AI creates new jobs, but the entry path is narrowing
An analysis piece argues AI is creating new roles such as agent developer and embodied-AI robot application technician, but the entry path for ordinary workers is not widening at the same pace. It cites a September McKinsey Global Institute report estimating about 11 million US workers may need to change occupations by 2035, with only roughly 14% having direct transition paths needing little reskilling and no significant pay cut. It also points to a Stanford Digital Economy Lab study updated in August that found 22–25 year-olds in AI-exposed occupations grew about 19% below peers in less-exposed jobs, mainly through reduced hiring, and to a Technical University of Munich experiment with 275 programming students where AI users scored better on exercises without matching knowledge gains. On Oct. 8, Reuters reported AT&T moved document review and legal research from outside junior lawyers to an internal 38-person LegalEdge team working with AI, cutting outside billable hours; Ropes & Gray lets over 1,000 lawyers count up to 100 hours of AI training toward billable requirements.
It highlights that the bigger labor-market risk may be less about total job counts than about whether newcomers can still gain the experience that turns them into experts.