The Skills Gap AI Is Making Worse, Not Better
The optimistic case for AI and labor markets holds that AI handles routine tasks, freeing workers for higher-value activities that require human judgment and creativity. It's a compelling theory. The evidence so far suggests the transition is considerably messier, and that AI is actively widening certain skills gaps rather than closing them.
The automation of learning tasks
Many of the tasks AI performs most efficiently are the same tasks that teach entry-level workers the fundamentals of their profession. Writing first-draft memos teaches junior lawyers how to structure legal arguments. Building initial financial models teaches analysts how markets and businesses work. Debugging basic code teaches new developers how programs fail.
When AI does those tasks instead, the junior worker doesn't just lose the task. They lose the learning that came with it. The gap between entry-level capability and the senior judgment that AI can't replace grows wider. Organizations that aggressively deploy AI for entry-level work may find, five years out, that they have a generation of nominally senior employees who never built the foundations those seniority titles imply.
The misalignment in education
Educational systems take 10-15 years to meaningfully change curriculum. The skills that were in demand when today's college students enrolled (data analysis, coding, digital marketing) are being partially automated by the time they graduate. The skills now in highest demand (AI governance, prompt engineering, systems integration, human-AI collaboration) weren't formalized enough to teach when current curricula were designed.
This creates a structural lag. Employers report skills shortages even in a period of elevated unemployment. Universities report full enrollment even as employers question the relevance of what graduates learned. Neither is lying. They're operating on different timelines.
What would actually help
The most effective response isn't a new wave of coding bootcamps or AI certification programs, though those have marginal value. It's a fundamental reconsideration of what skills are durable. Critical thinking, communication, the ability to work across disciplinary boundaries, and comfort with ambiguity are skills that have resisted automation through multiple technological waves.
The employers and educational institutions making that bet now are positioning their graduates well. The ones still teaching to the last generation of in-demand tasks are likely to find the gap between their graduates and what employers need getting wider, not narrower, over the next decade.