A widely cited Stanford study on AI and jobs just got a worrying update. Economists Erik Brynjolfsson, Bharat Chandar, and Ruyu Chen have tracked employment data from ADP, a payroll company that processes pay for millions of American workers, since generative AI took off in 2022. Their first paper, published last year, found that workers aged 22 to 25 in jobs highly exposed to AI had 13 percent lower employment than similar young workers in less-exposed fields.
The newest version of that paper, released in August, puts the gap at 19 percent. The jobs most affected are ones like software development and customer service, where AI tools can now do a meaningful share of the actual daily tasks. Jobs that AI mostly assists rather than replaces, including nursing and health aide work, are still growing for young workers.
That split matters. The damage is not spread evenly across the economy, it is concentrated exactly where AI substitutes for junior-level tasks instead of supporting the people doing them. Older workers in those same exposed jobs show no comparable gap at all.
The most useful detail in the research is how the effect actually happens. Companies are not laying off young employees at higher rates than before, they are simply hiring fewer of them. A company that used to bring on five junior analysts now hires two, and hands the rest of the workload to AI tools plus the staff already on payroll.
From a manager's spreadsheet, that looks like nothing happened. From a 23 year old's job search, it looks like the market vanished. Separate labor market data backs this up: entry-level postings and internships have been shrinking while applications per opening keep climbing.
Skeptics have pushed back, arguing this could just be a broadly weak hiring market for new graduates rather than an AI story specifically. That is a fair challenge, since overall entry-level hiring has been soft for reasons that have nothing to do with AI. But the Stanford team tested their finding against interest rate sensitivity, a hiring pullback in tech after its 2021 boom, and remote work shifts, and the gap held up in every version of the test.
If this trend continues, the business risk is not really about this year's hiring budget. It is about where mid-level and senior talent comes from in five or ten years. Junior roles have always been the training ground where people learn a job well enough to eventually run it.
There is also an opening here for companies willing to go against the trend. If large employers keep pulling back on junior hiring in AI-exposed roles, a smaller or mid-sized business that keeps training young workers, using AI to make them more productive rather than to replace them, gets first pick of talent that used to go to bigger competitors. That is not a consolation prize, it is a genuine advantage while everyone else cuts the bottom rung of the ladder.