I’ve recently read two articles about the subject of how AI is quickly destroying higher education and what can be done about that; one on The New Yorker (paywall-free version available) and another on The Chronicle of Higher Education (again, paywall-free version available), and they are quite similar in spirit: both quote similar sources, both deplore the current state of things, and both fail to provide any actionable idea at the end.
I have written an article on the subject of LLMs in education on my own magazine last December, called “Banning, Adopting, Reckoning”, and this article will largely expand those ideas. You can tell this is a subject that I’m very interested in; I have worked on-and-off as a private trainer for software developers for more than 2 decades, and I worry, but not about the students; I’m worried about the teachers.
My biggest beef with the articles on The New Yorker and The Chronicle is that all of those teachers in Ivy League schools keep on giving essay homework assignments to their students, and then keep on complaining that those same students keep on turning in ChatGPT-generated shit.
Well, duh.
My question is: why do you keep doing that? It’s a very simple question; a litmus test. Instead of grading them again and again through yet another paper request that you know will be generated with an LLM, aka a bullshit machine, why not try something else instead?
And here’s a crazy idea for those US-based professors, one that I did not come up with myself, but one I actually witnessed again and again in both Argentina and Switzerland when I was a student. In those places, in primary, in high school, and in university, we had to sit for hours in a closed room, with only a small set of allowed helper tools (a pen, a calculator, a dictionary, and not much else) plus maybe an apple and a bottle of water, and then (hear me out) we had to write our essays from scratch, right then, right there.
I know, crazy, right? We actually had to think by ourselves, without help. I wonder how could we survive such treatment. You even had just one bathroom break allowed of maximum 5 minutes per exam session, and there were teachers inspecting those same bathrooms for cheating devices (which consisted of, you know, small bits of papers and wall inscriptions; after all these were the late 1980s and early 1990s).
The topic would be randomly assigned (and discovered by students) right before the exam begins, and then we would have 4 hours to come up with some meaningful prose that explained our knowledge (or lack thereof) about the subject.
Here’s another crazy idea, one that is actually mentioned in one of the articles, albeit with a baffling and complete lack of enthusiasm: oral exams. My Swiss Matura exams in 1993 consisted of 10 exam sessions, 5 written and 5 oral, in various subjects; in my case those were French, English, Italian, Physics, and Mathematics. Of course, the subjects depended on your “orientation”, and in my case that was science. The written exams had the structure and organization I described above; the oral exams, again, would be taken individually, person by person, in front of two teachers, for 20 minutes. Students had 20 minutes of preparation before the session in a separate, closed room, without access to notes, computers, or anything else; just a pen and an empty pad of paper.
Good old just you and your brain. What about that.
Were these exams difficult? Why yes of course, they were. Did they take long to complete? Yes, almost a whole month, at a pace of 2 or 3 exams per week. For teachers, too, it’s a tiring and long procedure. But given the lack of access to help during them, you are required to learn your stuff, at the risk of failing your exams completely.
But wait: were such exam sessions a long and expensive procedure? Oh yes, and now you understand why US universities are sinking into this pit of despair. They have sacrificed their students in the name of economic efficiency, while at the same time drowning those same students in an endless cycle of student debt. And now those same students will go into an AI-enabled workplace where they will be (rightfully so) considered redundant, replaceable, and a simple commodity. Well done, America.
What was the advantage of having actual exams, you know, those of the kind where you had to think through them? Well, because, thanks to said complexity and length, teachers could actually gauge our knowledge, and grade us accordingly, without worrying of a bullshit machine getting in the way.
Now, what to do during the non-exam part of your education? Well, guess what, it’s very simple: keep computers and smartphones out of your classrooms, gather your students around you, and engage in a dialogue with them. Go back to the old Greek style of the Platonic Academy, where rhetorics become a part of the curriculum, where students are taught how to have healthy debates with one another without recurring to ad hominem attacks, and where ideas are discussed with facts and research. Science classes belong to labs; mathematics belong to whiteboards or blackboards; literature classes belong to libraries; and so on. You get the idea.
My point is: LLMs are not a threat, but an opportunity to change the way you teach. But, of course, if your business model consists of making easy cash, I have bad news for you; you’ll have to change. Ouch, sorry about that.
I concluded my article with the following words:
If anything, this author firmly believes that programming skills are second to those related to communication; most engineers coming out of colleges these days are unable to express themselves in public, to teach their peers, to write an essay or a blog post, to communicate their ideas to stakeholders, or to put together a simple documentation bundle without suffering a seizure in the process.
If we are going to have LLMs performing the grunt work of coding, we need students to become the architects of tomorrow, not just another coder selling their work at Fiverr for a living. Help those students build larger and more complex systems, while at the same time making them conscious of the process, and helping them develop those much required “soft skills”.
Will students use ChatGPT to prepare their classes anyway? You bet they will! The same way they have been using sand, chalk, parchments scrolls, books, logarithm tables, libraries, the Encyclopædia Britannica, sliding rules, thesaurus, calculators, Google, and Wikipedia to solve their assignments for the past 5000 years. But that’s not an issue, as long as the teachers conducting the classes adapt their teaching style to this brave new world of bullshit machines available for free.
The problem is not that LLM exist or whether they should be banned; is that US universities don’t want to evolve their teaching styles, and are very happy to keep charging those same exorbitant tuition fees for an ever-decreasing level of quality.
My message is, then, to prospective students: avoid those Ivy League schools, they’re clearly worthless. And to those businesses looking forward to hire those upcoming Ivy League graduates: brace yourselves for the limited set of capabilities you’ll see in action; or, as Duane Gran said last Monday,
Your future doctor is using ChatGPT to pass medical school, so you better start riding a bike and eating healthy now.