Rebuilding the Middle Class

The rebuild-the-middle-class thesis is Autor's prescriptive response to decades of empirical documentation of middle-skill decline. The argument, developed in his 2024 NBER paper and subsequent public writing, is that AI differs from previous automation in a crucial respect: prior technologies automated routine middle-skill tasks, pushing workers toward either the high-skill or low-skill ends of the distribution. AI, by contrast, can enable workers with moderate training to perform tasks that previously required expert credentials — the nurse who uses AI to diagnose cases that previously required a physician, the paralegal who uses AI to perform legal analysis that previously required an associate, the community college graduate who uses AI to do architectural work that previously required a licensed architect. If institutions channel AI toward this application, it could reverse four decades of polarization. If they do not, AI will simply extend polarization upward.