In my past life I spent some time in university administration, and one of my jobs at this public university was to take proposals for new degree programs that the university had approved of to the state board of higher education, for their necessary approval. In those proposals we had to include a section on forecasts for employment demand for graduates of the program, and there were a few quasi-official agencies who would produce such numbers. One of my PhD fields was labour economics, and I knew, and kept quiet about, the fact that these forecast job numbers were, respectfully, dubious. Not lies or anything meant to deliberately mislead, but numbers that were guesses meant to suggest careful application of economic science, but that actually were just extrapolation of trends. Nothing enhances a guess like a table of numbers.
The problem wasn’t that the labour market forecasts were being done by incompetent people. It’s that we just cannot know very much. The essence of future technological change is that we don’t know what it will be – if we did know then we would already have the technology. We cannot know whether some specific applied technique we teach now will be highly valued or useless ten years from now.
There’s a good op-ed in the Times today by Michael Steinberger on the uncertainty that the development of Artificial Intelligence is imposing onto labor markets, especially for recent college graduates. Degrees that ten years ago universities and state boards of higher education would have said were sure things are, in the end, not so sure at all. I don’t write much about AI (except for this piece on a hapless attempt by our university to get ahead of the game) and its future impact on work, or on arts and culture, because I would only be guessing as to what sorts of impacts it will have, or whether they will be big or easily absorbed.
So if I had any influence on higher education (I do not) I would be resisting calls to be “nimble” and “responsive to future industry needs” in degree programs, since shifting direction in educational focus at a university is slow and expensive, with unrecoverable costs should you guess incorrectly. Instead, we might avoid the highly specific, or temporarily in high demand, and let students choose from a wide variety of more general options that in their view will be interesting and stimulating and a good fit. Have faith in the knowledge that is decentralized amongst individuals, as Hayek would have put it.
Alas. Universities are in a rush to be relevant, hoping for actual degrees in AI no less, and our state government has said that degree programs that don’t pay decent wages at the moment ought to be cut (including, at my university, a bachelors degree in music in one of the very best public university schools of music in the country).
Nobody knows anything. It takes a couple of years to launch a new degree and hire faculty, then students need at least four years to get the degree, and then where are we? In the mid-2030s, with a labour market and technologies we just don’t know – can’t know – much about at all.
If a student wants to major in computer science, or epidemiology, or oboe, let them. Don’t try to steer them into fields for which you claim high potential but for which you don’t really have any idea. A student majoring in music knows there’s not much money there. A student being told that the new degree in AI is the ticket for the future is taking your word for it, and might borrow heavily on that promise, and we all need to be very careful about leading them on to what might be a dead end.
Cross-posted at https://michaelrushton.substack.com/
Leave a Reply