At least up to the present, AI has not caused large-scale unemployment. Peter McCrory, head of economic research at Anthropic, pointed out in a recent lengthy article that despite AI output growth exceeding 2,000% in 2024-2025, with approximately one-fifth of U.S. companies already implementing AI, the U.S. unemployment rate remained steady at a full-employment level of 4.2% in June 2026. His research shows that unemployment rates in jobs highly exposed to AI are not significantly higher than those in other positions, with recruitment for software engineering roles even rebounding. McCrory believes that AI is currently a 'skill-biased, labor-augmenting' technology, as every job involves elements such as interpersonal coordination and physical operations that are difficult for AI to replace, thereby amplifying the value of human expertise. However, young workers face increased difficulty in securing jobs in roles highly exposed to AI, though this phenomenon is more strongly correlated with the overall cooling of the labor market.