Why hasn’t technology disrupted higher education already?

Why hasn’t technology disrupted higher education already?

Matthew Yglesias — Read time: 10 minutes

Why hasn’t technology disrupted higher education already?

The key question about AI and school is about the past

There’s a lot of interest in the question of how large language models may impact education. I don’t have the answer, but I think there’s a more fundamental question worth considering: why have the past 50 years of technological change had so little impact on schooling?

The English word “lecture” derives from Medieval Latin’s “lectura” and is cognate with words like “lecteur” (French) and “lector” (Spanish) which mean “reader.” A lecturer, in other words, is a reader.

Today, giving a lecture that consisted of simply standing at a podium reading a book would be considered bad practice. But several hundred years ago, books were extremely expensive because hand-copying manuscripts doesn’t scale. What does scale, at least to an extent, is the human voice. So an institution could serve the very useful function of providing a place where students could gather to hear a person read out loud from a book and write down what the lecturer was saying, securing knowledge.

Give a gift subscription

An institution like that would need to have a lot of books on hand and a scholar would need ready access to books, so producing scholarship was highly complementary to lecturing. The scholars take books as inputs but also produce books as outputs. To earn money, they would lecture — which is to say, read the books — to students. The students themselves would benefit not only from learning about what the books say but also from some kind of formal certification. And thus was born the familiar university bundle that combines libraries, scholarship, teaching, and certification.

This is a somewhat rickety pile of in-principle-separate ideas that really does seem vulnerable to technological disruption. On its face, the relevant disruptive technology should have been the printing press, and the disruption should have happened three or four hundred years ago. But not only did the basic structure of the university persist, but most of the world’s leading universities are also much newer than the printing press. Harvard and Yale are really old by the standards of American institutions, but they’re not older than printing — and many other prestigious American universities date from the second half of the nineteenth century. By the time Stanford and the University of Texas were founded, it was already extremely clear that people could study books at home (or in libraries) and then take exams administered by certifying bodies that had nothing to do with teaching or research.

None of this is to say that large language models won’t disrupt traditional education, just to observe that education has been more resilient than one might think.

Learning has gotten much cheaper

While the price of education has infamously gone up a lot over the past 50 years, the price of learning has clearly gone down.

At one point back in the pre-vaccine days of Covid-19, our dishwasher stopped working. I am the opposite of a handy person, and I had no idea how to deal with this myself. But I Googled. And sure enough, there are multiple YouTubers operating in the home appliance repair space with sufficient specificity and dummy-friendliness that I was able to fix it.

More recently, I got a little idly curious about why the bokeh (those out-of-focus points of light in movies) sometimes consists of circles and other times ovals. It turns out to have to do with whether the camera lens is spherical or anamorphic. That led me down a YouTube rabbit hole on the history of anamorphic lenses, which were originally developed in order to make it possible to display films with a wide aspect ratio. The anamorphic lens essentially squeezes a wide image onto a square piece of film, and then an anamorphic projector unsqueezes it. These lenses have a variety of other optical characteristics — including oval bokeh, prominent lens flares, focus breathing, and other more subtle effects. In the modern day, it is perfectly possible to make a widescreen film with a spherical lens. But in part because of the historical legacy of anamorphic shooting of movies (but not still photography or television shows), anamorphic visual artifacts lend films a “cimematic” look. So many cinematographers continue to prefer anamorphic lenses even though they’re not necessary, and some filmmakers who aren’t using anamorphic lenses insert anamorphic-style lens flare to achieve a more cinematic effect, even in animation.

The point is, I learned a lot about lenses. And I learned it all for free, from YouTubers who were making videos to sell ads or to do affiliate marketing for lenses and other camera equipment. And when Matt Zeitlin tweeted about split diopters, I already knew where to go for more information about camera lenses.

Now I’m not going to say that watching these YouTubes is an adequate substitute for going to film school or apprenticing with a working cinematographer. But the fact is that today it is possible to learn a lot of things quickly and easily relative to 1992.

By the same token, my DIY dishwasher repairs still left the machine in less-than-perfect working order, and after about a year and a half of limping along, it was totally busted. I think it’s conceivable that someone with better training than me could have fixed it better. But the fact remains that a miraculous quantity of information is at our fingertips. If you want to learn how to tie a bowline knot (useful for dragging a dinghy behind your sailboat without the knot becoming so tight it can’t be untied) or what happened with the Great Vowel Shift in early modern English, the internet has your back. Learning is easier than ever!

And yet, in practice, learning is still hard.

In-person instruction has unique value

I lift weights regularly with a personal trainer. A person obviously could lift heavy things without a personal trainer and still get stronger, though of course if you just pick up some heavy weights with absolutely no knowledge, the odds are overwhelming that you’ll waste your time or injure yourself. But the basic information has long been available in books and magazines and on home video since before the existence of the internet. Today, we have weightlifting YouTube, we have Apple Fitness+ — we have tons of highly efficient alternatives to paying someone to watch you lift heavy things.

But the trainer is helpful not just for learning but for motivation. I consistently lift more for longer with a trainer. The trainer pushes me to do more, she provides a safety function, and more importantly, the fact that the trainer is there denies me the excuse of some safety concern as a reason to stop.

I think precisely because these are the ways the trainer adds value, the strongest, fittest people at the gym generally aren’t relying on trainers, at least not all the time. They just have a gym rat temperament, a lot of detailed knowledge, and a lot of strong muscles. Whether innately or through cultivation, they have the kind of mindset where they push themselves without external stimulus.

Precisely for that reason, they’re stronger than me and in better shape than I am. And a sort of naive read on the situation would be to say that if we look at the people who are really strong, they show that inefficient, non-scalable personal training has been disrupted by whatever hobbyist forums or other internet resources they rely on.

In the real world, though, that doesn’t work at all. I got sincerely interested in this lens stuff so I learned a lot about it, but I am not sincerely interested in exercising — I dislike it and lack inherent motivation. Instrumentally, though, I want to be a healthy person, and the in-person instructor is a tremendous motivating tool.

Tech improvements don’t solve the problem

A decade ago there was tremendous hype around the potential for Massive Online Open Courses (MOOCs) to replace traditional classroom instruction. Then it turned out that online for-profit colleges were mostly good for running scams on marginal students. The problem with MOOCs for the typical student is the same as with me trying to lift weights on my own: for people who have a second-order desire to get a degree despite a lack of temperamental suitability for school, the in-person instructor is invaluable. We learned that lesson all over again during the pandemic when a lot of districts went remote with bad effects. Motivation and self-discipline are valuable commodities, and an in-person instructor can help provide them.

I do think it’s fair to say that internet video is a step forward from VHS or simple text files on the web, and all of that is a step forward from print as the only medium for conveying information. And the printing press itself was, of course, a huge step forward.

It’s very easy to imagine chatbots improving on Google search as a way to look stuff up, and AI-powered individual coaching could be even more powerful than videos as a way to learn things.

But I do think the history of ed tech has been one where better and better information technology makes it much easier to learn things without really making much progress on the big problems of education, because the motivation/discipline piece of education is so central. In practice, I think the net impact of IT improvements on education has probably been negative. Today’s smartphone is a much more powerful and convenient learning tool than the public library of 30 years ago. But the 1992-vintage public library really did work very well. And today’s smartphone is also a much more powerful tool of distraction than anything that was available in 1992. Educators’ jobs have probably gotten harder rather than easier, not despite but because of the improvements in information technology.

The revolution, if one comes, is likely to be in the value of the learning itself rather than in how it’s done.

De-skilling is very real

My whole life has been an era of relatively slow productivity growth and misguided fears about technologically induced unemployment.

I think people today are generally aware of those macro facts, which in turn leads them to ignore some important micro facts. Something that I learned as a kid, both at summer camp and because my mom was a skilled photographer, was how to develop black and white film and make black and white photo prints. By the time I got to college, this skill set was clearly on the way out, but analog cameras still took much better photos than digital ones. I worked on a campus alt-weekly paper that had its own darkroom, and even though I wasn’t in the photo department, I was interested in photography and would sometimes pitch in and help out with darkroom stuff. I learned how to burn and dodge prints, about different chemicals, and how to use the negative scanner.

Once upon a time, these skills would’ve had real market value.

Now to be clear, it’s not like I was ever a superstar photo developer or that this was an incredibly lucrative line of work. But it was a real job skill and I did in fact learn to do it just fine. I would have been qualified, 22 years ago, for a non-minimum wage job that involved specialized skills and had nothing to do with writing. Today, that’s no longer the case.

Back in 1996, Paul Krugman wrote a column pretending to look back from 2096 and one of his predictions was that advances in information technology would lead to a huge decline in the monetary value of education.

Eventually, the eroding payoff of higher education created a crisis in education itself. Why should a student put herself through four years of college and several years of postgraduate work to acquire academic credentials with little monetary value? These days, jobs that require only 6 or 12 months of vocational training — paranursing, carpentry, household maintenance and so on — pay nearly as much as if not more than a job that requires a master's degree, and pay more than one requiring a Ph.D.

He envisioned this as leading to a huge decline in college enrollment, and thus a huge decline in the demand for economics professors like himself. Krugman-in-the-fiction was employed as a veterinarian and wrote essays about economics as a hobby. Even with veterinary medicine, though, the contemporary practice basically mixes two different things. One is that a vet has a knowledge base that can let her diagnose ailments and prescribe courses of treatment. The other is that the vet can just deal with the animal as a physical object that exists in the world and requires a certain delicacy of handling. Right now, we seem closer to a world in which computers can substitute more for the knowledge function than for the physical one. This would mean vets continue to be in high demand, but the practice of veterinary medicine becomes de-skilled — not something that can be done without training, but something that can require less training than today’s vets have.

And that would be the real revolution in education. The past couple of centuries have seen a steady increase in the market demand for certain kinds of skills. That’s meant people want to acquire these skills even if they don’t necessarily have tons of inherent motivation — they want training to earn a better living, and education as a formal process is a very useful motivational crutch. It’s at least possible that the next twist in IT will be to rapidly erode the demand for that kind of education, meaning people will be left primarily to learn things they are curious about where there is much less need for external motivation.

Update your profile

Only paid subscribers can comment on this post

Check your email

For your security, we need to re-authenticate you.

Click the link we sent to tka.lee@gmail.com, or click here to sign in.