Everyday AI Podcast – An AI and ChatGPT Podcast - EP 74: Should Schools Ban ChatGPT?

Episode Date: August 7, 2023

Should ChatGPT be banned from schools? It's a controversial topic that has people on opposite ends. Some schools have embraced it and others have removed it. Today we take a look at how students are using ChatGPT and if it's helping or hurting them in learning.Newsletter: Sign-up for our free daily newsletterMore on this: Episode PageJoin the discussion: Ask Jordan questions about ChatGPT in schoolsRelated episodes:Ep 35: How Students Can Use AI to Solve Everyday ProblemsEp: 55: How to properly leverage AI in the classroomEp 56: How To Govern and Properly Use AIUpcoming Episodes: Check out the upcoming Everyday AI Livestream lineupWebsite: YourEverydayAI.comEmail The Show: [e-mail] with Jordan on LinkedInTimestamps:[00:00:17] Daily AI news[00:05:30] Schools shouldn't ban ChatGPT[00:08:28] ChatGPT can be a valuable learning tool, but underutilized[00:14:25] Schools should encourage the use of AI[00:17:40] Need to prepare students for ChatGPT in the workforceTopics Covered in This Episode:- Importance of properly using Chat GPT as a learning tool   - Majority of students using Chat GPT for writing papers   - Need for students to learn effective use of Chat GPT- Tips for educators to combat student misuse of AI in the classroom   - Former guest provides advice to educators- Rebuttal against banning Chat GPT in schools and universities   - Lack of understanding by educational institutions   - Success story of a Harvard student using Chat GPT- Limitations of AI content detectors   - Ineffectiveness in detecting Chat GPT use in papers   - Closure of OpenAI's AI detection program- Concerns about AI impact on elections and criminal activities   - OpenAI CEO's concerns on personalization and high-quality generated media   - Criminals creating chat GPT clones for scams and phishing   - Study on people's inability to differentiate between deep fake voices and real voices- Encouragement for schools and universities to embrace AI technologies   - AI models can better prepare students for future careers   - Importance of developing skills in high demandKeywords:AI detectors, schools and universities, unreliable, OpenAI, shut down, lack of accuracy, encourage, AI models, ChatGPT, real world, future careers, banning, prohibiting, academia, develop skills, high demand, embrace, curriculum, studentsSend Everyday AI and Jordan a text message. (We can't reply back unless you leave contact info)

This is the Everyday AI Show, the everyday podcast where we simplify AI and bring its power to your fingertips. Listen daily for practical advice to boost your career, business, and everyday life. Meet Firefly AI Assistant, now live in Adobe Firefly, the All In One Creative AI Studio. Just describe what you want to create and the assistant handles the rest, orchestrating multi-step workflows across Photoshop, Premiere Express, and more in one conversational interface. You direct the outcome. The assistant accelerates execution. Should schools be banning chat GPT and other AI systems?

Or should they just be letting students use them and run wild? Well, that's one of the things that we're going to be talking about today. On everyday AI, this is your daily live stream podcast and free daily newsletter, helping everyday people like you and me, not just, understand what's going on with AI, but how we can all actually use it to our advantage. Like students, I know students are using chat GPT, but before we get into that,

as a reminder, if you're joining us live, let me know what you think. Should colleges and universities and schools be banning these AI systems? Let me know. Sometimes we feature some of our favorite comments in our daily newsletter. And as a reminder, if you're joining us,

on the podcast, checking the show notes, and actually checking in the comments here of the live stream. We have a couple of other podcasts that are talking about other aspects of this. But before we dive into this and we talk about it, let's first talk about what's going on in the world of AI news. Okay. All right. So the Open AI CEO has shared his concerns about AI influencing elections. And he has a lot of concerns. So Sam Altman, he is the Open an AI CEO. So, you know, the company that has given us chat GPT. So his kind of tweet that he shared, or I think it's called X now, so his X that he shared. Does anyone know what we call tweets now? Anyways, what Sam Altman said on the platform formerly known as Twitter, he said,

I am nervous about the impact that AI is going to have on future elections, at least until everyone gets used to it. And he's talking about personalization, one-on-one persuasion. Combining that with high-quality generated media is going to be a powerful force. All right. So that is the quote from Sam Altman. So a lot to unpack there, but we're going to save that for another episode. Let's go on to our next kind of news piece for today. So criminals, yeah, criminals. They've created their own Czech GPT clones. We've talked about this in the newsletter a couple of times. but there are some kind of dark web large language models popping up and starting to kind of gain some popular popularity. So why a new Wired report looks at some of these underground large language models.

Some of the more, I guess, popular ones are worm, GBT, and fraud GBT. So why would cybercriminals want to use large language models? Well, it takes the guardrails off completely. So, you know, tools that we use like chat GPT and Bard and, you know, cloud from Anthropic, they all have guardrails built in. So not only do these kind of dark web large language models like Worm GBT and FraGBT not have guardrails, but they are explicitly being used for scamming and fishing purposes. So if you want to know more about that, make sure to check out the newsletter.

Our third piece of news. deep fake speech is getting almost too good. So a new university college London study showed that more than a quarter of people could not identify if a voice was a deep fake AI clone or if it was real. All right. So these deep fake cloned voices aren't anything new. They've been around for a long time. But some recent updates have made it almost indistinguishable to know if a voice is AI or if it's real. I'll let you guys know I am real.

I'm joining you live. So thank you all for joining me live. So as a reminder, please give me your questions, give me your thoughts or questions on this topic of today. Now that we get back to it, should colleges ban chat GPT and other AIs? So as a reminder, before we dive into that, make sure to go to your everyday AI.com. If you want to know more on those stories that we just shared about, we're going to have a lot more in our free daily newsletter. All right. Well, good morning.

And thank you for joining us. Val, thank you for joining us. You know, one of the things that I love about the live stream is having great working professionals out there who are like me trying to keep up with ChatGBT. And to learn everything. So I always like to shout them out when they join us, Nancy saying, to ban or not to ban, what would Shakespeare say? I'm not sure, but ChatGPT is actually great at writing Shakespeare.

So let's get into it. Let's get into it. And again, as you have questions, please let me know. But should colleges ban ChatGPT? You know, sometimes, guys, I draw this out. and I make you guys wait a long time to get to the end of it. But I'm going to start at the end, all right? And we're going to work our way backwards.

And I'm going to say, no, ChatsyP or schools should absolutely not ban chat GPT. Here's why. That would be very like saying that we're going to ban the internet, right? which would be an asinine thing for a college or university to do. I don't think people fully understand or realize large language models yet, right, and what they can be used for. One of the best use cases, I think, it's not actually writing, right?

We think, oh, like, chat GPT can write all of this information. I think large language models are actually, one of their best in most underutilized use cases is researching an analysis, right? So let me also just put this out here too because it's related. I'd say very few universities, colleges, high schools understand large language models, right? They just think, oh, you know, we'll put some rules. in place or maybe we'll ignore it and it'll go away. It won't. But if you aren't even sure what a large language model is, let me just quickly share that, right? And I know for some of you guys, this might be

like very repetitive and redundant. You're like, all right, I use this every day. But I think it's important we set the stage. So a large language model put very simply is essentially a collection of everything that's ever existed on the internet. Right. So some of the more powerful large language models, we would say are chat GPT4, which is, they haven't said exactly, but it's a rumor to be about 1.8 trillion parameters.

So think of that, 1.8 trillion data points, right? So when we talk about banning chat GPT and a lot of colleges, universities, professors might be saying, students are going to use it to write. What about the researching side, right? That's one of the things I love most about chat chip tea is being able to cut down on the amount of time it takes to research something. Right?

If you're trying to learn a new subject, a new topic, whether it's for school or whether it's for work, a lot of times you have to go down a Google rabbit hole. And you end up reading 10, 12, 15, 20 different web pages, right? If you can properly use a large language model like chat chb-t, it is a learning superpower. The ability, if you know how to use a large language model to accurately get the information that you want and to help you to understand something I think right now is one of the most underutilized parts of a large language model. all right so i know i know we have a lot of uh comments here and i want to get to them so kevin is saying shameless plug i will be do i will be doing a lunch and learn on what uh l lm is this coming one say yeah kevin feel free to drop that in the comments uh fabian saying since the creation of wikipedia students have taken the easy path uh to do plagiarism absolutely absolutely all right

so now that you have my hot take let's let's unravel this a little bit because I say colleges and universities absolutely should not ban chat GPT, mainly because they don't know what a large language model is or how students are using it. Let's put this out there as well. I'd say the overwhelming majority of students, yes, they are using chat GPT to write their papers. All right. Sorry if you're a university professor or something right now and you think your students aren't, they 100% are.

they're not doing the greatest job at it because also a lot of students don't know how to use a large language model, but they are. Period. So if you think your students aren't, they 100% are. We actually have some great that I left in the comments here. If you're watching the live stream, and we're going to have it in the show notes as well. But there's a great, great tip from one of our former Gus, Kelsey Beringer, from I think that was episode, 55, just on how to properly leverage AI in the classroom.

She had a great tip on what professors or teachers can do to kind of combat that. So make sure to go back and give that a listen. So let's talk about a couple of things. I'm going to share some things on the screen here. But don't worry if you're listening on the podcast. I'm going to walk you through it, talk you through it. Okay? A couple big questions that we have to talk about.

because here's why. Students are going back to school, right? Here in the U.S., some colleges are actually starting, you know, this Monday or maybe next Monday, but most colleges in universities, high schools, are going to be going back to school within a week or so. So a couple of questions. Is chat GPT in other late large language models? Are they even good enough, right? Absolutely.

Absolutely. So we shared about this in a previous daily newsletter. But a student at Harvard, she was, I think pretty forthright about it, told her teachers as much, kind of like a little experiment. But ChatGPT wrote all her papers, this Harvard student. So you might think, oh, you know, for maybe a high school paper, sure, but not hard college, right? Or a challenging university, it could never work. Absolutely it did.

So a Harvard student used chat GPT for all of her papers got a 3.3 GPA. Right. Pretty amazing. So the next thing that's important to know is AI content detectors. All right. So this has been a new thing because this is kind of like what, semester four, I guess, since Chad GPT was rolled out in November of 2022. So this is technically the fourth semester where chat GPT has been available for students.

So a lot of people think, well, hey, there's content detectors, AI plagiarism detectors. And it's going to show if students are using chat GPT, right? And there's a pretty famous story that came out a couple months ago. where a professor flunked all of his students after chat GPT falsely claimed, they wrote their papers, right? So this professor, unfortunately, didn't fully understand how large language models worked and flunked a bunch of their students saying, hey, you use chat GPT. Here's the thing with content detectors, right?

So there's these new services, platforms that are selling the software to educators, to schools and universities saying, hey, your students are using chat GPT to write their papers, buy our software, use our software, and it's going to tell you if students are using chat GPT or not. Here's what, here's what people aren't saying. Those detectors don't work. they are problem I'm not going to say

garbage but they are pretty much useless okay here's another another kind of study or article to look at

when it was looking at the effectiveness of these AI detectors so this study revealed that they only had an accuracy

of detecting AI generated text 28% of the time, right? So what's that mean? That means flipping a quarter would be entirely more accurate at determining, you know, blindly closing your eyes and saying yes or no is more accurate than these AI detectors, these AI content detectors, right? They don't work.

They don't work. All right. So let's do a quick recap. Do I think schools should ban chat GPT? Absolutely not. Is chat GPT good enough to write papers at a high level? Yes. Have schools and universities try to ban it.

Yes. Are they successful in policing chat GPT usage in large language models? Absolutely not. Because these AI content. detectors, you know, Open AI actually publicly shut theirs down two weeks ago because they said it wasn't accurate, right? So what can colleges, universities do? Right? Yes, like Celeste said here in the comments, open AI just discontinued its AI detection program. Yeah, because they said it wasn't a viable technology. So what should colleges and universities do?

What do you all think? Drop me a comment, let me know as I unpack this just a little bit more. What do you think they should do? If you're listening on the podcast, again, check the show notes. There's a way to email us, you know, add me on LinkedIn, all those things. And I want to hear from you. What do you think schools and universities should be doing? I have some thoughts here as we wrap up the show.

They should be encouraging the use. of chat GPT and other large language models and other AI models as well, right? Because we haven't even talked about things like mid-journey for AI art or runway for AI video, right? But I think colleges and universities should be encouraging these things. I think they should be teaching large language models and other AI systems. They should. Yeah, some great comments. Brian just says teach it.

Simple as that, right? I agree. Maybritt saying, I think they should combine AI with traditional learning. No reason to ban or ignore it in her personal opinion. If I was a university professor, I would be requiring my students to use chat GPT or other AI models every single day to discourage the use of AI in the classroom is like discouraging students to use the internet to learn and find more, right? Because we also think we want to be preparing students for what it's like in the real world,

right? once you enter the workforce, once you get your real job. Because guess what, y'all? In real life, in the workforce, in real jobs, the companies that are moving the most ahead, the companies that are making the biggest splashes, the companies that are the most profitable, for the most part are those companies

that are already using large language models, that are already using different AI systems, you know, to research, to create, to ideate, to analyze their competitors. So if what's working in the workforce is using and adapting to AI technology in general, why would we discourage that in the university or in classroom settings? We should be encouraging it because if we're not encouraging it, we are doing students a disservice to try to ban it or to try to use these AI content detectors

that aren't accurate, right? We even did our own internal testing a couple months ago. We had about a zero percent accuracy because it's very easy to get false positives or false negatives on these AI content detection systems. But if we're not preparing students for, what it's like in the real world, we are doing them a disservice. So our hot take here at everyday AI is this. Students should be using chat GPT. University should be encouraging it and actually working it into the curriculum. I agree with another comment here that Nancy Hall just

said. She said fighting against it causes more problems than it solves. Absolutely. Absolutely. All right. So let me know. Let me know what you think. I hope this was a fun conversation. I hope it was also a little enlightening because I think there's so much, I think there's so much gray area when it comes to large language models in AI in academia. It's, it's one of those things that I think schools and universities are just trying to hope it goes away or they think it's, it's, uh, maybe. a fad, right? Or they think that, you know, maybe chatGBT is a, is a phase in our, in our journey. It's, but it's not, right? Large language models are here to stay. They are the future of the internet. Look at what Microsoft and Google are doing for traditional search. They're phasing traditional search out with a more hybrid search that is like talking to a large language, a large language model. It is a skill, right? Prompting is a skill. Knowing when and how to use different large language models,

it's a skill. And it is a highly sought after skill right now in the workforce, right? Even just prompting, right? Prompt engineering, right? Which is a funny, funny job title. But anyways, it's one of the fastest growing fields, if you will, is prompting. engineering, right? So if we are trying to ban or to prohibit learning these new AI systems, large language models, we are taking away students' abilities to get a job in the real world, because that's what companies want. That's what I want, right? Like if someone was applying for a job, the first thing I would ask them is, how are you using AI? Right? And if a student said, oh, well, it was banned, so I didn't use it, or oh, I just, you know, use it to write my papers, right? We should be teaching students the ins and outs of every important AI platform and system that makes companies go.

All right. Thank you for joining us. I hope that you enjoyed this topic on if schools should ban chat GPT. So as a reminder, please go to your everyday AI.com. sign up for our free daily newsletter. We're going to be recapping this conversation and a lot of other things that you might have missed, right? A lot happens even on the weekends. It's Monday morning here.

So make sure you check out your everyday AI.com. If you're listening on the podcast, we have some great other resources in the show notes. So check those out. And I hope to see you back tomorrow and every day with Everyday AI. Thanks. Meet Firefly AI Assistant, now live in a. Adobe Firefly, the Allman One Creative AI Studio. Just describe what you want to create in your own words and the assistant handles the rest,

orchestrating multi-step workflows across Adobe Creative Cloud apps, including Photoshop, Premiere Express, and more in one conversational interface. You direct the outcome while the assistant accelerates execution. Stand control with the ability to step in and refine at any time. See it today at firefly.adobie.com. And that's a wrap for today's edition of Everyday AI. Thanks for joining us. If you enjoyed this episode, please subscribe and leave us a rating.

It helps keep us going. For a little more AI magic, visit Your EverydayAI.com and sign up to our daily newsletter so you don't get left behind. Go break some barriers and we'll see you next time.