# The degree splits into two wage-premium tracks by the early 2030s: negative in AI-absorbed majors, positive in licensure-anchored ones — so national headcount masks a curriculum-level collapse

**Core claim:** AI does not shrink the degree market uniformly; it splits it by major. In AI-absorbed generalist fields — general business, marketing, communications, entry-level analytics — the wage premium for degree-holders turns negative by roughly 2031 (Unverified: date is extrapolated from current premium compression, no direct source), because AI demonstrably performs the routine cognitive work those degrees certified, and employers can verify competence directly through AI hiring screens rather than trusting the credential. In licensure-anchored fields — nursing, engineering, accounting/CPA, teaching — the premium holds or rises, because state boards and liability law still require credentialed humans. The result is a scissors shape, not a slope.

**The masking consequence:** national enrollment figures will look deceptively stable through this transition (Unverified: mechanism, not measured), because students don't leave the system — they migrate between majors and institutions. Total headcount holds while roughly half the curriculum at a typical tuition-dependent school loses pricing power and enrollment. This is why institution-level data will mislead: a school can show flat enrollment while its general-education and generalist-business programs bleed out beneath it, and administrators will misread program-level attrition as price sensitivity and cut aid in exactly the wrong places.

What would change this: if the entry-level wage premium in generalist majors stabilizes or rebounds through the late 2020s rather than continuing to compress, or if AI hiring screens fail to achieve employer-side adoption fast enough to substitute for the degree signal in those fields.