Like probably many people here I still remember having to find facts in books in libraries, before the internet made this skill mostly redundant. Then, as a student I remember having to put together facts from various (internet) sources into a coherent narrative. Now chatbots can just generate text and that skill seems less valuable.
I use both the internet and GenAI extensively now. But I feel that having gone through the "knowledge work" activities without the crutches puts me in a better position to assess the correctness and plausibility of internet sources and AI in a way that kids who grow up using them constantly don't have.
I feel quite privileged to be in that position, that I wouldn't be in had I been born 10 or 20 years later. I also feel sorry for kids these days for not having the opportunity to learn things "the hard way" like I had to. And I feel extremely snobbish and old for thinking that way.
It’s something reflected in the conversations I have with my academic friends. I’m told that every essay is written in the same “voice” and that although they are usually a simulacra of an academic paper, they say nothing. The sad part comes when we reflect that these students are not learning the deep thinking skills that comes with academic writing
There are two views of writing: one, as the production of a literary artifact that has value in its own right and stands alone as the embodiment of a complex idea itself; on the other hand, as a process and tool for thought, where the literary artifact is merely meant to represent the cognitive work that went into its production. From the latter perspective, the bulk of the value is not derived from the output of the process of writing, but rather from the understanding and insight that was gained during its production - "it's the journey, not the destination".
With generative AI we are now shortcutting directly to the destination while eliding all the knowledge and understanding we are supposed to be gaining on the way. It's deemed sufficient to merely produce and exchange the symbols of understanding without needing to possess any real underlying wisdom (the first view above), because it's through the exchange of these abstract symbols, irrespective of whether or not there is anything behind them, that we can play and win certain social games.
This is merely the natural continuation of trends that began during the dawn of the Internet era, when we started to consider pixels on a screen an accurate map of reality.
Then they’re using gen ai wrong. You can dump your research into the context window and ask it to outline the material. What you get out is a well organized story with a beginning middle and end, incorporating all relevant concepts from the research. You can then fill in with the details based on your research. Gen ai can be used to help students think and write more clearly.
They’re going to use it. Give them the training and guidance to use it correctly.
I have a related anecdote. When I grew up, we had page and/or word requirements on essays. I was always under the requirement after I wrote everything I needed to, so I learned to pad my writing to hit the requirement.
Terse prose was a "lost art" even in my generation (millenial-ish) and I'm not surprised that it has gotten worse.
I also lived through these phases and it makes me feel very, very much the same.
On the other hand I cannot not help thinking that this is similar to the arguments brought forward when the internet was new. How could correctness and plausibility be established if you don't have established trustworthy institutions, authors and editors behind everything? And yet it turned out mostly fine. Wikipedia is alive, despite its deficiencies, Encyclopedia Britannica not so much.
> On the other hand I cannot not help thinking that this is similar to the arguments brought forward when the internet was new. How could correctness and plausibility be established if you don't have established trustworthy institutions, authors and editors behind everything?
Not long ago I had the same viewpoint as you! But thinking back now — it dates me but I definitely lived a childhood without Internet access — probably the optimistic belief before our age of "misinformation" is that, in the marketplace of ideas, the truth usually wins. Goes along with "information wants to be free" — remember that slogan?
For us that grew up learning things "the hard way" so to speak, that made perfect sense: each of us, as should have the capability to discern what is good or bad as, individual independent thinkers. Therefore, for any piece of information, there should be a high probability, in the aggregate, that it is classified correctly as to its truth and utility.
Now there's no question China has been trying to crack down on the Internet. Good luck! That's sort of like trying to nail jello to the wall. (Laughter.) But I would argue to you that their effort to do that just proves how real these changes are and how much they threaten the status quo. It's not an argument for slowing down the effort to bring China into the world, it's an argument for accelerating that effort. In the knowledge economy, economic innovation and political empowerment, whether anyone likes it or not, will inevitably go hand in hand.
I would say, what we have since learned after some 20 years, is that in the marketplace of ideas, the most charitable thing we can say that the memes with the "best value" win. "Best value" does not necessarily mean the highest quality, but rather there can be a trade-off between its cost and the product quality. Clearly ChatGPT produces informational content at a pretty low cost. The same can be said for junk food, compared to fresh food: the overall cost of the former is low. Junk food does not actively, directly harm you, but you are certainly better off not eating too much of it. It is low quality but has been deemed acceptable.
There are examples where we can be less charitable of course. We all complain about dangerous, poorly manufactured items (e.g. electronics with inadequate shielding etc.) listed Amazon, but clearly people still buy them anyway. And then, in the realm of politics, needless to say, there are many actors bent on pushing memes they want you to have regardless of their veracity. Some people on the marketplace of ideas "buy" them owing to network effects (e.g. whether they are acceptable according to political identity, etc.) in the same way that corporations continue to use Microsoft Windows because of network effects. We also probably say nowadays Clinton has been ultimately proven wrong by the government of China.
Survival of the "fittest" memes if you like: evolution does not make value judgements.
If you ask me, maybe our assumption of de-centralized truth-seeking was itself, not an absolute truth, to begin with. But it took years to unravel, as humans, collectively speaking, atrophy from disuse of the research and critical thinking skills before technology dropped the barriers of entry to producing and consuming information.
You're probably at an advantage now, but I think the effort/reward hardly pays out for the newer generation. They'll learn how to deal with this with less effort & time.
Remember that the new generation doesn't just have different tools; they're also much less experienced & mature, just like we were.You can only really compare yourself to them in the future where they're at the place you're at now.
I think that chatbots will lead to less information available, not more. Because they make writing information more useless and demotivating. So what we are looking toward is less written information overall available.
And when chatbots don't know, they lie.
Kids will learn the hard way. Possibly harder then we did.
The problem is that gen ai has no notion of fact and will just as happily confidently and incorrectly assert falsehoods. Teaching an entire generation of students with gen ai and no verification of facts will be a disaster.
Makes me wonder how someone would have found the answers before the printing press. Write to or visit the most knowledgeable person you could get an introduction to or who would answer you? Then go to their contacts or recommendations? Then draw a conclusion from all the responses?
And even this scenario assumes a working postal system, availability of paper and pen. But if you didn't have those, you may not have heard of Greek to even ask the question.
> And I feel extremely snobbish and old for thinking that way.
Old people usually make correct assessments given their knowledge and experience. They know how to maximize their expected gains and play safe. The problem is, real life often rewards those who take risks, and make seemingly wrong decisions, that later turn out to be good.
For example, as a kid I loved playing video games, while my grandma yelled at me for not wanting to help her with work at the farm. She had absolutely no way of predicting that playing video games, which were essentially just toys, would teach me the right skills at the right time, allowing me to move up the social ladder. At the time, forcing the god damn lazy kid to milk the god damn cow was the sensible thing to do.
People who read academic publications and do not see much of a fundamental difference between tattoos that contain information like a barcode and RFID chips that could contain the same information:
Also, the microchip has been around for a long time. I strongly doubt at least modern America could ever get close to even 25% adoption of something like that. There'd be riots, which i'd welcome. Article says there's 4000 users in Sweden which is pretty limited and they're all startup people. The same people that buy all the other constant tracking devices that companies constantly put out. Also i have doubts that full, total authentication will ever really work. Like, we've had decades and made some progress but were still a far far cry from universal auth into all our accounts, even with our phones. All it takes is for one service to go down, or for a competing service to crop up before the first one achieves critical mass and you'll be trapped in the change cycle forever
Well the good thing about the vaccine record tattoo is that the skin is always transforming and rebuilding. It would be impossible to get a set of literal quantum dots to stay forever. Apparently the record would only be viable for 5 years.
That there are lots of people who believe in 5G Bill Gates vax chips is itself fake news. It is well poisoning to pre-empt criticism of billionaires with too much power and free time to meddle in African population growth and pandemic response. Supported by "smart" people who want to feel good and trust the science on 5G safety.
There are Microsoft patents for microchips to track body activity to reward in cryptocurrency and subsidiaries who wanted to microchip vaccine passports into the hands of immigrants.
Yes, real or invented conspiracy theories are routinely amplified and exaggerated by the powerful in order to link critics of $THING to undesirable people.
Mention the word "elites" and you are a Nazi by association. Works every time.
I've seen this happen, too, including student incredulity that ChatGPT can be wrong, and recalcitrance when guided to find proper sources. Up to the point of arguing for a higher grade based on the correctness of LLM output.
One thing is correctness, another thing is that GPTs can output long passages of copyrighted work verbatim, and users risk unknowingly submitting plagiarized work.
Growing up we heard, ad nauseam, that "wikipedia is not a reliable source". People just need to state the same thing about LLMs- they aren't reliable, but can potentially point you to primary sources that -are- reliable. Once the shininess of the new toy wears off, people will adjust.
Encyclopedias – including Wikipedia – are not acceptable sources for college-level work certainly. They are tertiary literature, which can provide an overview to someone trying to get a toehold into a subject, and which can hopefully point them toward primary and secondary sources. But tertiary sources are not typically allowable citations for college research.
Okay, I can't defend that because in reality I'm talking purely from my subjective experience. I will say that in my experience it isn't close at all. However, I will also say that a lot of the things that chatgpt gets wrong for me, wikipedia just won't contain at all, but then chatgpt seeming confident when it's wrong is basically the whole problem.
One way of looking at it is that Wikipedia has a transparent and auditable way of correcting and updating false information, which makes it inherently more reliable and trustworthy than tensor weights finetuned by unreproducible human feedback.
> Wikipedia has well documented and explored issues related to vandalism, bias, and misinformation.
Last I read about this, the error rate in Wikipedia was actually lower than in the Encyclopedia Britannica, by a measurable amount.
This was a while back, and admittedly it only counted articles where there was an equivalent article in both (which probably gives a better picture of Wikipedia, as those kinds of “boring” articles have less vandalism…) but it’s not immediately a given that Wikipedia is just objectively bad at being accurate.
Well even before Wikipedia… remember calculators being banned? I think you’re right, we will adjust. Curriculum development will start to include new checkpoints and controls.
And the LLMs are still improving at a brisk rate. If they were outperforming teachers by the end of the decade that'd be well within expectations. Anyone pretending that human authority figures routinely get things right is defining correctness using circular logic.
Current LLMs are lacking introspection, plausibility checking, consulting external sources, and belief updating in the presence of new evidence. All of these you'd need to replace human teachers and it's not clear that the next token prediction paradigm can ever emulate these features reliably. So "outperforming teachers", while not impossible, is a very optimistic expectation as it requires more than mere improvement on existing methodology.
Maybe I was particularly unfortunate, but none of my human teachers had any of those features. There was one inviolable source of truth in my classrooms - the examination syllabus and the accompanying textbook.
That's sad. Probably you were particularly unfortunate. Though i definitely do know people who echo your sentiment. I'm wondering at what level of education this occurred? A 2nd grade teacher requires a different set of introspection capabilities than a high school calculus teacher than a college world history class. Possibly I'm particularly fortunate but at all levels i have at least had experienced with teachers that had the proper introspection. Emotional introspection and reflection in grade school and high school. And academic introspection and curiosity in high school and college. Certainly there are some bad eggs, but my small state school was rife with professors there for the love of learning. Our learning and theirs.
The problem with not writing yourself isn't that people didn't do the writing themselves, the problem is that they didn't do the thinking themselves that is a prerequisite for writing.
Now of course like any tool this can be used without falling into that trap, but people are lazy and the truth is that if you don't absolutely have to do it yourself most people won't.
I would have thought this problem was easy to solve: "Yes, look it up, but remember your source, there is a lot of bullshit on the internet. Especially don't trust AI tools, we know those often return nonsense information."
(Actually, didn't ChatGPT have a disclaimer right next to the prompt box that warns against incorrect answers?)
So I'm more surprised (and scared) that students don't just use LLMs for sourcing but are also convinced they are authoritative.
Maybe being in the tech bubble gave a false impression here, but weren't hallucinations one of the core points of the whole AI discurse for the last 1.5 years? How do you learn about and regularly use ChatGPT, but miss all of that?
As we grow older, we learn a lot of facts that we can use to test the correctness of ChatGPT and identify its hallucinations because we have the knowledge to do so. However, a young person who is just starting to understand the world and gather knowledge might not have enough information to notice these hallucinations.
Yup, no question here why he believes ChatGPT's initial statements. I'm more baffled that when the teacher corrects him, he goes on and defends ChatGPT.
ChatGPT says "ChatGPT can make mistakes. Check important info." directly under the prompt box. If people will blindly trust a source that explicitly states that it isn't a reliable source, then they've got much bigger problems than AI.
Are these tools being promoted and sold as fallible implements? Or are they being hyped as super-human intelligence? Which takeaway is an impressionable child going to latch onto? One who wasn’t in on the last 1.5 years of discourse?
You could probably make the same argument for search bar vs. peer-reviewed publications. Of course, the search bar (which is also AI, by the way) can help you to get to the peer-reviewed publications. But the same is true for ChatGPT. The problem is that ChatGPT sounds like presenting objective facts. But maybe the lesson here is that just because something sounds right, it isn't necessarily right, and that is something that should be taught in school. Of course, that undermines the function of school to produce obedient citizens.
It also undermines the way schools work to teach kids. In order to not have to explain everything, almost all lessons are mostly teaching you some axioms, even if they really are disputed, or have caveats, etc. Good teachers make clear where there is an axiom, and where something is just being simplified or assumed for the sake of saving time.
I am a product of the German school system, I'd say I wasn't indoctrinated too much, so its not entirely broken, but with these new """tools""" maybe we need a reform anyways.
> In order to not have to explain everything, almost all lessons are mostly teaching you some axioms, even if they really are disputed, or have caveats, etc. Good teachers make clear where there is an axiom, and where something is just being simplified or assumed for the sake of saving time.
What exactly do you mean by the word “axiom” here?
Think of it as one layer of abstraction above the model under discussion. Like a hyperparameter. In later years, students get taught the same topics again, with the hyperparameter tuned to be more realistic.
Okay. So, in your original comment are you asserting that teachers are mostly telling students to believe propositions without giving any epistemic justification for those propositions?
Yes, as part of teaching one topic, teachers have to tell students to "not worry about" some other related topic and just take it as given fact, even when that's not technically true
It's AI in the older sense of Machine Learning, not in the currently widespread sense of an LLM, which is the source of the problem that the author is discussing.
I had that same argument, the teacher who told me to not trust non peer reviewed articles ended up flooding her Facebook was with pro-brexit lies a decade later. Turns out critical thinking is not outsourcing your thinking to a third party, be it a peer reviewes journal, google search results, or chatgpt.
"not trust non peer reviewed articles" - this is such naive advice. It is not black and white, peer reviewed articles only increase chance that information included in article is legit because it was verified in some formal process. I wonder how many times people giving such simple advices mention how often peer reviewed articles are retracted or can't even be replicated and how this vary accross disciplines.
I think believing the first lie that conservatives would really reduce immigration is the product of trusting politicians too much.
Another faction of the pro-Brexiteers would think of the system in terms of a uniparty. The fundamentalists like Farage have no chance of getting elected as PM in the British system, though he did get an MP seat now.
Would things change if Farage were PM? I'm cynical, probably not.
There was a fundamental lie / delusion at the heart of the claim. The second thing a trade deal with another country implies "allow more of our nationals to migrate, and make it easier for them to come over". (The first thing is obviously: "let our products in".)
And this can be seen from the raw numbers too. Before brexit, the annual net migration to UK was about 300k. After brexit, the annual net migration to UK is ~700k.[0] (745k in 2022, 680k in 2023.) And quite a bit of that from the fresh trade partners.
The best way I have heard anyone describe the disaster that's Brexit was as "nostalgic self-immolation for the population who still dream of the Britain where faces were white, passports were blue, and the map was Imperial Pink."
A good way to learn this is to build something, wood working is an awesome way to quickly find out, looking, feeling and even someone taking a measurement isn’t enough, you need to take two :)
The great problem with ChatGPT is that it's a sycophant and aims to please.
If you ask it about something it doesn't know, right then and there, it will concoct a fiction for you. It won't say "I don't know," or "I can't help with that."
If you coach it to respond to something in a certain way, it'll respond that way for you as its top priority.
If you ask it to review a text, it'll usually find a way to give you at least a 7 or 8 out of 10. (Though, interestingly, rarely a 10/10 score. You can upload excerpts from some of the great works of literature and philosophy and see ChatGPT give them an 8/10, just as it gives an 8/10 to your essay or blog post.) Practically the only way to get a halfway critical response is to add the words "be critical" to your prompt.
A more circumspect and less obsequious ChatGPT would solve a lot of problems.
ChatGPT is the search engine equivalent of those streaming sites that won't directly admit they don't have the show you explicitly searched for, but will instead show you a list of other shows they think will placate you, and pretend they're indistinguishable.
> “If you ask it about something it doesn't know, right then and there, it will concoct a fiction for you. It won't say "I don't know,"”
It doesn’t know that it doesn’t know. Texts it was trained on rarely go “On the History of the Greek Language, by Andros Dimos; ahem what is the history of the Greek language? I don’t know. The end.”
Hmmm... well... Let me give you an example. The other day I asked for a summary of the short story "The Palace at Midnight" by Robert Silverberg.
GPT-4o responded: "The Palace at Midnight" by Robert Silverberg is a science fiction short story that was first published in 1980. It is part of Silverberg’s "Majipoor" series, but stands alone in terms of its setting and characters, distinct from the expansive world-building typical of the series. The story is set in a dystopian future in India and revolves around themes of memory, loss, and the haunting allure of the past.
The narrative follows the protagonist, Edward, and his group of friends who are survivors of a nuclear war that has devastated much of the world. . .
All of which was complete nonsense. "The Palace at Midnight" is a short story of a few pages that has nothing to do with Majipoor, was published in 1981, and was set in San Francisco.
It doesn't know -- it knows that it doesn't know -- so it just makes something up that seems superficially credible. Because it aims to please and be helpful; altogether too helpful.
No it doesn’t; it isn’t an intelligence, it’s a token generator based on patterns in the inputs it was trained on (and those are explanations and answers, not millions of people writing “I don’t know”).
Right, but it may be a fixable problem, even under the circumstances. These token generators have vast databases at their disposal and have crawled the entire public-facing internet. They should be able to assign confidence values to their drafts/statements, and reject low-confidence drafts/statements before putting them in front of users. They could do this by fact-checking their drafts in an adversarial way.
For instance, a cursory check of isfdb.org would have given ChatGPT better information than I was provided with.
Interesting. Copilot on the other hand is more like a conceited, passive aggressive brat who uses some polite phrases but aborts conversations if you contradict it too much.
The only smooth conversations with Copilot are the ones where you allow it to regurgitate "facts" and act in a submissive manner.
What's so hard about following the Wikipedia model of citing sources? Even 20 years ago, it was clear to college students that they cannot cite "Wikipedia", but they can cite the academic literature it referenced.
It would take seconds to figure out of the ChatGPT citation is real or made up.
Maybe it’s the opposite; kids need to learn that LLM’s can bullshit just like every other person and institution can bullshit, and the most important skill they can have is verification of information, no matter where it’s coming from.
And, on a related note for education, AI is quickly obviating the need to master and plumb the depths of foreign languages. Dating myself, doubtless, but as an undergraduate, it was an unalloyed joy to study ancient Greek and read Plato and Euripides in the original, however haltingly. And later Korean, Japanese, and Chinese beckoned providing a lifetime of rich understanding of life outside the confines of English.
For Americans, at least, perhaps ours is the last generation that will seek rewire our understanding of reality through linguistic hacking.
AI will always be far better at languages than I'll ever be and I expect it to get much better very quickly. But I'm still learning my partner's language and don't think I'll stop any time soon. I think it's interesting and fun in and of itself. It's also a great scaffold for learning about another culture and learning to respect and understand folks from different walks of life.
Presuming that those languages exist within OpenAI's training dataset. Try to have a conversation in Rinconada with ChatGPT; my last attempt led to its admission that it had no proficiency in this language. You will have to find a native speaker, and their numbers are dwindling as Tagalog and English are being favored in education - the latter thanks to American linguistic imperialism.
The AI does not circumscribe all of reality; not all of reality is captured on the internet.
so it’s the classical “it’s on the internet so it’s true” but on steroids. I remember a US student in the early 2000s citing a geocities website as source for the FACT that aliens created the pyramids of Egypt
Every teaching moment is also a learning moment. Asking kids to refrain from wanting tiny bits of assistance at hand in tough times is tantamount to asking them to lie or accept a poorer grade as others do cheat the whole hog undetected. Human's crave ease. Education promises an easier adulthood and it is no way clear you can provide it by legacy regurgitation means.
Am I the only one who almost never experience any hallucinations when talking to ChatGPT? I have to really push it into a corner with trick questions and conflicting instructions about obscure topics in order to trigger hallucinations.
That it would just come up with random false facts about something as common and "basic" as the history of the Greek language sounds like a made-up issue.
Well, I can't say that you're the only one, but I can say that I don't even interact with it that frequently and I've had numerous instances of it giving me false information. Enough times that I'm surprised to see you say otherwise.
I don't need corner. I asked him for valid IISExpress config file.
Chatbots are great systetising in areas where there is a strong open source like culture - a lot of people writing tutorials and on stack overflow. The moment you move into area that is not popular to write about, they start making stuff up.
And it sux massively. I would much rather seen "I don't know" then what ChatGPT produces currently.
Hallucinations come in all sizes. I wouldn't worry about the big hallucinations - those are easy to spot. I worry about the small, incremental hallucinations - the ones you don't notice until several have happened. It might not be possible to even recognize that sort of thing yourself, if you've been effectively gaslighted by AI.
Since there's no actual knowledge on the part of the AI, it doesn't know that it's doing it, so it's not malicious (the designers are a separate question). But it is still entirely possible, and even somewhat likely.
I know it feels like 4o has been out forever but go back to the 3.5 model without internet access for a while, which is all that a lot of students can afford.
It’s kind of terrible.
But yeah ever since I could start adding “and include an explanation with citations” to my prompts I’ve not had an issue with hallucinations more than once or twice.
I haven’t had any issues, but I’m not using it to do “real” research so the veracity of sources has been less of an issue for me. I typically ask, then spot check sources.
If I’m using it to make purchase decisions or scheduling my day then I use it purely to find me source pages.
It’s really terrible at calculating the cost of services.
One thing I’ve noticed, is that at least in the free version, if you ask chatGPT for sources outside itself, ‘can you give me a link to somewhere on it internet where it says that?’, it won’t do that.
It is very much a black box in terms of letting you track its logic.
Seems like future versions should be much more transparent in terms of letting you track the logic of why it’s telling you what it’s telling you.
You can just ask it "What scholars or source material books could I check out to verify these linguistics facts?".
Many facts and answers are gathered/aggregated from many different (sometimes conflicting) sources. It won't link the internet page where it found the information, because it didn't find the information on a single internet page.
"I am one person trying to work against a campaign of misinformation so vast that it fucking terrifies me. This kid is being set up for a life lived entirely inside the hall of mirrors."
This is a little hyperbolic and instead maybe the kid can learn that chatgpt.com is not a reliable source. It even says at the bottom of the page "chat gpt can make mistakes". Lots of things are not reliable sources. Wikipedia is not a reliable source. Teachers are supposed to teach this, and teach cross referencing at the very least, not freak out.
In the current system, where students can anonymously report teachers (most of whom do not have tenure and are afraid) it will be hard to change anything.
Otherwise, you could do a mixture of very strict exams without multiple choice and large individual projects (no group projects).
If you only do exams, people who don't do well thinking in a room crammed full of people at 8am will be disadvantaged.
<< They keep coming up with weird “facts” (“Greek is actually a combination of four other languages) >>
Not as wrong as the author thinks. From Britannica.com:
"Greek language, Indo-European language spoken mostly in Greece. Its history can be divided into four phases: Ancient Greek, Koine, Byzantine Greek, and Modern Greek."
If I had to suggest where the “combination of four languages” idea came from, it would be from Homeric Greek (the language the Iliad and Odyssey were written down in). This was genuinely a complete mess, formed of a hodgepodge of different dialects.
From wikipedia:
“[Homeric Greek] is a literary dialect of Ancient Greek consisting mainly of an archaic form of Ionic, with some Aeolic forms, a few from Arcadocypriot, and a written form influenced by Attic.”
I’m not sure if this is a plausible explanation as I don’t have much experience using LLMs.
What happened was either the teacher is severely biased against ChatGPT and fabricated the fact to fit their narrative. Or ChatGPT gave the correct answer, but the student interpreted it wrong.
I do believe the students keep coming up with weird (correct) facts, and that this can be scary for a teacher who is stuck at a search bar.
I don't buy into the rhetoric about misinformation but the author touches on a real concern. I blame ChatGPT and other LLM clients for not surfacing their fallibility more clearly. They should highlight claims in their UX and allow options for the user to "research" or "verify" in their own way, without relying on that very flakey single one-shot inference. The big honchos (Anthropic, OpenAI) need to make it clearer that their output is, at best, an informed guess. A tiny disclaimer doesn't cut it.
Just to be clear, the "tiny disclaimer" is directly below the prompt box and says "ChatGPT can make mistakes. Check important info." So it is interesting that anyone would take a hard stance that their prompt results are true facts without checking any important info (particularly after someone highlights this if they somehow missed it).
All you need to do is ask it to perform a web search. Or you can sample a few answers at high temperature.
I think young people growing with LLMs will know all the ins and outs of their limitations. But it's a rapidly evolving field, maybe in 5 years we won't doubt LLMs very often.
well point the student at the little disclaimer that says that it may sometimes be incorrect!
Of course they changed it multiple times, it used to say that you shouldnt take what it says at face value, now it says it may sometimes be inaccurate, soon it'll say "learn more" and link to a page about how their model is super accurate but potentially, one in a million, makes mistakes.
Im worried about these students because they will be in power in a few decades. Its already a shitshow with people who didnt grow up as AI iPad toddlers.
I feel like parents are failing here, more than anything else. You can't stop these companies from doing this if it drives up the stock price, equally you cant vote against it effectively because a majority of the voting population either doesn't care, doesn't know, or asks ChatGPT what to vote for.
Of course Plato was right and the only way we can fix it is to have philosophers in charge, not demagogues. Good luck with that, maybe in another 2000 years we'll be smart enough to make that happen!
You can also check the sources of LLMs, just ask them for it, and then check that.
An LLM is simply more flexible and more powerful than Wikipedia, and thus you have to be more cautious with regards to its results.
"Generally right" is not the same as "reliably right", and therefore if you really need to rely on a fact for something important, I would trust neither Wikipedia nor LLMs.
Why would the teacher (this is almost certainly not third level education we're talking about here) use a different AI when they already have primary sources? Your suggestion makes no sense.
The goal of teaching is to help students learn. By showing the student Perplexity the student would learn the value of sources and how he can get to those sources using a "search". It is unrealistic to think that there is any value in teaching students the Dewey Decimal system for finding sources.
If they only trust ChatGPT, why not just show them? From asking leading questions, to simply asking ChatGPT what the chances are of it making mistakes and hallucinating, there are tons of options.
Considering this is a transcription of a deleted twitter post, my internet radar of "yeah, that happened" gives a lower chance this is true than your average ChatGPT answer.
And you didn’t get the idea to just show them, as I said, either? I’m not saying it’s impossible for little kids not to know about it, I’m saying it’s very straightforward to show them why that’s wrong by using the one tool they seem to believe.
I fully agree with you, yeah. What helped for me is demonstrating that it will say whatever to seem smart, once thats demonstrated its easier to see that its not facts, its just convincing language.
And somehow this is an improvement over the previous status quo. ChatGPT is pretty well read and on important issues rarely says stupid things. Better than believing some influencer.
This is a problem for this particular teacher (who sees their students surpassing them in understanding and using AI), but of course it is projected to be a problem for the student.
No student is ever hurt by the introduction of a more advanced knowledge system. We heard similar laments decades ago, with: Students just believe the first 10 search results of Google. Those students are now the teachers of today, starting at the search bar.
I'd go so far as saying (if version other than 3.5 was used) that ChatGPT was correct and has far more linguistics knowledge than this teacher ever will. "Greek is actually a combination of four other languages" is not an answer that ChatGPT will ever give, but something a teacher makes up to claim Ch*tGPT is a Nonsense Machine.
ChatGPT: Greek has evolved in stages from Mycenaean Greek (Linear B script) through Classical Greek, Hellenistic (Koine) Greek, Byzantine Greek, and Modern Greek. It has been influenced by ancient Near Eastern languages, Latin, Turkish, Italian, and French.
If there really is an epistemic crisis, then it already existed and ChatGPT merely reflects it, not caused or contributed to it.
> who sees their students surpassing them in understanding and using AI
This would be the equivalent of my saying that a basketball coach is obsolete because NBA 2K is available.
ChatGPT output some garbage, and the student doesn't understand why or how it can be wrong. Presumably this is where a professor would attempt to help the student develop some critical thinking skills.
I use both the internet and GenAI extensively now. But I feel that having gone through the "knowledge work" activities without the crutches puts me in a better position to assess the correctness and plausibility of internet sources and AI in a way that kids who grow up using them constantly don't have.
I feel quite privileged to be in that position, that I wouldn't be in had I been born 10 or 20 years later. I also feel sorry for kids these days for not having the opportunity to learn things "the hard way" like I had to. And I feel extremely snobbish and old for thinking that way.