Is AI shutting young people out of their first job?

Irene Chang is 21 and has just graduated in engineering from Georgia Tech, one of the most respected technical universities in the US. Since September she has sent around 450 applications for junior data analyst roles. She has had 19 interviews and no offers.¹

Irene has a hunch: many of the tasks a new analyst used to start with are now done by AI, and done fast. If you or your kids are about to look for work, the question is hard to avoid: is AI shutting young people out of their first job?

The short answer: AI isn't putting a whole generation out of work, but in some fields it is narrowing the way in. In the US, recent graduates face 5.7% unemployment, against 4.1% for all workers.¹ Whether AI is the main reason is still an open fight. At the end you'll find what you can actually do if you're looking for that first job.

Why so many people fear for newcomers

For the International Monetary Fund and many economists, young people will be the first to feel AI at work. The IMF's head, Kristalina Georgieva, has called it a "slow-moving tsunami" that will affect around 40% of jobs worldwide.² Three arguments sit behind that alarm.

The data. The jobs AI does best are exactly the ones losing their youngest workers. In the US, 22-to-25-year-olds in the most AI-exposed occupations now have 19% less employment than people their age in less exposed jobs.³ The gap grows every year, and experienced workers in those same occupations don't show it.

The companies. Employers are hiring fewer graduates, and they say AI is part of the reason. In the UK, graduate job ads fell 45% in a year, to their lowest level since records began in 2016.⁴ The cut isn't coming through layoffs. It comes through jobs that are simply never posted.

The learning. What a beginner used to do is what AI does best: summarising, drafting, basic coding. Stanford economist Erik Brynjolfsson puts it this way: AI learns from knowledge that has been written down, the same knowledge a new graduate brings.¹ Veterans have what only experience gives. And if nobody does the junior work, nobody learns to become senior.

All three arguments are serious. But look at the data more calmly and each one tells a less simple story.

Is AI really taking entry-level jobs?

The data, under a different lens. The fall in junior hiring started before ChatGPT existed. A study from the universities of Warwick and Oxford looked at hundreds of millions of hires in the UK, US, Canada and Australia.⁵ By 2025, the share of hires going to early-career workers was 8% to 11% below 2019.

The same study found a clearer culprit than AI: remote work. Training a beginner from a distance is hard, so many companies prefer someone with experience. Once remote work is taken into account, the AI effect almost disappears. That won't comfort Irene, but it changes her question: maybe the problem isn't that an AI does her job, but that nobody wants to teach her the job over a video call.

The companies, by what they do. The firms investing most in AI are hiring more juniors, not fewer. A study of more than 21,000 US companies found that the heaviest AI spenders grew their entry-level headcount by 12% in two years.⁶ It comes from Ramp, a payments company, using its own data.

Planning fewer hires is also not the same as making them. Some economists believe companies are blaming AI for cuts that really come from interest rates or from hiring too much after the pandemic.⁷ In 2025, only about one in twenty US job cuts was officially put down to AI.

The learning, in a new shape. The first rung isn't disappearing; what it takes to climb it is changing. A junior who can check and fix what AI produces adds value sooner than one who can only do the work by hand. And the same Stanford study shows that in jobs where AI helps people rather than replacing them, employment is holding steady or growing.³ Knowing a field well and using AI with judgment is, for now, a rare combination.⁸

So what can someone looking right now actually do?

What you can do if you're looking for your first job

• Get real experience early. Do internships, your own projects or a portfolio that shows what you can do: a degree helps, but on its own it's no longer enough.
• Pair a field with AI. Knowing one sector well and using AI with judgment is the profile employers struggle to find.⁸
• Learn to review, not just produce. Checking and correcting what AI produces is already an entry-level task.
• Look for places where you learn in person. An office or hybrid role with someone to teach you can be worth more than a remote one.⁵
• Look beyond the obvious sectors. In the UK, construction, travel, teaching and manufacturing were the only sectors where graduate job ads went up.⁴

The question nobody can answer yet

Nobody knows yet whether this is an AI effect or a sign of the times. The serious studies disagree because they measure different things: some look only at the most exposed jobs, others at the whole economy. We don't know if it's temporary either. The Ramp study sees hiring rise six to twelve months after a company adopts AI, so one door may close before another opens. What we do know is that anyone job hunting today can't wait for economists to agree.

So, is the door closing?

AI isn't shutting the door to the first job, but it is narrowing it and changing what it takes to get through. The bigger risk isn't for Irene, who will find her place, but for companies that stop training juniors: in ten years they won't have any seniors.

Now it's your turn, in one click: do you think AI will make it harder for you (or your kids) to land a first job?