Evidence available so far does not support claims that artificial intelligence is already causing large, economy-wide job losses, according to a Stanford Institute for Economic Policy Research review. The brief finds a softer labor market across occupations, while identifying possible pressure on younger workers in AI-exposed roles as an unresolved area of concern.
The review compares unemployment among occupations with different exposure to AI. Since 2022, unemployment in the most exposed fifth rose by 0.77 percentage points, while the least exposed fifth experienced a slightly larger 0.85-point increase. The authors interpret that pattern as general labor-market weakening rather than a distinct wave of AI displacement. Employment in highly exposed occupations has remained broadly stable, and growth in coding-heavy work has slowed but stayed positive.
Job-posting data likewise provide little evidence of a broad contraction tied to AI. The brief says postings for software developers grew faster than those for other occupations over the preceding year. It also cites research finding that firms adopting enterprise AI increased employment by 10% over the following two years, with the effect concentrated among companies spending the most per employee.
Corporate layoff announcements that cite AI deserve careful interpretation, the authors argue. Some limited reductions may reflect automation, but others may free cash for AI investment or correct pandemic-era overhiring. Human-resources executives reported effects appearing through combined roles and decisions not to hire for work containing many automatable tasks. Such changes can affect individual workers without producing a clear aggregate shock.
Recent graduates present a more complicated picture. Their unemployment rate reached 5.6% in early 2026, 1.6 percentage points above its level three years earlier. Research has documented falling employment among early-career software developers and customer-service representatives after 2022, while employment among older workers in the same occupations held steady or grew. Other US and UK studies have reported similar weakness in hiring young people for AI-exposed positions.
Causation remains difficult to isolate. The US Federal Reserve began raising interest rates months before ChatGPT’s November 2022 release, and some research locates the hiring decline after that monetary shift but before the chatbot appeared. Pandemic overhiring and the move to remote work may also have favored experienced candidates. With additional controls, one prominent study did not find notable entry-level declines until 2024, when AI adoption and model capabilities were further advanced.
The review therefore rejects both complacency and an immediate-apocalypse narrative. Aggregate measures show no sweeping AI employment collapse, but they can conceal concentrated disruption. Determining whether early-career weakness reflects AI, macroeconomic forces or both remains an active empirical question.
