Resources
Join the discussion: Ask Jordan and Jules questions on AI and education
Related Episodes: Ep 168: AI in Higher Education is Broken. How to Fix it.
Ep 252: What schools need to do now to benefit from an AI future
Upcoming Episodes: Check out the upcoming Everyday AI Livestream lineup
Connect with Jordan Wilson: LinkedIn Profile
Try Our Free AI Prompting Course: Register for our free Prime, Prompt, and Polish AI Course!
The Implications of Generative AI in Higher Education
As the world grows increasingly digitized, schools and universities grapple with maintaining academic integrity in this new landscape. AI content detectors were posited as a potential solution for this issue, but recent discussions highlight their ineffectiveness. Moreover, public examples and personal tests challenge the validity of these detectors, leading to a reexamination of how academic honesty is monitored and reinforced.
Understanding and Leveraging AI in Education
Despite these hurdles, it's crucial to view AI not as a threat, but as a tool for learning and innovation. Rather than relying on AI for answers, it should be utilized intentionally to gain new perspectives. Businesses that neglect the importance of training and innovation in AI risk becoming irrelevant as AI technologies increasingly permeate various industries.
Augmented Intelligence: The Future of Learning
Comfort with AI technologies is a necessity for future leaders. As such, incorporating AI learning tools into the classroom early is vital, tailoring our education systems towards augmented intelligence grounded in creativity and critical thinking. This crucial shift necessitates a corresponding change in assessment methods, focusing less on rote learning and more on innovative problem-solving and a comprehensive understanding of AI principles.
The Role of Generative AI
Generative AI, particularly ChatGPT, creates both opportunities and challenges in education. Understanding how to prompt and interface with these systems can enhance learning experiences, especially if introduced early on in K-12 education. Nevertheless, it is essential to understand and address inherent biases in data and the potential for overreliance on AI tools. The key lies in achieving a balance: utilizing AI to enhance learning while avoiding mental atrophy due to overuse.
Controversial Practices and Ethical Concerns
Like any pioneering field, the development of AI is not without controversy. Concern has arisen over tech giants training AI systems using YouTube subtitles without creator consent, sparking conversation about the ethical and legal implications of AI advancements.
AI Integration: The Next Steps
Despite the ethical debates, strides continue to be made towards better AI integration in education. Innovations like Google Labs' AI image generator, Imagine 3, and generative AI educational labs, herald the dawn of a new era of learning. The slow adaptation of higher education systems to these changes, however, remains a pressing issue.
Conclusion
In conclusion, as businesses and educational systems alike grapple with the rapid developments in AI technology, it's clear that generative AI skills will become essential in the very near future. How we adapt our current educational structures and methods to this wave of change will undoubtedly shape the leaders of tomorrow. Acknowledging the growth and potential of AI as a tool for learning and decision-making is an investment worth making.
Topics Covered in This Episode
1. AI Content Detectors in Education
2. Importance and Need for AI Training
3. Embracing Technology in Education
4. ChatGPT and its Impact on Higher Education
5. Ethical and Legal Issues with AI Training Practices
Podcast Transcript
Jordan Wilson [00:00:16]:
How will the future of higher education work when AI seems to be disrupting everything? Right? I think so much of what we talk about here on the Everyday AI podcast is what happens in the business world, but there's something very important that happens before that is our future leaders of tomorrow, the future people running our businesses are learning and they're growing and they're getting trained in colleges, but not all colleges and universities are on the same playing field when it comes to generative AI. So I think today's conversation about the future of generative AI in the classroom and how it'll actually work is a very important conversation to have. Alright. I'm excited for today's episode. What's going on y'all? My name is Jordan Wilson, and I'm the host of Everyday AI. This show is for you. This is a daily livestream podcast, free daily newsletter, helping us all learn generative AI so we can leverage it to grow our companies. Alright.
Jordan Wilson [00:01:09]:
So if that's you, if you're listening on the podcast, thank you for joining us. Make sure to check out your show notes and go to your everydayai.com. In today's free daily newsletter, as we do every single day, we'll be recapping the episode with a lot more information that you need as well as everything you need to stay up to date in the world of AI. Before we get into today's conversation on the future of generative AI in the classroom, let's kick it off as we do every single day by recapping what's happening in the world of AI news. So first, we were able to slip this into our newsletter yesterday, but it came after the podcast. It's worth mentioning, well, some tech giants, according to reports, were secretly using YouTube subtitles to train AI models. So tech companies are faking are are facing backlash for using YouTube titles, subtitles without creators' consent to train their AI models, raising ethical and legal questions. So according to reports, subtitles from nearly 200,000 YouTube videos across nearly 50,000 channels were used by companies such as Anthropic, NVIDIA, Apple, Salesforce, and many others.
Jordan Wilson [00:02:14]:
So the dataset was named YouTube subtitles, and it includes transcripts from educational channels such as Khan Academy, MIT, Harvard, as well as popular shows like The Late Show with Stephen Colbert and many, famous creators. So Google, the owner of YouTube, obviously claims that its use of videos for AI training is permitted under agreements with creators, though OpenAI and other big tech companies have neither confirmed or denied similar practices. And there's also some high profile YouTubers, including mister Beast, Marques Brownlee. Pew PewDie reportedly had 100 of their videos subtitles harvested without their consent or permission. So, you know, we'll have more on this in the newsletter as well. So, our next piece of AI news, Google Labs started teasing its next version of its AI image generator, Imagine 3. So there is no release date yet, but Google has, for the first time, started to consistently share, generations created with their new AI image generator. So it's very infamously, I guess, Google's generative AI tool, Gemini, faced backlash for generated contra for generating controversial images that depicted, as an example, people of color in historically inaccurate roles leading to its temporary suspension.
Jordan Wilson [00:03:32]:
So that was its old image center, Imagine 2. They fixed it, relaunched it. So we'll see what happens with Imagine 3. But from the samples, which we'll be sharing in our newsletter, it does look much better than, as an example, DALL E 3 from OpenAI, but it does look, obviously well behind leaders like Midjourney. Our our last piece of AI news for today, an OpenAI cofounder has launched an educational lab, to help people better learn AI. Speaking of today's episode. Right? So Andrej Karpathy, a prominent AI researcher and a cofounder in OpenAI, has just announced the launch of Eureka Labs, an AI and education company. So Eureka Labs aims to create an AI native learning environment by integrative by integrating generative AI with traditional teaching methods.
Jordan Wilson [00:04:18]:
So Carpathi described the venture as, quote, a new kind of school that is AI native and emphasized the scarcity of passionate and fluent subject matter experts. So, yeah, we'll have more information on that in today's newsletter. So make sure to go to your everydayai.com and check that out. Yeah. Interesting because I've been saying this for, like, a year and a half. Like, there needs to be more people focused on education in AI. And a lot of times when these, you know, cofounders of big companies like OpenAI and others leave, they believe to start another large language model company or, you know, an AI consulting firm. So this is the first time we've seen someone really focus on AI education.
Jordan Wilson [00:04:55]:
So speaking of education, that's what we're here for today. I am personally very excited, for today's show to talk about the future of generative AI in the classroom and how it will work. So, please help me welcome to the show. There we go. We have him. Jules White, who is the let let me get the correct title. The senior adviser to the chancellor on generative AI at Vanderbilt University. Jules, thank you so much for joining the Everyday AI Show.
Jules White [00:05:21]:
Yeah. Thank you for, having me, and this is an exciting topic that I'm I'm I'm excited to discuss.
Jordan Wilson [00:05:26]:
Oh, man. I'm I'm so so excited for today's show. So, you know, for our livestream audience, make sure you get your questions in now. So, Rolando joining from South Florida or Tara from Nashville, Jay, Chrissy, Douglas, everyone else, thanks for tuning in. Get your questions in now. Maybe we can get them answered. But, before we get into that, Jules, can you tell everyone just a little bit about what you do there, in your role at Vanderbilt?
Jules White [00:05:49]:
Yeah. So I'm a professor in computer science, and and what I tell everybody is November 1, 2022, if you'd stopped me on the street and you'd said, this thing called ChatGPT is gonna come out, and here's what it's gonna do, I would have told you, trust me, I will not be alive when I see that level of advance in computing, and then it came out. And so I've spent my time thinking about how do we teach people to really use this and innovate with it? So I created a prompt engineering course that's on Coursera that has almost 300,000 people. And then within Vanderbilt, I lead a lot of the university wide initiatives on generative AI, and how do we incorporate it both into our operations, but also our education.
Jordan Wilson [00:06:24]:
And let's just start at the end. Sometimes I drag people on. Right? And you already gotta listen through the AI news to get to the good stuff right here with Jules. But let's just skip straight to the end. How is this going to work, Jules? Because I feel that AI and especially in higher education, it's so polarizing. Either colleges and universities are going all in and doing a fantastic job, or they're still banning it, not teaching it, not allowing students to use it. So how is this ultimately going to pan out?
Jules White [00:06:53]:
Oh, well, there's no question how it's gonna pan out. It's gonna pan out that we realize that it's an incredible tool to, you know, teach students how to use it to support their learning and to create all kinds of completely new learning experiences. So, there's there's no question that people are gonna get past the perspective of it's a tool for cheating, and they're gonna learn how to teach people how to use it effectively. And a simple example of this is, like, people by default go in and use it, and they don't exploit the generative aspect of it. That is they don't get options. So a simple way that a student can use this, they can go say, give me 5 different ways to solve this problem and compare and contrast them. And the moment that the student sees that, it's not giving them an answer anymore. It's giving them something to think about, but then they have decide which way am I gonna choose and why.
Jordan Wilson [00:07:36]:
Yeah. And, you you know, one thing that you mentioned there is talking about new learning experiences. But I don't know. My my point, you know, obviously, as an outsider, is new learning experiences and and traditional higher education. Those 2 don't always overlap very well. Right? Like, that's that's my opinion. It seems like sometimes, change can be very slow in higher education. What are the challenges? Right? So, you know, obviously and and and we'll talk about this.
Jordan Wilson [00:08:05]:
I know that, you know, Vanderbilt is on the, you know, more on the innovative and pushing AI side forward. But what are those challenges for the rest of those, colleges and universities in the middle on being able to adapt to those new learning experiences?
Jules White [00:08:19]:
Well, I think the biggest challenge is that, like, the moment ChatGPT was turned on, it created the largest educational gap and need in the world. Basically, everybody overnight, because it's gonna affect every discipline, needed to understand how do you think and solve problems with it. And, basically, nobody knew how to do it. And so what you have is you have a very small number of people that really understand how to go and use it, effectively and how to solve problems with it, how to use it to support learning. And so the first thing that those institutions have to do is they have to close that gap and they have to get their faculty and staff up to speed on knowing how to use it themselves, but also, like, really understanding the techniques that are gonna be helpful in learning. And then it's a very interdisciplinary thing. So this is not like, the problem that you saw is a lot of places went and they said, okay. This is something our computer science or AI or data science people do.
Jules White [00:09:07]:
We'll let them do it, and we'll that's it. But, no, this is something every discipline needs to know. Right? And so you have to start thinking about every department, and that's a really big lift for a lot of places.
Jordan Wilson [00:10:09]:
I was in, you know, college 20 years ago, and even at the time, I I, like, I thought, like, is this the best way to learn? Right? If I'm being honest, I wasn't the most, you know, attentive student. You know, I was actually usually in class. I was working. But then, you know, we like, so much it seemed like so much of our final grade, I don't know, 80% or more was based on writing papers.
Jordan Wilson [00:10:49]:
And I was a journalist, so that was very easy for me. Is that aspect of higher education going to change? Right? Because when we talk about cheating, I think at least okay. Colleges, obviously, every single student out there is using ChatGPT to write their papers. Right? So is just how higher education happens going to, drastically change if it hasn't already?
Jules White [00:11:12]:
Well, I mean, I think it will drastically change. There's no question. How it will change, we don't really know yet. And I think it's gonna be, you know, completely dependent on each discipline. So, like, in computer science, like, in my senior level class, right, from the beginning, I was like, okay. Everybody should be using this. I'm gonna teach you how to generate high quality code with it. I'm gonna teach you the issues, like the fact that it can't generate something of the scale that you need.
Jules White [00:11:35]:
So you're gonna have to think about how you design things to incorporate what it can and can't do. But it also made me dramatically raise the bar in terms of the sophistication and quality of what I was expecting them to be able to do. And so you you're gonna see things like that play out, but you're gonna also see, like, let's say, in an intro class, we're gonna have to figure out, and we're still working on it. Like, do we still teach people to write code by hand? Do we teach people to read code? Do we teach people to write code assisted with this? Like, what is what and we don't know the answer to get. And there's so many disciplines. You you pick your discipline. There's gonna be the equivalence of that. But, that's what's being discovered right now.
Jules White [00:12:13]:
And, honestly, that's the opportunity for education. That's what makes my job exciting. Right? Because I I get to go and be part of that and figure that out.
Jordan Wilson [00:12:20]:
Yeah. It's a great debate. We've talked about that. Right? Even NVIDIA CEO, Jenson Wong, said, you know, maybe kids won't be coding in the future. Maybe they shouldn't. But something that you said earlier that I wanna unwrap a little bit here, Jules, is is the importance of not just prompting, but using AI across different sectors, different fields. Right? You said, oh, early on, they're like, oh, let the, you you know, the CS, the computer science people, or the IT, you know, departments handle this, but it's not really that. Right? You've had a prompting, you know, very, famous, popular prompting course.
Jordan Wilson [00:12:52]:
We've had, you know, more than 6,000, you know, people take our live prompting course, you know, dozens of Fortune 500 companies. How important is it just for for students especially just to be able to learn how to prompt, to learn how to, you know, interface with different generative AI systems? How important is that as a skill set versus things that maybe we've traditionally been taught, you know, through the decades?
Jules White [00:13:16]:
Well, I I think that it's I I don't know if I can say that, exactly how important it is, but I would say that I don't think that we can undervalue it. Like, I I don't think that I think it could be a foundational skill that every single student coming in should know in college. But the truth is is, like, I just launched a generative AI for kids course on Coursera yesterday. And I think the truth is is, like, k through 12, we should be teaching kids from the beginning how to think about and engage with generative AI in the right way, in the ways that are gonna support their learning, that they understand what it can and cannot do, but also that they start thinking about, like, when do I reach for generative AI as a tool to help me solve a problem? When do I not? And if I do reach for it, understanding those building blocks that they can put together to solve the problem. And I think it's like a foundational skill that needs to be very early. In fact, I think if k through 12, you're still teaching coding, you're nuts. Like, you should be teaching prompting. 100%.
Jules White [00:14:11]:
Way more important.
Jordan Wilson [00:14:13]:
I love that. And and speaking of of teaching. Right? And I know it's probably easier said than, said than done for both, you know, high school teachers, professors at the university level and colleges. But for those teachers and professors that can teach AI or they can use AI, where should they begin? Because I can only imagine how difficult it is to, you know, both, teach and encourage students, to to use AI, which I think is, you know, one of the most in demand skill sets that employers want. But not to be overreliant on it and to still actually learn things. How can people find that balance? How can educators find that balance?
Jules White [00:14:55]:
Well, I mean, as a shameless plug, I would say the starting point is you take my prompt engineering for chat gbt course. Once you've done that, I would say the starting point is you help people learn to use it as a tool for exploration and giving them multiple perspectives on something. This is a starting point for learning is when you go and you can say, okay. I, as a human being, am biased. I'm biased about data. I'm biased about all kinds of things. So I'm gonna use the tool to help me overcome that bias. So you take some data and you say, okay.
Jules White [00:15:22]:
Give me 3 different conflicting perspectives on this data. And then as a human being, you have to think about them and which one do you agree with and why. And so it engages your brain, engranges your thinking. I think the most basic thing is you stop asking it for answers and you start asking it for perspectives. And the moment you get multiple perspectives on a way to solve a problem, data, the email you're gonna write, anything, you as a human being have to employ your own aesthetics and critical thinking to decide which one you're gonna use and why. And often, you'll learn from the different, you know, interpretations that you get out of it. And I think that's the starting point, is you teach them to exploit the generative capacity of it.
Jordan Wilson [00:15:59]:
Let's I wanna hit this from another angle and talk a little bit about the, quote, unquote, business of of higher education. Right? Again, I think, you know, in the post COVID world, I think it made it difficult, for for teachers, for universities, for everyone. Right? You know, and having more hybrid classes and how can you still have an engaging learning environment. But how would you say generative AI affects the future of the business side of schools? Right? You know, because I think maybe people are going to be able to get good quality education that they weren't able to get before because of generative AI. What are your thoughts on that?
Jules White [00:16:39]:
Well, I think that you're always gonna have a a cohort of people who are really good self learners, and this gives them a tool that just, you know, is gonna be an incredible support to learning. Like because it fills the gap that they didn't have before, which is the ability to go and ask the question. You know, if you go to the at the end person, you can ask the question to the faculty member, you can get a follow-up. But I think most people, like, they're coming to the university to learn, but also for the environment of being there with other people to talk with faculty members who have different personalities and things. And I think you're still gonna have that. But I also think that what you're gonna have is, like, this new experience where we have all these other ways of helping people learn. Right? If you don't ask essentially, you know, we've got our material that we wanna deliver, but it's gonna be like a template. Right? And when you as a student get it, you're gonna know how to go and immediately personalize it for you and inject your learning style and all these things.
Jules White [00:17:39]:
So it's gonna be a more of a collaborative effort. And rather than me just deliver material, which, you know, you must accept in its current form. You're like, okay. Let's reshape this so it fits, you know, what I want. You know, it's like when I tell my son, I'm like, okay. Let's go have it quiz you on state capitals. And then he is like, okay. But let's inject baseball into it.
Jules White [00:17:58]:
So ask me about baseball teams, and I have to know the state capital of the state that the team is from, you know, and then things like that.
Jordan Wilson [00:18:07]:
I think I think that's a great example. Right? And just the power of personalized learning, and I think back, wow. I wish I had that when, you middle school or when I was in high school and how much easier learning would be. But do you think that maybe there is an overreliance? Is there a risk on, you know, kind of, like, quote, unquote, both sides, both both university, you know, leaders and students using AI too much. I I find myself asking that all the time, and there's always these memes on the Internet where, you know, someone enters 3 bullet points and they say, you know, make this into a long, you know, form for for my boss, and then the boss uses AI to turn it back into 3 bullet points. Is there a danger in higher education of both kind of, like, quote, unquote, students and professors being too reliant on on on generative AI?
Jules White [00:18:57]:
Well well, I think absolutely yes if they're using it the wrong way. So I think of it in terms of not artificial intelligence, but augmented intelligence that's gonna help augment and amplify our critical thinking and and problem solving skills. And I say it's like an exoskeleton for the mind that you put on. But at the same time, you don't want your mind to atrophy. So you have to use it in a way that benefits you in your learning and your thinking. So, like, that that example of giving options is is a great one. Like, if you're going and you're saying write the email for me, you stop thinking. You start copying and pasting that email.
Jules White [00:19:27]:
I see that students send emails that are like, you know, dear insert faculty member's name, and then it has this really nice thing where you write me a recommendation. If you say generate 5 possible emails and talk about why I might send 1 versus the other one, then the person's learning in the process, but they're also thinking about which one do I like and why, and you're not atrophying. And so I think the key is is that we have to teach people to use it in a way that they don't stop thinking, but they think more about the problem they're solving, not less. And a lot of people start thinking less.
Jordan Wilson [00:19:58]:
Yeah. And I obviously don't think that that's just a higher education university. That's something I think about all the time. Right? That generative AI has so much power and potential, but it also has, maybe if we're too over reliant on it, it maybe keeps us from actually learning new things and practicing our skill sets. But that's another show for another day. I wanna get into some, maybe some hot take territory here, Jules, because I know no one can predict the future, but this is something I think about all the time. Right? I've been able to learn so many things over the past couple of years with the use of generative AI. And, obviously, you know, there's more and more even things being, you know, given or offered to kids.
Jordan Wilson [00:20:37]:
Right? Microsoft Copilot putting GPT 4 0 in the classroom. Same thing with OpenAI. You have Khan Academy, all these personalized learning experiences. Are colleges and universities in the future decades in the future? Is it still gonna be the same? Are we still gonna need all these colleges and universities, or is there gonna be a certain point, you know, where in in in Andre Karpathy or, you know, we're just all, you know, getting degrees with, you know, Khan Academy? Like, what does that look like with AI?
Jules White [00:21:06]:
Yeah. Well, I think that, you know, like, I think about my son, and I'm like, right? Is somebody gonna put a VR headset on him with a virtual teacher? And he goes and does everything. And I think, no. He's gonna go to school, and he's gonna get sweaty on the playground, and he's gonna, you know, get in trouble for you know, messing around in line and having fun with his friends. And he's going to sit on the carpet and all the other things that come with that and go to pep rallies. And that's part of the experience, and he learns from everything that's part of that. And he is able, I think, to engage with the learning better because he does all that. I think if he's sitting there in front of the screen, I think university is kind of the similar way.
Jules White [00:21:40]:
I mean, like, there's so much that goes into it. But, you know, will we use these tools to enhance the learning experience that goes on a 100000 percent? Now can you already go to Coursera and get a vast number of of amazing, you know, courses and things? A 100%, you can. And you could go and give yourself a tremendous education just from your laptop, just sitting there on something like Coursera and then supporting it with generative AI. And for some people, absolutely, that will make sense. But for many people, that's not gonna make sense, and that richer experience is what they're looking for. So I think you're always gonna have, both sides to it. Now I think where it gets more interesting is when you start thinking about things like training and compliance and certification. All those types of things, I think, are the ones that are more at risk in the near term.
Jordan Wilson [00:22:26]:
Right now, is there, you know, too few qualified students coming out of colleges? Right? You see all these studies about how generative AI skills, over the last especially over the last year are exploding in demand. And it almost seems like there's not, you know, and different studies say different things, but seems like there's not enough, you you know, quote, unquote kids, young adults graduating college right now to keep up for this demand and the skill set. Do you see that as a problem? And if so, how can higher education keep up when we know sometimes change can be a little slow?
Jules White [00:23:01]:
Well, I think I think the answer is yes. I mean, ideally, every single student that's graduating from, I would say, high school has these basic skills and understands how to use generative AI because I think it's gonna be a fundamental need to be effective in most jobs. Like within a a year or 2, you're gonna really need to know this. And over and in 5 years, like, it's absolute necessity. Now at the same time right now, you can be a world expert with like a year and a half of experience in this space, right? So I would say there's also an incentive right now to be teaching your students this to put them way ahead of the curve earlier. But I think long term, everybody's gonna need to know it, and we have to engage and start teaching these fundamental skills early. I would ideally like people coming into the university who already have a lot of those fundamental skills. And then they go into their discipline like we do now, and we refine those skills in the shape of of what's needed in their discipline.
Jules White [00:23:57]:
But, how do we get like, do we need more now? Yes. Is it gonna take time? Yes. Partially, because we haven't innovated within the disciplines yet, and I think that has to happen.
Jordan Wilson [00:24:07]:
Yeah. Yeah. I I think innovation and innovating quickly is is so important. And I was kinda chuckling there because what you said there, Jules, is so true. If you have a year year and a half of experience right now, you could be, you know, viewed as a, as as a as a leader in the field. You know? I I always, like, scratch my head sometimes when a, you know, company doing a 100,000,000,000 dollars in revenue reaches out to us and, like, hey. Can you help us with prompt engineering? And I'm always like, wait. What? And then I'm like, okay.
Jordan Wilson [00:24:32]:
Maybe maybe it makes sense. But a couple of questions that I would love to get to, and this one is kind of related with what we just said. So, Douglas asking here, speaking of of rapid innovation, how do teachers in courses keep up with rapid developments? Yeah. How can you get things approved right at the university level when it can take months? And then it's like, oh, that's, you know, that's so old now. How can that work?
Jules White [00:24:56]:
Well, I think if you go back and you look at the fundamentals about like, you know, we've we've always done this as faculty. I mean, in computer science, right? Everything's changing continuously in computer science. It always has been. Right? And so my job as an educator has always been to distill what's happening down into what are the core principles that are more timeless. And so when I created my prompt engineering for JiGPT course, everybody said, well, isn't that all gonna change in 2 months? And then the answer is no. It didn't. Those fundamental, you know, principles and how you think about solving problems with generative AI have not changed. And we need to focus on those things, and faculty need to focus on educating them on the themselves on those things.
Jules White [00:25:33]:
And universities need to think about how do they take those core principles and rethink learning, rethink how they deliver learning, create content, all of those things based on those core principles and capabilities. And then from there, you know, yeah, the tools will change. Like, you'll have, you know, Claude 35. You'll have 4 o. Those things are sort of irrelevant. It's like swapping different engines in the car. But if everybody already knows how to drive and they know what a car is, it doesn't matter if you swap engines. It's just like you get some more performance out of it or fuel efficiency or whatever it is.
Jordan Wilson [00:26:04]:
That's a great analogy. Yes. Swapping the engine out. Another great question here, from Tara. So thanks for joining in. Saying, have you seen faculty shifting their assessments toward things AI can't do? If so, what creative ways do you endorse?
Jules White [00:26:20]:
Yeah. That's that's a great question. So the answer is yes. I mean, I've seen faculty shifting their their assessments in all kinds of directions. I mean, one is that just faculty expecting so much more out of students. Saying that, like, my goal at the end of the day is to help you be able to build and create amazing things. I'm just gonna scale up the size of what I'm gonna look for, but also the quality of what I'm gonna expect you to be able to do and how how do you be able to integrate all these different parts together. Because the moment you scale up, what you tend to have to do is you have to think about how do you integrate and make all these things cohesive.
Jules White [00:26:52]:
And, so that's one area that I've seen a lot across disciplines. I've also think faculty get really creative with it. So there's a faculty member in French at Vanderbilt, and she was having students generate dialogues between historical French figures, and then they would have to go and discuss the accuracy of the language for that historical time period and the and what was happening and, like, did it get it right and how and are there subtle issues in the language. And, like, that requires students to really engage. It's fun because it's creative and interesting, but it also requires them to creatively engage with it. And then I've also seen, you know, people just go in and say, well, okay. We're gonna, you know, talk about things in class and other things, and we're gonna bring discussions from class into assessments. And now if the student wasn't there in class engaged and listening, they're not gonna have the appropriate contacts to even feed into the generative AI to begin to answer the question.
Jules White [00:27:53]:
So there's all kinds of creative ways, to go about it. But I think at the end of the day, our goal should be to figure out how do we use these tools to make students think even more deeply, you know, rather than worry about, like, are they cheating? I mean, like, there's a straight Stanford study that came out that basically said, we monitor cheating before ChatGPT. We've been monitoring after ChatGPT. There hasn't been a change. Just the way that students who do wanna cheat decide to cheat. And I think it's also a little bit of a disservice because, like, people pretend like all students are cheating, and then they're not.
Jordan Wilson [00:28:24]:
I'm gonna open myself up for one here, but since we're talking about it, I have hot takes on this, so I wanna know yours. Should universities be using these AI detectors? It seems like that's the the crutch, you know, so many universities have been using over the last, like, 2 2 to 3 years. Should they be using those? Are those actually real?
Jules White [00:28:44]:
A 100 my perspective my my personal opinion is a 100% no. I think there's snake oil. And when you have OpenAI produced a detector and they took it off off the market because they said, hey. It's not it doesn't work. We had an AI detector turned on unilaterally by the vendor on campus at Vanderbilt, and it started giving scores on assignments. And we really quickly formed a a team to go and assess, are we gonna do this? And basically what we asked is like, can you give us the data used to verify and come up with these numbers? They said no. We said, can you give just details? They said no. It's proprietary.
Jules White [00:29:16]:
And then what we did is the next logical step is we took a couple of our thesis that were done long before any of this fed them in, and immediately, they got flagged. And so for us, we looked and we said, look. This is gonna turn get thousands of students sent to the honor council. You know, thousands of assignments would be flagged. And what we were told was, well, we would you know, it it the the vendor told us, well, we would expect the faculty member to look at the score, but then assess all the characteristics and make a decision, except that AI detectors tend to be a number. It's like you go into court, and it says 83% guilty. Alright. He's guilty, but there's no evidence.
Jules White [00:29:56]:
Right? Because it's all baked into some model that you can't see. You know, at least plagiarism, you can say, well, here's what they wrote and here's the paragraph, and then you can decide. So I think AI, AI detectors, 1, it's like a a Band Aid for not wanting to change, but 2, it's just a tremendous disservice to students.
Jordan Wilson [00:30:15]:
I'm glad I'm I'm glad we agreed on that a 100%. Literally, I've I've been saying this for a year, my exact, phrases, snake oil. Yeah. The whole OpenAI thing, they took theirs down because it showed it was only 26% effective, which is less than flipping a coin. So, yeah, if if you are listening out there, I'll be more direct. AI content detectors do not work. There is no such thing. We've personally done tests and busted them all very easily.
Jordan Wilson [00:30:41]:
So I do wanna get to one more question here before we start to wrap this up, Jules, because, you know, you are in charge of right now, you know, kind of educating, the the next generation of leaders there at, at Vanderbilt. But I feel a lot of the lessons that you're learning can be applied to the the business world, the rest of us. Right? So as someone that is constantly having to learn and innovate and, you know, understand and implement new concepts, what have you learned in the educational setting that maybe is very applicable to the business world and what we should all be focusing
Jules White [00:31:15]:
on? Yeah. Well, I I would say the first thing is, like, you really need to teach how to use it intentionally. So, you know, that use it to gain perspectives, not use it to give you answers because you don't want your workforce checking out and becoming, like you said, overreliant or not checking its work. So I think that's one. I think the second thing is not to underestimate the magnitude of this change. I mean, just to go back to what I said at the beginning, I didn't believe I would live to see something like this. And, you know, you see so much discussion of what is your return on investment, and that's thinking of it the wrong way. Like, it's gonna be some incremental 20%, 50%.
Jules White [00:31:48]:
This is one of those technologies that creates completely new things that we haven't thought of before that create tremendous growth in an area and then make some area totally obsolete or irrelevant. And so I say the ROI is risk of its irrelevance. And so you need to determine what is your risk of irrelevance. And if you're not investing in training people, and that is critical as you're training people, you're supporting experimentation and creativity in this, then your risk of irrelevance is huge. So I I think those are sort of the 2 things. And if we go back to risk of irrelevance, I mean, we wouldn't be talking about our user university is gonna be here if this wasn't a fundamentally destabilizing, you know, capability. I think that they're the ones and I I should probably modify in my answer. The universities that engage and embrace and innovate with this will absolutely be here.
Jules White [00:32:38]:
Those that do not will be maybe not in existence anymore.
Jordan Wilson [00:32:43]:
So we've we've talked about a a lot in today's episode, Jewel, from, you know, the future of of learning, how we can take lessons from the classroom to the boardroom, examples of what you and your peers are doing at Vanderbilt and AI content detectors. We've covered so much. But maybe as we wrap, what's the one most important takeaway that you want our audience, to hear and apply today when it comes to the future of generative AI in the classroom?
Jules White [00:33:10]:
You know, I think I think the the future is augmented intelligence where you teach somebody how to use it in the right way to augment and amplify their critical thinking and creativity. But it's fundamentally rooted in their critical thinking and creativity, and you wanna really, really build on that aspect of it. So a lot of the things that we do right now are based on, like, the way technology worked before, the way we assessed before, And we're gonna have to reshape that. And the fundamental thing that we're gonna have to think about is creativity and critical thinking. It's fundamentally gonna be the most important piece. And then all of this technology is gonna be how do we harness that, amplify it, magnify it, give them new outlets for creativity.
Jordan Wilson [00:33:49]:
I think those are such good insights, today. So, Jules White, thank you so much for joining the Everyday AI Show to help us really make some some, meaning of all this mess of what's happening in the world of AI in education. We really appreciate your time and your insights.
Jules White [00:34:08]:
Yeah. Thank you so much for having me.
Jordan Wilson [00:34:10]:
Alright. And, hey, as a reminder, that was a lot. It was like a course and a half on AI and education. So if you haven't already, please make sure to go to your everydayai.com. Sign up for the free daily newsletter. Myself, a human, I'm gonna go sit down, relisten to this, and write some of the most important takeaways and more that we didn't get to. If this was helpful, please consider sharing this with your network or, you know, subscribing and leaving us a review on Spotify or Apple. So thank you for tuning in.
Jordan Wilson [00:34:38]:
We hope to see you back tomorrow and every day for more everyday AI. Thanks y'all.