Baumol’s cost-disease captures the economic certainty of productivity costs that innovation cannot modify, reduce, or outright eliminate will rise over time. Human labor proves the most incurable. As it is always necessary, it is eternally inflationary. Its indispensability makes it resistant to algorithms and silicon. Human labor, once the hallmark of economic progress and individual prosperity, has turned into a toxic expenditure for industries requiring more of it.
American higher education currently operates as an unsustainable expanse human labor stationed in 2- to 4-year academies, many of which exhibit congenital defects including inequitable enrollment practices, insufficient advisers and counselors, inadequate financial aid support, or curricular irrelevancy. All have garnered more national exposure, yet their endemic cost-disease and the cumulative effects warrant greater coverage.
Most attention over the past decade has gravitated toward affordability issues for students, and understandably so, but rarely considered are the productivity costs incurred by the institutions being blamed for those issues. This neither exonerates nor diminishes institutional liability. Nor does this perspective at all dilute the take that cost-disease has been overestimated when explaining excessive tuition rates. Its impact on the two-year college, however, does prove concerning.
Two-year colleges often accommodate deficits of their underserved populations beyond the academic: some sites have opted to provide daycare services to accommodate parents, other locations make food pantries available for malnourished students, while other campuses offer onsite health care services for uninsured students, including a rise in mental health services. And these populations—the parents, the low-income, the first-generation, the racially diverse, the non-English speaking, the functionally illiterate, the undecided, the uninterested, the all the above—will comprise a larger share of declining enrollments and demand more assistance programs, greater comprehensive counseling, and more expansive academic supports, all human-driven. Students who have had sparse access to learning technologies and few opportunities to practice self-sufficiency, pushing many to become more fully dependent on human-based services. This shift will subsequently require from educators a more expansive, perhaps more explicit, hidden curriculum—in loco parentis—for students whose upbringing had survived without one, but whose immediate future demands it.
These trends and roles make teaching and learning—that which we call education—a supremely human act, a dynamic that artificial intelligence can neither replicate nor replace. Although software can cultivate students’ online navigation and recall abilities, a digital environment cannot outclass what flesh and blood has long refined. Genuine interaction will remain human: nuanced, contradictory, and temperamental—qualities that software can at best mimic.
But this human triumph pays an unavoidable cost: provided appropriately to scale as it ought to be, higher education’s operating costs would prove prohibitive, requiring an impossibly expansive labor-force that would never be able to meet demand. Capacity drops. Quality declines. Downtrends continue. Rinse. Repeat.
Even if artificial intelligence prevails in higher education within another generation, a sustainable market for it might not. State funding will never rebound; unsustainable health care and pension costs will make sure of that. Referendums will eventually run dry of support from a disillusioned public, whose purchasing loyalties have begun prioritizing relevancy over an obscure allegiance to a brand name. Productivity costs will continue their rise, thus limiting the scale of commodification required to meet demand. Online learning serves as a reliable proxy, as its revolutionary role after 30 years has settled more naturally into a learning platform, storage unit, and conduit for faster, more novel communication. Online education’s expedition not only to improve human learning but also to replace human educators likely capsized during the high-seas reign of fraudulent for-profits.
Artificial intelligence has not failed on account of too few attempts, a lack of ingenuity, or diminished capability. Online delivery has become a curricular staple, required of many faculty programs, but it was no more a threat yesterday as current AI software will be tomorrow. The tried and true Socratic method persists, as it always should.
The most pertinent question is at what continuing costs?