The junior hiring crisis everyone blamed on AI? Turns out we've been looking at the wrong screen this whole time.
The Summary
- Researchers analyzed 243 million new hires and 407 million job postings across the US, UK, Canada, and Australia, finding remote work exposure predicts weaker junior hiring far better than AI exposure does
- The narrative that AI is wiping out entry-level jobs may be hiding what's actually happening: companies can't figure out how to train and supervise junior workers remotely
- While tech CEOs warned about AI eliminating half of entry-level white-collar jobs, the data points to a management problem, not a technology replacement problem
The Signal
For two years, the story has been simple: AI is eating junior jobs. Anthropic's Dario Amodei warned AI could eliminate half of entry-level white-collar positions. Google DeepMind's Demis Hassabis said he was already seeing hiring slowdowns for junior roles and internships. The narrative fit. New technology threatens to automate the grunt work that used to be the proving ground for young workers.
Except the data tells a different story. Peter John Lambert from the London School of Economics and Yannick Schindler from the Ellison Institute of Technology pulled résumé data covering 243 million new hires and job posting data covering 407 million positions across four countries from 2017 to 2025. Their finding: work-from-home exposure is a much stronger predictor of weaker junior hiring than generative AI exposure.
"Our findings point strongly towards WFH exposure as a better predictor of the decline in relative early-career hiring."
Here's what's actually happening. Companies that went remote discovered something uncomfortable: they don't know how to train people who aren't in the room. The informal knowledge transfer that happens when a junior employee sits near a senior one, overhears conversations, sees how decisions get made in real time. That's gone. What replaces it? Scheduled Zoom calls that feel like an imposition on everyone's calendar. Slack messages that go unanswered for hours. Training that gets documented but never actually absorbed.
The research challenges the growing narrative around AI's employment impact, at least for entry-level roles. The problem isn't that ChatGPT can do what a junior analyst does. The problem is that remote work broke the apprenticeship model that corporate America ran on for decades, and nobody built a replacement system that actually works.
Key implications:
- Companies are avoiding junior hires not because AI replaces them, but because remote work makes them harder to manage
- The entry-level hiring crisis is a training and supervision problem masquerading as an automation problem
- Aggregate employment remains stable and recent assessments show limited AI impact on headline numbers, but structural changes are hiding beneath the surface
This matters because we've been solving for the wrong problem. If the real issue is remote supervision and training, then the solution isn't about competing with AI or finding "AI-proof" skills. It's about rebuilding how companies onboard and develop talent in a hybrid or remote world. Or accepting that certain companies will need to bring people back to offices specifically for the first 2-3 years of their careers, which is a harder sell in 2026 than it was in 2019.
The Implication
If you're running a company, this research suggests your junior hiring problem isn't about AI threat, it's about management capability. The companies that figure out remote apprenticeship structures, async knowledge transfer, and distributed mentorship will have access to junior talent pools their competitors are avoiding. That's a competitive advantage hiding in plain sight.
For early-career workers, this is actually good news. The bottleneck isn't that you're replaceable. It's that companies haven't built the infrastructure to grow you remotely. Seek out companies that have cracked this problem, or roles that come with explicit in-person training periods. The crisis isn't that the work is going away. It's that the learning paths are broken.