For over a year, headlines have warned that artificial intelligence is stealing entry-level jobs, particularly in white-collar sectors like software development. The narrative seemed intuitive: as generative AI tools like ChatGPT exploded in popularity, youth unemployment rates climbed. But new analysis of millions of job postings across five Western economies reveals this direct link may be a mirage – and the real culprit looks more like traditional economic pressures than technological disruption.
The Hiring Data That Tells a Different Story
Exclusive analysis of job advert data from LightCast, covering the US, UK, France, Germany, and Netherlands, shows no clear evidence that AI is behind the slowdown in early-career employment. When examining hiring patterns rather than employment figures, both the “disruption hits young people most” and “disruption started with ChatGPT” narratives collapse.
The decline in hiring for both junior and senior positions began not in November/December 2022 when ChatGPT launched, but in mid-2022 as interest rates rapidly climbed. This pattern suggests a hiring slowdown in response to worsening financial conditions, not AI displacement. The respective slowdowns in hiring for entry-level and experienced positions show minimal difference – it’s a cooling across the board, not a specific reluctance to hire juniors.
Why Youth Unemployment Spikes Anyway
If hiring is slowing equally for all experience levels, why does youth unemployment spike higher? A new report from Google economists Zanna Iscenko and Fabien Curto Millet for the US Economic Innovation Group explains this paradox: an economy-wide hiring freeze always shows up as rising youth unemployment because older workers simply remain in their jobs, while new labor market entrants have nowhere to go.
Stephen Isherwood, joint chief executive of the UK’s Institute of Student Employers, confirms this perspective: “Most of the employers I talk to say that actually a lot of the data and the headlines are conflating a tough economic climate, nervousness about hiring, cost pressures… I haven’t actually spoken to a single employer who says ‘d’you know what, AI’s taken these jobs, so we’ve reduced our intake because of it.'”
The AI Adoption Correlation Fallacy
The analysis reveals another critical insight: the types of firms most likely to adopt AI – generally larger companies concentrated in tech and professional services – are also exactly the types of firms most sensitive to rate hikes and other macroeconomic shocks. The same basket of theoretically AI-exposed occupations that pulled back on hiring steeply in 2022 also did so when COVID hit in 2020, long before large language models were on the scene.
While hiring did contract slightly more steeply in mid-2022 at AI-adopting firms than non-adopters, that gap has since narrowed and in some cases reversed. This contradicts what we’d expect if AI job displacement were driving the pattern, but aligns perfectly with AI-adopters being coincidentally more sensitive to an economy-wide shock that then spreads to all firms.
The Physical AI Revolution: Real Displacement in Robotics
While generative AI’s impact on white-collar jobs may be overstated, physical AI – the convergence of artificial intelligence and robotics – is quietly transforming industries in ways that could have more tangible employment implications. According to the International Federation of Robotics, over 4.7 million industrial robots were in operation in 2024, with annual installation growth exceeding 500,000 units – twice the rate of ten years ago.
China accounted for 54% of all new robots installed in 2024, and healthcare showed nearly twice as many new robots installed compared to the previous year. Stephan Schlauss, global head of manufacturing at Siemens, notes that “AI-enabled robots that pick and place different parts and materials in our assembly lines reduce automation costs by 90 per cent.”
This physical AI revolution extends beyond traditional manufacturing to sectors like healthcare, agriculture, and logistics. Amazon uses over 1 million robots in its fulfillment centers, while companies like Foxconn and Apple increasingly rely on robotic automation. The robotics sector saw 381 deals transacted in Q1 2025 alone, a 20% increase from 2024.
The Graduate Job Market Reality
Meanwhile, graduates face a challenging landscape regardless of AI’s direct impact. Global hiring remains 20% below pre-pandemic levels, with UK graduate hiring reduced by 8% in the last academic year. Some surveys show 140 applications per graduate vacancy in the UK.
Emily Chong, a recent graduate from University College London, expresses the frustration many feel: “For most of our lives we have been told that working hard and getting good grades will get us to where we want to be. I realise this isn’t the case.”
London mayor Sadiq Khan warns that the capital will be “at the sharpest edge” of changes wrought by AI given so many white-collar jobs are based in the city, predicting that “entry-level jobs will be the first to go.”
AI’s Real Impact: Productivity and Process Changes
The evidence suggests AI’s current impact may be more about changing how work gets done than eliminating jobs entirely. David Gewirtz, a software developer who used Claude Code to port an iPhone app to Mac, found the experience required significant oversight: “Vibe coding trades creativity for coordination and oversight.” He notes that “AI shines when developers relentlessly lead, test, and correct.”
This aligns with broader observations that AI serves as a force multiplier for experienced professionals rather than a replacement for entry-level workers. The technology may be contributing to hiring slowdowns indirectly – not by displacing jobs, but by creating uncertainty and the belief that existing staff can become more productive with AI tools.
Looking Ahead: What Comes Next?
While current data doesn’t support dramatic AI job displacement claims, the landscape continues to evolve. As one analyst notes, “Just because we don’t see evidence of big AI-linked job displacement trends yet doesn’t mean we won’t.” The agentic phase of AI – where systems can take more autonomous actions – may present different challenges in the medium term.
For now, businesses and policymakers should focus on the immediate economic factors affecting hiring while preparing for longer-term technological shifts. The convergence of economic pressures and technological change creates a complex landscape where distinguishing cause from effect requires careful, data-driven analysis rather than sensational headlines.