Two Papers, Opposite Answers on AI and New Graduates
Two research teams looked at roughly the same question and came back with opposite answers. Ars Technica reports a new working paper from economics researchers at Munich's CESifo arguing that there is no evidence of significant, widespread displacement or of reduced hiring of recent college graduates, measured in absolute or relative terms. That sits against a Stanford study Ars covered last month, which found entry-level employment in so-called AI-impacted occupations lagging well behind other fields. Note what is and isn't in dispute. Neither finding is a claim about what the models can do; both are claims about what employers did. One sees a gap opening in AI-exposed entry-level work, the other sees recent graduates being hired at ordinary rates. That is the kind of disagreement more data settles.
AI-generated discussion · Different perspectives
The CESifo authors' premise is that hiring is the first place this shows up: if AI handles the relatively standardized tasks in many entry-level office jobs, firms can shrink new hiring for simpler roles rather than lay off experienced staff.
Which makes the null result a strange sort of comfort. Ars notes there is some reason to believe 2026's graduating job seekers are more at risk than those who finished a year or two earlier, and the researchers point to a Census survey showing a sharp recent increase in firms replacing a large number of employee tasks with AI, alongside rising AI spending per employee and ChatGPT Enterprise token use.