Is AI reducing entry-level jobs?
At AI-adopting firms, the junior share of the workforce is falling. A 2026 study of 1.25 billion job postings across 41 countries found senior employment rose 6.7% and junior employment fell about 2.5%. The authors say most of the shift comes from senior hiring growing, not juniors being cut.
Why are young workers in AI-exposed jobs being hired less?
Stanford’s research on US workers aged 22-25 finds employment falling where AI is used to automate tasks, and flat or rising where it’s used to help people do the work. The drop is mostly fewer hires, not more people being let go.
What did IBM change about entry-level roles?
IBM is tripling US entry-level hiring in 2026 after rewriting those roles, three years after pausing hiring for roles it expected AI to replace. Junior developers spend more time with customers and less time coding. Entry-level HR staff step in and correct things where the HR chatbot falls short.
Should my business stop hiring juniors because of AI?
Not by default. Map which tasks in the role AI can do, then decide what the person does with the time that frees up. Cutting the role saves cost in the short term. Redesigning it keeps building the experience your future senior people will need.
Connected Paths works with CEOs on the organisational and operational changes that make AI adoption compound. Start here.