Computer Science and AI
This is a very good article about the use of AI in computer science courses. The article ends with a discussion on “what sort of future problem solvers and software developers we want to graduate from our programs,” which I think is the absolute right question to ask now.
I’ve said for years that when the measure becomes the goal it ceases to be a good measure (Goodhart’s Law), which has been true in education for decades, certainly since the late-1990s national push for standardized tests. I feel that if the goal of a CS program (or any academic program) is to have students produce a specific set of work, it has failed Goodhart’s Law. The article mentioned headhunters looking at Github pages of applicants to see what projects they work on outside of class, and the response is spot on:
...without any recognition of the fact that interesting innovative solutions to hard problems come from people who have experiences outside of class that don’t involve computing, who do something besides code all the time; like be on the debate team, sing in a choir, get involved in theater, take a dance class, study history, read poetry.
Teaching civics students to pass a test instead of engaging with the government is failing them. Rather, we should focus on having students develop a specific set of skills to be successful in their fields.
There are still people in CS fields who started with punch cards. Mechanics working on electric vehicles now almost certainly learned their trade by working on internal combustion engines. There is no future-proof curriculum that doesn’t teach critical thinking and adaptability because fields are fluid.
It’s why I also get so frustrated at the anti-education social media posts like the one above. The tax code changes constantly. Teaching students to do taxes can be important (I do it every semester), but so is learning critical thinking and how to unpack government jargon. Taxes will change. Do you want students who can adapt to those changes because they understand the process, or do you want ones who can follow a set of rules that will become outdated after a handful of years? Do you want students who can only submit their taxes or ones who can advocate for fair tax policies?
The best educators are the ones who can answer, “What do I want my students to be?” instead of, “What do I want my students to do?” We are educating whole people, future members of society, parents, friends, problem solvers, not computer scientists or historians. There can, and should, be a vast difference between what a person does and who they are.
I’m not sure when the AI bubble will burst, but I’m confident it will. We can have students who are adaptable to that reality (and the recession it will certainly bring) or not. The programs that address that head on are the ones I want my students, and future children, attending.