The Broken Bottom Rung: Entry-Level Work, Durable Capability and What Schools Should Prepare Young People For

Most public debate about AI and work asks how many jobs will disappear. For education, a more useful question is which jobs are changing first and what that does to how people learn to work. Evidence from 2025 and 2026 suggests that AI is concentrating its effects on entry-level roles in exposed occupations: employment for 22 to 25 year olds in the most AI- exposed jobs in the United States fell well below that of comparable peers, while older workers in the same jobs were largely unaffected. At the same time, firms most exposed to AI have grown faster and pay a sub‐ stantial premium for AI skills, and the jobs AI reshapes increasingly demand distinctly human capabilities. This article draws on a practitioner newsletter series written for school leaders in 2026, read alongside labour market research and the task-based economics of technological change. It argues that the central risk for education is an expertise paradox: AI raises the value of deep domain judgement while automating the routine entry-level work through which that judgement has tradition‐ ally been built. It examines what employers report about skills, the weakening signal of credentials, the changing status of caring and voca‐ tional work, and how systems such as China's are restructuring higher education. It proposes a three-layer model of preparation, combining durable human capabilities, AI fluency centred on supervising and check‐ ing AI output, and domain expertise built through authentic work with employers, and sets out implications for schools.