Disambiguation: Generative AI for Higher Education and the Future of Work - TRANSCRIPT
Michael Fauscette
Welcome to disambiguation. I'm your host, Michael Fauscette. Each week we interview experts in AI generative AI and business automation to help business leaders understand how to use these tools for the biggest business impact. And our show today we look at generative AI in education and its impact on the future of work. I'm joined by Keeley Metz. Haley, welcome. Oh, thank you for having me. Could you just give us tell us a little bit about yourself first, before we get started? Sure.
Keeley Meetze
I'm the founder and CEO of The Keeley Company, which is a company that focuses on international marketing solutions, as well as AI and education, integration. And I am currently a master student for psychology, culture and evolution, Master science at Brunel University London.
Michael Fauscette
Nice. And welcome from London. I know you've only been there for a little while. So that's one of my favorite places. So alright, well, so let's get into this a little bit. And I, you know, there are a couple of topics that we were going to talk about. But let's start with education. Because I, there's been so much noise around education, particularly in the US, but I'm sure everywhere else, too. So, you know, from your perspective, what are the practical applications of generative AI and applications in education? And how can it be used to make the learning experience better.
Keeley Meetze
So when I think about practical applications, it's really hard because AI moves really fast. So what I like to say is just the way to know how to use it is just to use it, because it's updating all the time, the way we look at it in my company is by using it to expand on knowledge networks. So we know that students learn based on what they already know. And the idea of mastery learning comes from that is when you have gaps in your knowledge, it's really hard to fill those gaps with more knowledge building on those foundations. So what we look at with affiliate company is how can we use the AI to assess what our students already know and understand? And then how can we adjust our assignments, again, using AI to address those gaps and more utilize the class time for more effective learning. And that's kind of how we're approaching it with us. But like I said, everything changes so quickly. So there's a lot of different it depends on what you teach and what you're trying to do. But I think the most important thing you should ask yourself first is like, what am I trying to do? What is the point of this assignment? And then how can I use AI to do that the current AI?
Michael Fauscette
I know, one of the one of the articles I read not long ago was talking about the way you could use use AI or use chat bots basically as like individualized tutors. For, you know, for your particular learning. Is that Is that something you've seen? I mean, is that a discussion that you've heard?
Keeley Meetze
For sure. Yeah. I mean, you personalized and individualized learning is one of the great things that AI can do for us. One thing I think people have, when I'm talking about AI, so let me just say I'm talking about large language models usually. So AI can like be abroad thing, but I what I focus on is online chat to be at the being chat, cloud with me topic and Google's AI, large language model that came out Bard. But those things can be used, and they have their own knowledge. But at the end of the day, they aren't understanding they're not intelligent in that way. So to trust the outcome of them is one thing that like people don't really understand is that they hallucinate, they, some of them will tend to be T used to not be able to search the internet, but just this past month, they included that and they can scan documents and things now so that's where I'm like, you know, things are always moving. But yeah, being able to develop new analogies develop new examples about things that maybe the teacher doesn't have knowledge about, but the student has a lot of that foundational knowledge about is one of the main ways that I think I advocate is the best way to use it. That makes sense. There's, there's so many things that students like in this day and age like, for example, I know nothing about anime, I'm not an anime person. But there are some people who are really involved in anime. And if you could give them an example from anime, somebody that they know a lot about to explain a concept that maybe they wouldn't understand in another context. That's an amazing way for these largely because we use
Michael Fauscette
that makes sense. So put it in the context of whatever that students area of expertise and knowledge are or, you know, likes and dislikes, I guess, anyway.
Keeley Meetze
Yeah, exactly. I think that's strong. That's a really strong thing. I like what you said about like that student's individual knowledge and expertise because I think that's one thing that can get really lost in traditional education is just like treating every state Then you know, to do one task or train one different way, when what's great about humans, and what's great about individuals is that we all have our own nation, we all have our own expertise. And the way that we can foster and build on that with our students is by making our class concepts applicable to things that they're interested in, in the things that they're going to do. So I think that the large language models can really help facilitate that learning.
Michael Fauscette
Yeah, I mean, it seems like that, that is almost a way to go out of this kind of industrial model that we've had, especially in the US around education, we sort of took it to the lowest common denominator and teach across rather than teaching down into the areas of expertise and knowledge, then that, makes a lot of sense. So we're talking about benefits and challenges to me, what have you seen from, you know, from the benefit and the challenge of incorporating it into the classroom? And how can how do you help teachers overcome some of those challenges?
Keeley Meetze
For sure, so I think the benefits, like I said, have to do with applications are immense. And they, I don't even know all the benefits that could happen, because I haven't even used all the cases. So are all the different, you know, functions. And its paths that are possible. So I think for each teacher, It's same with the students, it's individual. So the things that you could use these large language models for, you're only going to be able to discover if you practice and try to use these large language models. So I think that's the benefit is that it's so easy to tailor and to adjust to what you're trying to do. But at the same time, that also leads into the challenge, which is just like, there's so much unknown, and it's so much just like you really need to have, as a teacher, you don't really have a lot of extra time to do all this extra study and all this extra adjusting of your assignments and things like that. So I think that's really where the challenge comes in is just like, there's so much that you can do, it's just picking out what you need to do and how you need to use it. So that's what my company like aims to do is so I partner with these teachers to make it a little bit easier to kind of give them a more narrow focus on like, Okay, here's what you're trying to do here, how can we work together to like plan assignments, and that will utilize these AI. So I think that is the challenge. But I also think that's a fun challenge, because it also allows the teacher some learning, and the students will be able to tell you a use cases that you didn't even think of. So I think that is another point to advocate for encouraging students to use these models and encouraging students to use this AI because they're going to come up with innovative ways that then will be good for you as the teacher to use as well. So I think those are the challenges. I think most of the challenges also come from trying to fight it. I feel like a lot of teachers are scared of it, understandably. So you see all these like headlines like students are divided homework, students are cheating. You know, I'm a master student right now. And the teachers are so freaked out by plagiarism and all of these things. And I think that's where people run into problems is because they don't really understand how it works. And then they don't understand how we can make it work for us. And I think that's kind of where that where we're at right now, with everything being so new. Yeah,
Michael Fauscette
I mean, I see that in business, too. I think, you know, a lot of people that, that push back on it, they don't really understand even why they're pushing back there, they're just kind of follow the crowd when it goes to that. And, and yet, if they, you know, they start to apply it, people come around pretty quickly, like, I know, it was talking to a professor who had started to use chat GPT themselves for different parts of what they do, like, you know, grading and, and different, creating different kinds of content. And they sort of flipped completely when they when they got to that point, because they all of a sudden understood why this was beneficial to them. And I mean, do you see teachers using this for as a part of their process of, of teaching is that do you think that's one of the ways that we're going to get around some of this pushback?
Keeley Meetze
Maybe, for sure, that's another thing that I like, tell teachers, I'm like, it will make your life so much easier. Like I can come up with a whole syllabus, like just tell it what you what your cover, maybe put an old syllabus in there or just it's a conversational model. So you just talked about use it as a TA use it as a secretary use it to, you know, draft emails, but again, it is going it does require interaction with the user. It's not an autonomous agent like, and it's not, you know, completely smart. It doesn't have the expertise that these teachers have. So I think when teachers start to use it, and they start to see, like you said, that professor did the benefits of them, they'll be able to more clearly understand the benefits that it will provide for their students as well. So I mean, like my favorite things are coming up with examples that you wouldn't have been able to come up with before. One of my favorite ways to get around the essay thing is tell the students use the chat UBT to write the essay had the students edit the essay, which requires them reading the essay, understanding where the essay went wrong, and then editing that and then have them rewrite So using chalk up, you know, you can expand, you can scaffold it in so many different ways. And it just takes practice, and it just takes use. And each year, you're gonna get better. But I think the one thing you shouldn't be doing is saying, don't use these at all. If you use these, you're going to be kicked out of school, there's no way to detect them. There's zero detectors out there that work, it disproportionately affects students that don't speak English. There's a lot a lot of problems that come along with this whole trying to avoid it. I say, don't fight the calculator, the calculator came and we just changed how we taught math and we changed what was important in math. I think Chuck EBT, and large language models are going to be the same way, you're not going to avoid uses, they're gonna be every job, you might as well prepare students to use them and also make your life easier as well, for sure. Yeah,
Michael Fauscette
that makes sense, right? I mean, I know, you know, as a writer, myself, sometimes staring at a blank page, it's just hard, right? So but if you use the, if you can use a chat bot a chat, you know, GPT what, whichever large language model you choose, just to give you a draft star, even an outline, right? I mean, just something to start working with sometimes can accelerate that your learning and your research into whatever it is you're trying to do. So that yeah, that that actually, I mean, I think that's great advice. But and, and from a from a teacher perspective, figuring out ways to incorporate it into your process then helps you be able to explain and teach much more effectively, I would think, yeah, exactly.
Keeley Meetze
And I think teachers will be surprised at what this thing can do. Like, I remember the first time I use chat, GBT, three, I don't even think it was 3.5. Back in January, I remember just being like, every single day, you know, like, and it's not going to be perfect, like you said, like, you can have it help you brainstorm, but nothing at the end of the day is going to be like the product that you can come up with, you're the expert, you're, you, as a teacher know what your bullet points are trying to get across, I just think just use it like a tool, you know, just like spellcheck, just like any of these other tools. If you ask it to do a certain thing, and it doesn't do it right, then you know, and if you ask her to do a certain thing, and it does, then you're happy because it did what you wanted it to do. So playing with it, and like learning from it, I think is you know, teaching by learning, all of the things, all of those things are super important. And I think that, yeah, the only way, the only way to know is to try right. So that's, and there's no risk in it as well. So of course,
Michael Fauscette
we get in there trying, right? So the one of the things, I mean, we're gonna move on to talk a little bit about future work, too. But this kind of brings me to this question around. You know, are we shortchanging students today? Because we're not teaching them? And so that doesn't prepare them for the future? Because they aren't going to have to use these tools at work? Most likely, anyway. Yeah.
Keeley Meetze
Yeah, I think 100% I think if you're trying to avoid it, or tell students that they can't use it, then when they get into the workplace, where they, they're against people who or they're competing for a job with people who have been using them, and have had this like training and experience, they're going to be disproportionately affected. Because every company I said, AI in the future of job report by the World Economic Forum, 85% of companies have admitted that they are using them, or they plan to integrate them within the next year, these AI tools. So and one thing I've been looking at is the history of education, especially in America and traditional education, you know, and I'm talking primary secondary education was this factory farming kind of, you know, point of view where we like just teach kids, it's like a babysitting system at the same time nine to five, so the, you know, keeping them on a schedule, and training them, giving them foundational critical knowledge, things that they will use and apply throughout their lives. And in higher education, AI is just going to be another one of those foundational things. So and critical thinking and being able to understand what the AI does, and evaluate the outputs is going to be even more of a foundational knowledge that students will need. And then when they get into higher education, you know, becomes more niche, and you are more prepared, you're preparing more for the job market, but at the same time, even if you're a marketing major, it doesn't mean you're going to go into marketing. So you're really just going to learn these like foundational skills, like I said, and these technical skills and technological literacy is like the number one thing that companies are going to be looking for. And by being scared of the technology and telling students start to use it. It's not applicable to what they're actually going to face in the real world. So yeah, I think you're exactly right about that.
Michael Fauscette
talk a little bit more about future work in general then. So what do you think? What are the other implications of using generative AI and its effect on what you know what works gonna look like? I mean, no, there's a lot of, of, you know, a lot of fear around Oh, it's gonna take your job and you know, all the companies read just going to run on AI and the rest of us will have time off or something, I guess. I mean, how do you see this? What do you how do you see the future work with generative AI, and obviously, newer things are going to continue to evolve in that, but how does that impact the future work in your, in your world.
Keeley Meetze
So that was, the first thing I was gonna say is that don't be scared that it's going to take all the jobs because there will be a net creation of jobs, it's just jobs are going to look a little bit different. And honestly, what these AIs can do is going to probably, hopefully, take away some of the boring parts of our jobs, a lot of the secretarial work, and even I mean, people were like, oh, technology's taking over, it's gonna change all of these things. But, you know, even today, like most jobs use Excel, like most, you know, white collar jobs are using Excel or like spreadsheets, and they're analyzing data that way. So that's what this AI does is a lot of that data analysts analyzation. And, like I said, the most critical thing is going to be teaching students critical thinking skills, evaluating, you know, products that the AI will come up with, and in every single job, it's going to look different, but a lot of companies are going to be investing in this. And so I think, you know, one of the happy things that I like to think about is it's going to make human work human again. So there are things that humans can do that AI you can't trade it to do happiness, creativity, soft skills, you know, personal interaction, understanding emotion. And those are the things that make us human. And I think we kind of lost that a little bit when we went into this, like industrial revolution, just like humans became machines, basically. And I wouldn't say I also study evolution. So that's what I'm here doing at Brunel is studying evolutionary psychology. And I think from that perspective, you know, getting back to what makes us human, and focusing on these niches and focusing on that creativity is really where jobs are gonna change. And I think that's for the better, honestly.
Michael Fauscette
Yeah, I like that. I like the idea of, and have used the term quite a bit human machine collaboration, I think that makes a lot of sense to me, because it is, it is a collaboration, it's an augmentation of what you can do. And there are things that the AI can do that, take it off your plate, and then there are things that the AI isn't capable of, and, and maybe someday will be, but it certainly didn't today. Yeah. that I think makes a lot of sense. And you mentioned the, the empathy piece of it. I mean, that's one of the biggest things in customer service support, how do you empathize and show that empathy. And, you know, artificial empathy just doesn't work very well yet.
Keeley Meetze
And people can tell you no, and also like, that leads me to another point of like, the future of work is globalized. Like we are connected more than ever. And honestly, like, the most important thing is learning how to globally cooperate, cooperate with our fellow humans. And I actually had an interesting conversation with some of my cohort the other day where they were saying, we were talking about something else, but the need for global cooperation and the need for understanding you know, each other and different people from different cultures, people that look different from you. And one of the guys that something Buddy, he was like, We need a common enemy that isn't human, we need aliens, you know, I was like, I never, that's such a boyfriend thing. Like, that's something I've never thought of before in my life. But I think that not saying that AI will be this alien that humanity needs to come up against, hopefully. But I do think that seeing these machines as something that can help us and also show us the value of uniting and working with other humans, I think what other will be another positive outcome that comes from this? Hopefully, if we do it, right,
Michael Fauscette
yeah. I mean, it's just another member of your team. Right. So you're exactly add the AI to your team. So you mentioned a little bit about, you know, some of the negatives. And I know one of the things that people talk about a lot are ethical considerations around AI. So what do you think, what ethical considerations should we take into account when we're using generative AI in education, and also just in the workplace in general?
Keeley Meetze
So one of the main things I've read about ethical considerations, the main one is that, you know, when you have software developed by and created by a team, it's going to reflect the biases of that team, because it has human interaction human have by humans have bias. So I think in the same way that you kind of have to monitor student work for maybe some uneducated or ignorant biases, you might have to do the same with the AI output. So I don't think much will change there. That the thing that would be positive, though, is that you can use that to explain biases in a way that's different that isn't like blaming a certain human you know, you're like looking at why these things are biases and why these things exist. And the more that we integrate it and use it across, you know, foundational learning More students are going to be using and more students are going to be pointing out, hey, this isn't exactly right. So I think the ethical considerations come from just biases that have always existed with humans. And then also, you know, people talk about copyright. And they talk about plagiarism, and they talk about things like that. But it's like, these models are trained on that data, but they won't, a lot of them have been told, like you cannot, like if you if you ask him to give me a paragraph from a book, they can't, you know, produce that. So I think it can be used as like a tool to kind of like see the difference between player or player, plagiarism and, and copyright and those things like that. Also, I think a lot of the ethical things come from where they're getting that data, generating data, but an idea that I had, actually recently that I like, woke up in the middle of the night and like, put in my notes app was that, you know, we can come in partnership with creating this data, like when we're with our students, and like, you know, that's something that maybe other companies should like look into is like, you know, data is going to run out. And that's going to be the big, the big next problem that this AI is going to have to face is that they're going to run out of data. So thinking of innovative and creative ways to create new ethical data, I think there's something that could be really implemented in the classroom between teachers and students.
Michael Fauscette
I think, obviously, that a lot of bias just exist in the data itself. And so, you know, helping recognize what that is, and coming up with ways to train it more effectively, I guess. I know, one of the methods of grounding, and which obviously helps with, you know, with hallucinations, but also just, you know, having that capability to pull real time data and to give more context to what you're working on. Could be one of those ways to overcome some of those considerations. I think, too. So. Yeah, that makes that makes sense. So I, you know, we talked a little bit about human machine collaboration already. But But how do you, you know, how do you get a balance between automation and human involvement when we're talking about generative AI? I mean, we talked a little bit about the blank page problem. And I guess that's one of the ways it can help you jumpstart some things. But what else do you see? I mean, how, what's the right balance between automation and human involvement? Do you think?
Keeley Meetze
I think it, like I said, really depends on the human and what the human is trying to do, because AI does not have its own goals. So that's like, the first thing is that that's why it's not an autonomous agent. Because I mean, it can't, at the moment, do anything that you don't tell it to do already. So that's also another argument against the plagiarism thing is, like, even if a student is using the AI to write the essay, the AI is not writing the essay on its own, it's using input from the student. So that is the students work, which I mean, there's, it can get more complicated than that. But at the end of the day, right now, where we are right now, it requires human input, and it requires some kind of interaction. And I think that it's like using a TA or using, even just like an assistant that doesn't have the expertise that you do, it's just using that like in that way. So that's where the human interaction is gonna come in, it's gonna give you an output, you're gonna have to be the one to evaluate that output. And, you know, evaluate if it's right or wrong. And like I said, I you can change like that into assignments, you can have students say, if it's right or wrong, address the student's knowledge about it. So I think all of it has to be human AI interaction right now. Like, it just depends on how you the human want to use it. So
Michael Fauscette
yeah, I mean, you mentioned some of the new features in in open API's tool, check GPT. And a couple of things that they've done in and I see this and other tools like perplexity, for example, does this too, but gives you citations when it when you do research, it pulls up the you know, it will generate something for you to read. But it also then tags that with here are the five URLs of the things that you know, that I searched to come up with this, which to me seems like a learning opportunity, really. And I use it that way myself, right. I'm doing research, I have it generate something and then I follow those links, because it's certainly more effective than just doing a simple Google search or whatever.
Keeley Meetze
Definitely, I use it for all of my research. Like that's like if I like come up with a new research idea or come up with a new problem as the first thing I do is I talked to the Chatbot about it. I prefer because the chat up to using the internet is so new. I haven't worked with that one as much. If I want to search something on the internet and do some research, I use Bing more than anything barred. Honestly, I was kind of surprised. Like, it's not that good. Like it's actually quite bad. And I was surprised because when I first found out about AI was from like a Google lab with like this whistle blower all of these things. So I think Google definitely has some things under lock and key that they just haven't released to the public. That is much better, of course, but what they have publicly released is like not that good. So if I'm using it for research I use Bing. And they don't, it does provide links, but also with chat to be to you. And both, I mean, thing is better, because you can just directly click on the link and then evaluate the source yourself, which had to be to win, it provides citations. Before this was before it was connected to the internet, it was providing me citations with that didn't exist, which is also a learning opportunity, because then students will need to go through and they actually have to do the work themselves. And if they turn something into you, that isn't a citation, you know, that they didn't do that extra step. So I think, yeah, it's definitely, you can use it in all different ways. That's just another one that I have this because I do research. So yeah, it's definitely Yeah,
Michael Fauscette
same for me. And I think the, you know, having the real time connection, so you can actually, you know, it does ground the model much more effectively. And you starting to see that and more and more kinds of applications and, like, I use perplexity a lot. Now, because of the same thing. It has a mode where it basically uses, you know, rag retrieval, augmented generation so that it can ground itself and make it give more context, which, which I think is good. And I noticed that in, in chat GPT now that it has, you know, live internet connection. I mean, it had it in the plugins, but you had to do it manually, which was kind of tricky, I think, unless you knew that right. But yeah, but now at least it does have that available, and it's much, much easier to use.
Keeley Meetze
And also, I think it's better for students, because like, you know, I remember when I was like an early researcher, when I was a student, and the teacher told me, okay, here's your first research project, go find some good sources. And even if they gave a lesson on how to find sources, it's like, I stare at my computer. And I'm like, I don't even know where to start, at least this will give students a place to start. And we'll give them sources that are actually valid, and then they can trace and see which sources are actually good for whether they're trying to do if they're valid, what they're trying to do, and and that critical thinking that being able to evaluate is such a important thing that I think a lot of teachers, not teachers, sorry, a lot of education itself has lost a little bit. You know, they're just trying to teach these foundational things like, you know, of course, the foundational math, science, history, English or whatever. But within those days, I feel like a lot of it has kind of lost that. Okay, but why? So that's one thing I wrote a paper on was like this elaborative interrogation, using the AI is asking why getting to this deeper understanding, and then it builds out, like I said, those knowledge networks, and that is what's happening
Michael Fauscette
to me. So you have kind of a unique position right now, because you've gone back and you're in in school, but then you were also working in that market and working with education institutions to implement this as well. So what advice do you have as a student to students what what advice would you give them to figure this out, and how they could, you know, learn and apply better?
Keeley Meetze
Same advice I give to teachers use it, like use it, I would say I, I spoke a little bit with you about this before we started Michael, briefly, like my school that I'm currently at. I love burnouts, right. But they have really hounded and scared a lot of the students about this idea of plagiarism using these AI tools. And there isn't any AI detection, like I said, out there right now. So I think they kind of have taken the wrong approach to that. And I think it has really freaked a lot of students out and students are just like, I'm not using it at all, because I'm scared that I'm going to use it in a wrong way. And even I working in this space was feeling a little bit like, I'm scared, I'm going to accidentally plagiarize. And I know that that's not like a thing that's ever even crossed my mind before. So I think even if you're not using it in school, use it for things that are fun for you use it to journal use it to write blogs, use it to I mean, I'm saying these are things that I do, you know, use it, don't make quizzes for yourself, play with it and see what it can do. I think it like I said, with the mastery learning even me if I don't understand like a concept, I use the AI to give me different examples of that concept. And that has really upped my learning. So it's up to the user, it's up to the individual student how you learn best, but if you know something works for you use the AI to create that, like you use it for your own studying for sure. Yeah,
Michael Fauscette
I mean, even just from my perspective, as a researcher, I use it all the time for exactly that, right? It helps me if I have an idea, then I can get more context around that idea, then I can develop that and help that evolve better just by virtue of having that extra amount of information and context around it. So that I think makes a lot of sense. Well, so that we're out of time, but I you know, first of all, thank you for spending the time I really appreciate it and thank you for having me. There. It's I think this is a really interesting subject particularly around how we, you know, get students up to the right speed and help them be more productive in their, in their career student but also in their in their future job, right. So one of the things I'd like to do at the end is, is just get a recommendation from you. So could you give you have somebody that's, you know, a thought leader and author, some somebody that has influenced you around this subject of generative AI?
Keeley Meetze
Yeah. 100% Ethan Mollick and his wife Lilach? I don't know if I'm saying her name, right. But both they are like a power couple of, I would say this century, they are really focused on pedagogy, which is another word I can't say, right. And they're at the Wharton School of Business in Pennsylvania. And they are the pioneers of using it utilizing AI in education. Ethan has a great blog that I read all the time. Anytime he posts I have like notifications on my phone. His Twitter is amazing for use cases, follow him on Twitter. Those two people are the Founding Fathers, I think of AI and education. So I would definitely recommend any of their work. They've written some amazing research papers. They have an amazing YouTube. What is it YouTube should not channel but a series YouTube series where they talk about integrating AI and education. And they're working on it every single day. So they're at the forefront. So definitely those two. Great,
Michael Fauscette
yeah, that's great. Thanks. I really appreciate it. And again, thank you for joining. I appreciate you taking time out of your extremely busy schedule as both a student and an executive.
Keeley Meetze
So I wear many hats. It's true. Thank you. Thank you.
Michael Fauscette
And that's the show for this week. Thank you all for joining remember to hit that subscribe button. And for more on AI and other software, research reports and post check out the area on research.com/blog and slash research reports. And don't forget to join us next week. I'm Michael Fauscette. And this is the disambiguation podcast.
