Implications of AI to schools

Again, AI is not getting rid of degrees or learning or education.

It's getting rid of cheap methods.

Scantrons and bluebooks were always a way that made it cheap for institutions to produce results. Now those methods kinda seem silly, right?

500 person freshman lectures seem kinda absurd now, right?

Teaching via adjuncts that had 3 days notice for the class and are paid nothing is kinda scammy, right?

R1s professors whose tenure evals have nothing to do with teaching is kinda wrong, right?

The Oxbridge model of 5-10 person classes with a proctor is what the education with AI is going to be about. It's small, intimate, and expensive.

>Also, just like how calculators are allowed in the exam halls, why not allow AI usage in exams?

Dig deeper into this. When are calculators allowed, and when are they not? If it is kids learning to do basic operations, do we really allow them to use calculators? I doubt it, and I suspect that places that do end up with students who struggle with more advanced math because they off loaded the thinking already.

On the other hand, giving a calculus student a 4 function calculator is pretty standard, because the type of math they can do isn't what is being tested, and having a student be able to plug 12 into x^3 - 4x^2 + 12 very quickly instead of having to work it out doesn't impact their learning. On the other hand, more advanced calculator are often not allowed when they trivialize the content.

LLMs are much more powerful than a calculator, so finding where in education it doesn't trivialize the learning process is pretty difficult. Maybe at grad level or research, but anything grade school it is as bad as letting a kid learning their times tables use a calculator.

Now, if we could create custom LLMs that are targeted at certain learning levels? That would be pretty nice. A lot more work. Imagine a Chemistry LLM that can answer questions, but know the homework well enough to avoid solving problems for students. Instead, it can tell them what chapter of their textbook to go read, or it can help them when they are having a deep dive beyond the level of material and give them answers to the sorts of problems they aren't expected to solve. The difficulty is that current LLMs aren't this selective and are instead too helpful, immediately answering all problems (even the ones they can't).

I have never had to rote learn anything mental since at least mid childhood. I can't remember before then.

If it's something I need to do regularly, I eventually learn it through natural repetition while working towards the high level goal I was actually trying to achieve. Derivatives were like this for me. I still don't fully know the periodic table though, because it doesn't really come up in my life; if it's not something I need to do regularly, I just don't learn it.

My guess is this doesn't work for everything (or for everyone), and it probably depends on the learning curve you experience. If there are cliff edges in the curve that are not aligned with useful or enjoyable output, dedicated practice of some sort is probably needed to overcome them, which may take the form of rote learning, or, maybe better, spaced repetition or quizzing or similar. However at least for me, I've not encountered anything like that.

If I was to speculate why rote learning doesn't work well for me, I don't seem to experience a feeling of reward during it, and it seems like my ability to learn is really heavily tied somehow to that feeling. I learn far more quickly if it's a problem I've been struggling with for a while that I solve, or it's a problem I really wanted to solve, as the reward feeling is much higher.

It’s mind boggling that image generators can solve physics and chem problems like this—but I will note that there are a few slight mistakes in both. (An extra i term in LHS, a few of the chemical names look wrong, etc.) Unbelievable that we’re here, but it still remains an essential to check the work.

As someone who had a college English assignment due literally just yesterday, I think that "the vast majority" is an overstatement. There are absolutely students in my class who cheat with AI (one of them confessed to it and got a metaphorical slap on the wrist with a 15 point deduction and the opportunity to redo the assignments, which doesn't seem fair but whatever), but the majority of my classmates were actively discussing and working on their essays in class.

Whatever solution we implement in response to AI, it must avoid hurting the students who genuinely want to learn and do honest work. Treating AI detection tools as infallible oracles is a terrible idea because of the staggering number of false reports. The solution many people have proposed in this thread, short one-on-one sessions with the instructor, seems like a great way to check if students can engage with and defend the work they turned in.