How Will AI Impact Higher Ed?

Virtually every observer of American higher education agrees that it is in trouble, and most think the short to midterm future for universities is pretty bleak. Most emphasize growing disenchantment with the academy on the part of governmental funders, most conspicuously the Trump-era federal government. Still others point to both the enrollment decline of the past 15 years along with the shrinking supply of college-age Americans in coming years because of declining fertility rates.

Another factor arising that could be both a threat and an opportunity for colleges is artificial intelligence (AI). Will it magnify higher education’s troubles or help foster a period of expansion and prosperity?

Warning: I am an economist, and my profession’s record at forecasting future events is pretty dismal. An early economist, T.R. Malthus, in 1798 predicted a coming era of extreme poverty and near starvation, whereupon Britain then had the longest sustained period of rapid economic growth the world had ever seen. In the early 1940s, many prominent Keynesian economists predicted a resumption of the Great Depression at World War II’s end: It never came.

So let me hedge my bets by offering both a pessimistic and an optimist take on the possible impact of AI on higher education’s future.

The Pessimistic Perspective

During the first great Industrial Revolution, beginning in Great Britain around 1750, new technological advances increased incomes and jobs for most people, but there were some losers, too. Most famously, in the cotton textile industry the invention of machine-based technology to spin and weave cloth (i.e., the spinning jenny and power loom) meant some home-based spinners and weavers lost their jobs to new factory-based workers operating much more productive machines. The losers were often illiterate or marginally educated, while the more educated classes generally participated in the growing income arising out of increased production.

Similarly, in the U.S. over the last half of the last century, employment and job opportunities in Rust Belt manufacturing communities declined for factory workers with modest amounts (at most, high-school diplomas) of formal education, while other parts of the country, embracing tech-heavy new industries using lots of brainy, highly educated people, seemingly prospered. Compare Silicon Valley’s robust job growth with employment in the auto and steel industries in states such as Michigan, Ohio, and Pennsylvania, which started losing out to new foreign competition. For example, Ohio’s personal income per capita in 1940 was about 10 percent above the national average, whereas today it is over 10 percent below it, and many Rust Belt cities (think Detroit or Cleveland) have dramatically smaller populations today than they did in 1950.

Technological change until recently almost always meant that machines substituted for manual-labor tasks previously inefficiently performed by humans. But now AI has rapidly grown in sophistication and importance, so machines (computers) using AI are replacing college-trained workers. “The Bull Market for Economists Is Over. It’s an Ominous Sign for the Economy” read a New York Times headline recently. The author added, “Earning a Ph.D. in economics has long been a reliable path to affluence and prestige. Not anymore.”

Right. Why hire a bright economist for over $100,000 a year to predict, say, interest rates a year from now based on some mathematical modeling, when a much-lower-paid worker with a high-school diploma can retrieve similar results from a computer using AI?

Previously, machines replaced human physical labor; now, perhaps, they are replacing human brainpower enhanced by high levels of education—and not just in economics. For the first time in history, perhaps, new technology’s benefits are biased against the brainy and book-educated—the people who go to college.

Since college is primarily viewed as an investment in “human capital” improving vocational outcomes, AI could be devastating, since human brainpower is being replaced by sophisticated computer-based machinery. There’s the “dismal science” of economics at work!

An Optimistic Interpretation

However, the July 26 issue of the Economist offered precisely the opposite interpretation with its article “The Economics of Superintelligence.” (Gated for subscribers here.)

Productivity will explode as a consequence of this most dramatic technological advance in hundreds of years. While the owners and producers of AI technology will derive huge financial benefits, other Americans will gain, as well, from growing incomes. Moreover, AI probably is not much help in dealing with an overflowing toilet and cannot do welding or roofing. AI doesn’t do beautiful painting or, at this stage at least, write beautiful symphonic music. With rising incomes, there will be enhanced demand for plumbers, fine painting, and roofers. Indeed, AI might favor those doing some forms of manual labor or even studying the arts and humanities.

Indeed, demand for colleges might rise as more affluent Americans embrace exploring the sources of truth and beauty via a liberal-arts education. Increased funding arising from prosperity can also fund more advanced high-level university research in scientifically oriented areas. Higher income-tax revenues can enhance subsidizing universities.

Additionally, as incomes rise from technological advances, more Americans will go to college for reasons that extend beyond traditional learning: Upscale residential colleges might flourish as increasingly affluent Americans use universities for an extended gap period of relaxation between the perils of adolescence and the trials of adult life. Colleges will more than ever become like country clubs, with some culture and learning thrown in.

As worldwide incomes rise, the demand for high-quality but also fun higher education in America will soar, leading to robust growth in foreign enrollments in American colleges.

Some Alternative Thoughts

Maybe we are overpromoting AI a bit, and both its impact on society as a whole and on America’s colleges and universities will be notable but not revolutionary. After all, for two or three generations, computers have substituted machines for workers performing mathematical calculations—roughly what AI purports to do. And while computers have changed our lives importantly, it is not clear what they specifically did for university growth—probably enhanced it a bit but not in a revolutionary fashion.

Alex Green, a teacher at Harvard’s Kennedy School, has argued persuasively in the Wall Street Journal (“AI Robs My Students of the Ability to Think”) that the ubiquitous presence of AI in classroom settings has meant that students have substituted the use of a machine (AI) for thinking and brain-centered evaluation of alternative ideas. Human learning therefore is being retarded. My sense is that Green is right and that students are not exerting their mental capacities to evaluate alternative approaches to ideas and problems. Thus, not only is AI substituting for human intellectual resources, it is reducing the size of those assets—turning humans into clueless prisoners of a technology they have embraced but do not understand.

If Green is right, why send kids to college if they are not exercising their human capital—brain power—to evaluate alternative solutions to real-world problems? Are humans becoming servants to machines? Why go to college if not to learn how to think?

But Angus Fletcher makes yet another appealing argument in his new book Primal Intelligence. Humans possess something called “intuition” that often guides us to doing innovative things not easily modeled by machines that think. Human ingenuity has given us an extraordinary tool, AI. Whether it is strengthening the role of universities in our lives or hastening their weakening is currently an unsettled matter.