For the past few years, the story about young people and work has had a clear villain. Graduates everywhere — but especially in the US, UK, Canada and Australia — have struggled to land first jobs, and the slump has been worst in white-collar fields like software engineering. The obvious explanation wrote itself: artificial intelligence. Tools like ChatGPT, the argument goes, can now perform the routine tasks once handed to juniors — drafting memos, debugging simple code, summarising documents — while still falling short of what experienced professionals do. Why hire a 22-year-old when a chatbot will do the work for free?
A new working paper by economists Peter John Lambert and Yannick Schindler complicates that tidy narrative. Drawing on roughly 243 million hires and 407 million job postings between 2017 and 2025, the authors searched for another shock that could match the same fingerprint: a force that hurt junior hiring more than senior, white-collar more than blue-collar, and young software developers most of all. Their counter-proposal is the take-off of remote work.
The theory is straightforward. Early-career workers need supervision. They need someone to look over their shoulder, answer a half-formed question, or notice that they are quietly stuck. They also absorb 'tacit' knowledge — the unwritten norms of a profession — by sitting near people who already have it. Remote work adds friction to all of these processes. Bringing a junior up to speed over Slack takes more time and effort than doing it in person, and that extra cost makes firms less willing to take the risk in the first place.
Here's the catch that makes the research more than just another opinion. AI exposure and remote-friendliness are heavily correlated: the same kinds of jobs that AI can assist with are also the kinds that can be done from a sofa. So the two variables had been compounding each other in earlier studies, making AI look guiltier than it was. When Lambert and Schindler control for remote work, the apparent AI effect on junior hiring largely vanishes. The remote-work effect, by contrast, holds up — and crucially, it had already begun shrinking entry-level hiring before ChatGPT's late-2022 debut. Jobs that score high on AI exposure but require an in-person presence, such as receptionist roles, have actually held up comparatively well.
The findings cut against several comfortable assumptions. Not every weird trend in the post-pandemic economy is AI's fault. And the long-running 'return to the office' debate, usually framed as grumpy older bosses versus flexibility-loving Gen Z, looks different in this light. The data suggest that the loudest beneficiaries of an extra day in the office aren't the executives demanding it — they're the juniors who, once back in the building, become visible enough to be trained, mentored and eventually promoted.
The paper does not argue that remote work is uniformly bad. Earlier research by the same author found that fathers in remote-friendly occupations take on more childcare; their partners' earnings and employment rise; remote work has even nudged up birth rates in some places. The story is one of trade-offs, not villains. But for the specific cohort entering the workforce right now, the trade-off has been brutal: a labour market that is merely cool for everyone has been icy for the youngest, and the freezer door, it turns out, may have been propped open not by a chatbot but by a webcam.
Source: https://www.ft.com/content/
Everyone blames ChatGPT for entry-level jobs vanishing. But a new study of 243 million hires says the real culprit is something far less futuristic: working from home.
Across the US, UK, Canada and Australia, hiring of junior workers has been falling steeply since around 2019, especially in white-collar fields like software. The obvious suspect has been AI: tools like ChatGPT can do a lot of the grunt work that used to be handed to new graduates.
But economists Peter John Lambert and Yannick Schindler analysed roughly a quarter-billion hires and 407 million job postings and found that once you control for how 'remote-friendly' a job is, the AI effect on junior hiring largely disappears. The remote-work effect, however, stays strong — and it had already started shrinking entry-level hiring before ChatGPT existed.
Think of an office as an apprenticeship machine. Juniors aren't only paid to produce work; they're paid while older colleagues teach them. Remove the building, and the machine quietly breaks.
If you're picking a college major, an internship, or a first job, this reshapes the calculus. The fully-remote dream role might actually be the riskiest move early in your career — not because you're lazy, but because the people who could promote you literally won't see you working. Burn-Murdoch's blunt line: the biggest beneficiaries of an extra day in the office aren't the bosses demanding it. They're the juniors.
The 'return-to-office' culture war has mostly been framed as old bosses vs. young workers. This research scrambles that story: it suggests Gen Z may have negotiated themselves out of the exact environment where careers are built. Watch for hybrid schedules to harden, for companies to start advertising mentorship as a perk, and — if the pattern holds — for a 'lost cohort' of workers in their 20s who entered the labour market just as the apprenticeship ladder was being quietly dismantled.