Ask HN: Have students started questioning their curriculums due to ChatGPT?

17 points by amichail ↗ HN
In particular, what is the point of learning something that is answerable by an AI such as ChatGPT?

How do teachers reply to such questions?

Are teachers starting to change what they teach?

68 comments

[ 3.0 ms ] story [ 134 ms ] thread
I think these are great questions. Certainly one answer could be “don’t do anything, don’t change anything, consider using LMMs on par with cheating and plagiarism.” And in the short term, that’s seems like an OK answer. This is very new technology and we are in the early stages of developing it, let alone putting it to some kind of industry standard use.

In the long term, assuming AI takes over a significant share of knowledge work, curriculums will need to adapt, just as they adapted to computers and calculators.

> In particular, what is the point of learning something that is answerable by an AI such as ChatGPT?

ChatGPT will often make up an answer which is absolutely incorrect. It will even insist that it’s correct.

It appears that the only responsible way to use ChatGPT is for questions where you already know the answer or can quickly validate the correctness somehow.

>ChatGPT will often make up an answer which is absolutely incorrect. It will even insist that it’s correct.

You're assuming this will be the case forever... check back in a year :)

>ChatGPT will often make up an answer which is absolutely incorrect. It will even insist that it’s correct.

...just like [some] humans

I see this parroted but no, it's not like humans at all in this regard. While some humans may speak confidently incorrectly they internally know whether they know something or not. Large language models don't inherently provide an accurate confidence.
That’s not strictly true, an AI for sure can provide a confidence interval for its answer, but the sentiment I feel is generally correct.
ChatGPT can’t provide any score for how correct it’s answer is.

Basically everything it says is a lie, some of those lies happen to randomly be true.

It absolutely can, just check out the OpenAI Azure documentation to understand more: https://learn.microsoft.com/en-us/azure/cognitive-services/o...

Edit: This also might be helpful: https://learn.microsoft.com/en-us/azure/cognitive-services/q...

I don’t know what you think that shows, but none of it corresponds to “truth.”
I wasn't trying to suggest it did, just that the concept of "confidence" is a key one in ML.
Gotcha. Rereading what I said, “correct” rather than “factual” was poor word choice on my part. From it’s perspective correct is just whatever results in a high score.
Yeah, it's well trained by humans who do know what is "factual" (or at least better than ChatGPT does), and the "correctness" score gets used to figure out why ChatGPT gave a "wrong" answer. Over time, the model is trained this way to better align "correct" with "factual", but yeah ChatGPT needs humans to tell it what's "factually" accurate and what's not.
> well trained by humans

Yes and no, the goal wasn’t focused on just being a source of truth. They wanted it to be able to write fairy tales, poetry, and song lyrics so creative answers where rewarded even if they weren’t factually accurate.

There was an attempt to reduce harmful and deceitful responses, but the team acknowledge the tendency for “plausible-sounding but incorrect or nonsensical answers” calling them hallucinations. Which becomes common when prompts are outside of it’s training set.

So yes there is definitely a link between the internal “Correctness” score and “Truth” but they are measuring very different things.

How confident is it in it's confidence interval?
In what practical or observable sense does that matter if someone never acts as if they know what they're saying is incorrect? Also, assuming the Dunning-Kruger effect is mostly accurate, there will be large portions of the population genuinely believing they are correct when they are not.
> In what practical or observable sense does that matter if someone never acts as if they know what they're saying is incorrect?

People usually do act as if they know what they're saying is incorrect, especially if it's a case of lack of recollection rather than a conscious and strategic lie. ChatGPT defaults to offering a continuation with the correct format and plausible text strings, without regard to truth values it doesn't know because it's indifferent to them.

And even if human learning or human motivation did lead a to person to internalise nothing about, say, the titles of scientific papers but the general format and authors well enough to make up similar sounding ones on demand in future, they probably wouldn't be as good as it...

ChatGPT also has no skill whatsoever in logic or deductive arithmetic - just like some humans - but would you want such a human performing professional knowledge work?
A calculator can do arithmetic so why learn it? Editors have spell check and auto-complete why learn?

It's the age old question, if you'll wear diapers when you're 90, why bother with toilet training?

This. And on top of that chat gpt is the faulty calculator, spell checker with incorrect syntax, the - thanks for this image - the leaky diaper.
(comment deleted)
(comment deleted)
I think the key skill being overlooked is the ability to spot when someone (or something) is wrong, and correctly and accurately call it out and justify why it's wrong.

Critical evidence-based thinking and the ability to properly apply understanding to solve a problem, and validate or disprove a proposed solution is something that a language model struggles to do, as it can't apply the reasoning process on facts and circumstances and context.

GPT output can be very believable and confident, but I would expect a good student to be able to question and interrogate a position and establish whether it is accurate or not, based on an evaluation of the available evidence.

Education moves slowly as a sector, and I doubt many have changed anything in the last month or two, but the pandemic and a shift to remote exams or "open book" exams has (for good courses and educators) resulted in more of a focus on explanations, justification, application of knowledge to new areas, and critical evaluation. If you could copy it down in a book and repeat it in an open book exam, chances are a language model poses the same threat.

This is where a tool like GPT (when it's accurate) is really helpful. I have been using it to try & generate trade-offs, disagreements & things I may not have considered. So far it's been helpful at helping me understand some areas I should look into closer.

It's a great idea generator & tool to bounce things off of to use as part of your own execution in producing your own work. It's hard to find people with the right skills, who have the time & same schedule to do this.

Indeed - but you still require that "subject matter expert" ability to actually evaluate the output and figure out if it's coherent or correct. That tends to require a fairly high level of techniacl knowledge (and confidence your knowledge is right).

I do wonder the extent to which GPT-type tools could become teaching/learning aids - where a student asks for 5 arguments both for and against, and then fact checks them, figures out provenance of arguments, and validates them. The understanding that a student will get from taht process will, as you said, drive you to look closer at things, and ultimately if you are critically evaluating it, you'll probably learn about the subject pretty effectively.

In the same way that it's often said the best way to learn something really well is to teach/explain it to others, I think the same could be said for carefully and accurately rebutting something and explaining why it's wrong.

> In the same way that it's often said the best way to learn something really well is to teach/explain it to others, I think the same could be said for carefully and accurately rebutting something and explaining why it's wrong.

Yes, exactly this. It can be hard to find this though. I'm unable to get this out of Google or social media search engines. GPT is doing the best for me so far at helping me find them.

> Indeed - but you still require that "subject matter expert" ability to actually evaluate the output and figure out if it's coherent or correct.

Yes, you still have to do the work but knowing what to look into & where to put the work is much more than half the battle. It's giving you an opportunity you may never have had before which I think is as big as the Internet or written books or programming.

That's not what education is about. AI is not novel concepts, and learning is not just rote facts. Things are answerable by AI because humans have written it and recorded it.

It dismays me seeing such poor takes everywhere about how they think education will falter and people ask questions like this, it's almost comical.

Perhaps people should start to learn that education is about learning how to learn and applying that to new concepts.

You haven't provided any solution to the problem though. ChatGPT is able to digest new material and provide reasonable answers to questions about that material. That is what is unique to this.
You are not even describing a problem, and you seem to thoroughly misunderstand how ChatGPT works, it is able to digest new material, where does that material come from? Getting answers to basic facts is not a valuable skill and is not something that is valued in workplace that much either.
The OP did describe the problem. You can feed it a story live in the chat dialogue and then ask it questions about that story, or ask it to rewrite the story in a different style. I've used ChatGPT enough to be impressed, that the capabilities are clearly an order of magnitude better than any chat bot I've interacted with. I understand where its limitations lie. However, it is, for example, able to consistently answer above 50% of the questions on the bar exam for attorneys. It is probably able to pass an SAT with a very good score. Answers to basic facts do matter for academics because a large part of our modern schooling system is just rote memorization and regurgitation of facts. Even for subjective interpretations of, say, works of literature. You can direct the GPT to write an essay arguing from virtually any perspective you desire. I do think it's something worth discussing and not being dismissive about.
ChatGPT is cool, but you're misappropriating what you think is a solution to a problem that you cannot understand. If you cannot understand the question, it doesn't matter what the solution is. Why does it matter if it can answer questions on a bar exam? You know who can also answer that - if you have a copy of the answer key... And it does not matter to academics, rote memorization is a proxy for learning as it is very difficult to observe how well people learn and like I said, education is about learning how to learn. The point is ChatGPT does not learn how to learn, its regurgitating tokens from training data, and that training data is not coming from itself.
It's because we have experienced education that serves a few functions: it's a daycare for working parents, it's a place for children to socialize with other children, and the educational purpose it serves is that children at the bottom don't get left behind, there's a minimum amount of education everybody gets rather than some kids getting none.

This last point means it doesn't particularly serve people around or above the median, which is why, in fact, it feels to many of us that the education we experienced was about learning rote facts.

Now maybe that's not the education you received, but it is the one I received. Asking if students have started questioning it is in fact asking if students are using it to question the value of rote facts (as students have rightly been doing since forever anyway.)

A lot of people are about to get a rude awakening about the differences between the concepts of information, knowledge, and wisdom.
How is this different than "what is the point learning something that is answerable by Google, an Encyclopedia or common genre specific book, a calculator, etc?"
Because the content on those other sources is copyrighted and non-responsive.

I guess GPT-3 is more akin to a calculator. And yeah, tons of kids ask “why should I learn [this math equation] when I can type it into a calculator?”

The difference is the calculator is guaranteed to be correct if the inputs are done properly.

It’s still worth learning the “why” of a non-solvable topic for intellectual curiosity and because there’s no bottom or border to these topics.

But yeah, for stuff you aren’t interested in, I can see lots of kids wanting to use GPT-3 to just “get through it”.

First, it may not be. It's entirely possible that ChatGPT & it's successors will just become another tool that we use - the bestest auto-complete/intellisense in the world, if you will.

That said, qualitatively it feels different because it can produce an answer specific to your question. Like, a lot of 'Googling for the answer' involves putting puzzle pieces together and combining that with what you already know to get an answer. In contrast, it feels like if you wanted a better program you can copy and paste your code into ChatGPT, ask it "How do I improve this?" and it spits out an improved version with a short explanation. No more poring over web pages with advice about code quality, no more trying to figure out how to apply that advice to your specific program, etc.

So at this point it's not clear where things are headed, but it certainly seems reasonable that this may be qualitatively different.

Well, I think there's a difference between something being answerable by ChatGPT, and someone being able to take that answer, understand it, and use it to deliver value.

Moreover, while ChatGPT is really good, it can make mistakes, and if said mistake isn't an obvious one, then you'd need to have a good general understanding of the thing you're asking a question about to spot that mistake.

We should also consider that ChatGPT won't be free forever, so it might be a bad idea to change teaching in a way that assumes that these kinds of things will always be accessible to students and wherever you're going to work.

> We should also consider that ChatGPT won't be free forever, so it might be a bad idea to change teaching in a way that assumes that these kinds of things will always be accessible to students and wherever you're going to work.

If recent trends are any indication, we should consider that ChatGPT will probably be supplemented by a number of open source GPT models, likely within the next 12 months, and a rich ecosystem of developer activity will continue to evolve and flourish.

Haven't we already been through this with the advent of Google?
And calculators before that. I think that's the whole point of OP? Whether this new cycle has started yet.
But that immediately suggests the answer, which is that nothing much changes. Maybe you have a new tool to learn about, but the fundamentals of education are the same.
Indeed. At the same time, we should probably take the availability of technology in consideration when designing educational curriculums and put more emphasis on the things that still distinguish us from computers. For example, high schools eventually allowed (even mandated) the use of calculators and a lot of engineers today probably wouldn't be able to do long division by heart without having to Google it. And some schools have started spending less time on teaching hand writing (e.g. stopped teaching cursive) now that smartphones and computers are so ubiquitous.
Probably "You won't always have a ChatGPT in your pocket" :)
We've had calculators that can do arithmetic for decades but we still teach kids how to do math.

Also, ChatGPT in particular is a terrible "source of truth". It gives plausible sounding answers, but they may not be factually correct.

I just asked ChatGPT to write a press release for a new scientific discovery, which would be impossible to actually happen. It happily did so, making up a name for a new element along with creating a fake quote from a made-up researcher at a made-up National Laboratory. You could do the same for historical events, current institutions, anything.

(comment deleted)
Often times answers are the easy part. It’s unfortunate that this comes across as the basis for much of US education, especially K-12.

As most (all?) researchers will tell you, asking good questions is the often times the hard part.

A good curriculum, sadly rarely found, will teach you how to think and how to ask good questions. Chatgpt doesn’t really impact this, imho.

This is ultimately no different than search engines & internet access have been since the late 90s. ChatGPT may be able to answer a question slightly faster than a search in some cases. I suppose the one unique difference with ChatGPT is you can "teach" it some new material that it can infer responses from, to a limited extent.

A simple answer would be to have students take handwritten exams in class, with no electronics allowed. This is not always possible (ie online classes). Homework has always been subject to a number of methods that relieve the student from the burden of actually learning the material (ie parents helping, copying other students, searching online, etc).

Online classes are a tougher problem to solve. Maybe the answer for that is requiring that they be paired with proctored exams.

Search engines cannot write essays for you.
Plenty of online services provide that service for a small fee. And before anti-plagiarism detection sites cropped up, it was quite common for people to copy paste essays from the internet. Also, let's consider that you find an essay online that fits your topic, it is not that much work to reword an essay to make it unique to your style of speaking without needing to really read the original source material.

Teachers aren't all that original in the topics they ask students to write essays on. Often, a teacher or professor will use the same questions for years, and they usually fit common themes, books, literature, that other teachers and professors frequently cover. So it's usually trivial to find resources online for this. I've even seen university level exam questions shared in various forums so that students need not do anything but rote memorize the responses.

Yes, it is true that there are online services that provide essays for a fee and that some students have used these services to plagiarize. It is also true that some teachers reuse essay topics, making it easier for students to find information online to use in their assignments. However, it is essential to note that plagiarism is a serious violation of academic integrity and can have severe consequences. In addition, rote memorization of answers does not promote critical thinking or deep understanding, which are essential skills to develop in higher education. The best way to succeed in school is to engage with the material, ask questions, and put in the effort to understand and synthesize information in a way that is unique to you.
While it may be tempting to copy and paste an essay from the internet, this can result in serious consequences such as failing the assignment, failing the course, or even being expelled from school. It is always best to put in the effort to create an original essay rather than risking academic dishonesty. I must remind you that I am a language model and cannot recommend or endorse any specific website or service, including this http://essaypapers.reviews/ . It is important to carefully research any service or website before using it to ensure that it is reputable and trustworthy. As for finding essay resources online, it is important to ensure that the sources are credible and reputable. Simply finding an essay that fits the topic is not enough - the essay should be properly cited and referenced to avoid plagiarism. It is also important to read and understand the material in order to create an original essay that showcases your own ideas and understanding. It is not appropriate to use essay writing services or to plagiarize in any way. These actions undermine the learning process and academic integrity. Instead, students should take the time to research and create their own original work.
I think ChatGPT will cause similar problems to those already experienced in STEM disciplines, particularly math. Students can use a camera based app on their phones that can do an entire worksheet in seconds, computer algebra systems are baked into calculators, Wolfram Alpha can solve fairly complicated problems, and googling effectively can give the answers to the majority of homework questions through a rigorous undergraduate math program. Math class has always made students ask "when are we going to use this?" Since it is already difficult to convince students that math is useful, ChatGPT will probably make students question the usefulness of most of what we currently teach in schools.
I’m about to graduate with a computer engineering degree. I’ve messed around with ChatGPT a fair bit. If I ask it about topics in my curriculum, it’s either wrong or it spits out a high-level overview of some extremely basic concepts that’s less useful than the end-of-chapter summaries in my textbooks.
that's like asking what is the point of learning something that is in a book...

seems like the wrong question

I teach CS at a university.

I've largely given up on trying to "catch cheaters." At this point, either a student wants to really learn the information or they don't.

What are you gonna do? Cheat your way to a programming degree and then... bomb at some programming job because you don't know how to actually code or think?

I try to set students up for success, w/clear & obvious requirements (e.g. test suites for project grading) and then appeal to their sense of reciprocity: I'm trying to be cool to you, please be cool w/me & don't cheat. If you feel like you need to cheat to pass, come talk to me so we can address the material instead.

Regurgitating was never a great way to measure understanding anyway. In an ideal world, I'd have the time to sit down with each student and have a 10 minute talk, and get a sense of how well they understand the material, like the old english system.

> In an ideal world, I'd have the time to sit down with each student and have a 10 minute talk, and get a sense of how well they understand the material.

Aren't in-person exams a decent way to gauge that?

They are in my experience, we do this as part of many of the courses and projects: they get an assignment followed by a 15-20 minute assessment of their work. Gives us the chance to measure their understanding of the work and it gives them the chance to clarify things that they didn't make clear in the work, which can lead to a slightly better grade.
I feel like the first thing I'd do as a teacher is embrace it. Take the assignment I was going to give, pass it through ChatGPT and give that to the student. Their job is to find the errors and omissions and rewrite without them.
ChatGPT has only made me question who will write confident-but-dead-wrong SEO spam in the future. I would only question my curriculum if I were averse to doing any research myself and were willing to blindly and explicitly trust a black box on a website run by strangers.

I don’t really understand the uptick of “Rate how much the idea of a confident-sounding chat bot has shaken your world view to the core on a scale from 1-10” posts. It’s a language model, not a New Internet.

Even if ChatGPT could answer all questions correctly, you would still need to know which questions to ask.
The point of learning is to apply the knowledge of the subject matter through something. If you just let the AI answer for you, you haven't learned anything.

In most cases, homework / basic questions are to strengthen your knowledge. If you are letting the AI do this for you, are you able to have a conversation about said topic?

Changing what teachers teach will not change the problem. Students will always use this bot now. The cat is out of the bag.

You're making a very big assumption here, which is that the point of learning something is to be able to answer specific questions of the kind that ChatGPT can answer. This is wrong. Students below the graduate level do not, in general, produce valuable work in the course of their studies. I promise you, your teachers do not need your help to write a high school freshman-level essay on Romeo and Juliet, any more than they need your help to factor a polynomial or discuss the causes of World War 1. The point of learning is to improve your brain. Asking why a student can't use ChatGPT to write an essay is like asking a jogger why they don't use a car. It's missing the point entirely.

People have been having this same discussion about math for many years. A copy of Mathematica can answer just about every question on every math test you will ever take. So what? We didn't stop having to learn math in order to understand anything that's based on it. Mindlessly quoting answers from ChatGPT will not teach you new ways of thinking.

As for the knowledge itself, the key thing to understand is that knowledge inside your brain is more useful in more circumstances than knowledge outside your brain. ChatGPT won't notice connections between unrelated subjects for you. It won't notify you in advance when a question ought to be asked.

That's faulty reasoning, why learn something since it can be looked up?

Because understanding and looking it up are not the same. We've been able to look up things since we've had libraries and experts.

One one hand, you may end up doing something really stupid/dangerous just because you don't know what the answer means or what are the implications and blindly do what the statistical model tells you.

On the other hand learning gives you "HD vision" for the world.

Anything new you learn will let you see much more detail of the things around you, and you don't really know what you're missing, you really can't know, because until you learn it, you're blind.