A new report from Anthropic is cutting through the noise in the AI jobs debate, offering a more precise look at where artificial intelligence is actually affecting work. The study, “Labor market impacts of AI: A new measure and early evidence,” by Maxim Massenkoff and Peter McCrory, moves beyond theoretical speculation to track real-world AI use in professional settings. The key finding: AI is reshaping the labor market, but not through mass layoffs — instead, it’s quietly narrowing the on-ramp for entry-level workers.
The report introduces a measure called “observed exposure,” which combines theoretical task feasibility with actual usage of Anthropic’s Claude AI in workplace settings. This approach gives more weight to automated use over simple assistance, making it a better gauge of substitution risk. The authors argue that AI disruption won’t arrive as one dramatic event but will spread through specific tasks, occupations, and hiring decisions long before broad labor market indicators reflect it.
The data reveals a stark gap between theoretical potential and real-world deployment. For computer and math occupations, theoretical exposure reaches 94%, but observed current coverage is only 33%. That reality check matters because it shows AI’s labor market effects are still constrained by factors like legal risk, compliance, software integration, and quality control. As noted in our coverage of Hassett’s forecast of a robust job market despite geopolitical pressures, the overall employment picture remains stable for now.
The occupational rankings confirm AI is hitting specific sectors hardest. Computer programmers top the list at 75% observed coverage, followed by customer service representatives and data entry keyers at 67%. At the other end, 30% of workers are in occupations with zero measured coverage, including cooks, bartenders, dishwashers, lifeguards, and mechanics. AI is clearly concentrated in digital, language-heavy, structured work where outputs are easier to generate and integrate.
Importantly, the authors detect no systematic increase in unemployment for workers in the most exposed occupations since late 2022. That directly challenges the popular claim that generative AI has already produced a broad labor market collapse. Outside research from Yale’s Budget Lab and earlier work by Acemoglu, Autor, Hazell, and Restrepo point in the same direction: AI adoption affects hiring patterns and skill requirements at the establishment level, but aggregate effects remain difficult to detect.
The strongest warning sign appears in entry-level hiring. For workers ages 22 to 25, the report finds that entry into highly exposed occupations fell by about half a percentage point, translating to a 14% drop in job-finding rates in the post-ChatGPT period relative to 2022. The authors are careful not to oversell this result, noting it is only barely statistically significant. But it is the clearest signal in the paper that disruption may already be affecting the hiring pipeline.
This finding is reinforced by outside evidence. In “Canaries in the Coal Mine?” researchers Erik Brynjolfsson, Bharat Chandar, and Ruyu Chen used ADP payroll data to show that workers ages 22 to 25 in the most AI-exposed occupations experienced a 13% relative decline in employment, driven mainly by weaker hiring rather than a spike in separations. Different data, similar signal — the early damage appears to be hitting the career ladder first. This aligns with broader trends highlighted in our analysis of 2026’s top job markets for new graduates, where early career opportunities are shifting.
For employers and policymakers, the report underscores a critical point: a stable unemployment rate can mask a shrinking on-ramp into professional work. If firms retain experienced workers and automate pieces of junior work, the damage shows up in hiring data long before layoff statistics. The organizations that respond best will be those that protect the pipeline of human judgment before it thins out beneath them.
