Name: Brainstorm AI Singapore 2024: AI in education—fighting cheating and preparing smarter students
Uploaded: 2024-07-31
Duration: 19 min 29 s
Description: Khairul ANWAR, CEO and Founder, Pandai Tim BALDWIN, Provost, Mohamed bin Zayed University of Artificial Intelligence Sonita JEYAPATHY, Co-Director, Center for Pro Bono & Clinical Legal Education, National University of Singapore Moderator: Jeremy Kahn, FORTUNE

Khairul ANWAR, CEO and Founder, Pandai Tim BALDWIN, Provost, Mohamed bin Zayed University of Artificial Intelligence Sonita JEYAPATHY, Co-Director, Center for Pro Bono & Clinical Legal Education, National University of Singapore Moderator: Jeremy Kahn, FORTUNE

Brainstorm AI Singapore 2024: AI in education—fighting cheating and preparing smarter students

Conferences

Khairul ANWAR,  Pandai, Tim BALDWIN, Mohamed bin Zayed University of Artificial Intelligence, Sonita JEYAPATHY, National University of Singapore, Jeremy Kahn, FORTUNE

Sunita, let's start with you. Um at, at the National University of Singapore, your law school. How are you finding students are using A I? And how are you preventing them from misusing the technology? Why you need to do certain things? Why you need to learn certain critical skills with or without the assistance uh That all helps to solidify for the student whether or not to use this particular tool. So yes, there will be students who will, who if they are not better aware, they will try to use it, but sometimes that's not really their fault as well. They need to be made to understand why certain tools should be used perhaps at a later time after they've acquired certain critical skills without the tools. And do you have a specific policy about use of chat bots? Um So we uh there is anti plagiarism obviously uh in in any tertiary institution. And the knee jerk reaction when Chat GPT came out was um what are we going to as educators? What are we gonna do? Are all our students gonna be plagiarizing things from Chat G BT? Um But then after a while things settled down a little bit more and we realized that we could leverage it as well. I mean, there was a lot of experimentation but it could be a tool that could assist students, it could be a tool that could assist educators as well. So we're still in the throes of experimentation, but students basically know they can use it, they need to be aware of plagiarism policies and they need to also take personal responsibility for how they use it. Right? And when in doubt, talk to the educators, right, think about what the institution allows you to do. Absolutely, Cairo, you've created a kind of a I tutor. Can you talk a little bit of how this works and how do you make sure the tutor is providing information in a way that the student is actually learning as opposed to sort of enabling them to sort of shortcut through the learning process and maybe not actually take on board the information that they should be acquiring. All right. All right. So just a little bit of context. Panda is an education app to help, you know, school students outside of school or after school to do, you know, revision study to access new educational materials. And Panda started in 2020. Now, we have more than half a million students using it. So um we incorporated I and one of the features is a chatbot that helps students with their homework, with their assignments and all that. But what we've done different is that as unlike chat G BT, the chat board in Panai actually accepts, you know, students lingo, it is trained on the curriculum. And what's important is that it's designed not to give the outright, but you have to give you step by step solutions. So um to ask the students, you know themselves, what's next and what do you understand now? And where do you think the next step is? And it's as if a personal tutor is sitting next to the student asking the students questions rather than, you know, giving the answer outright as if it's a calculator. Right? So the danger is that with, with tools like JA I and CT like 25 mission is that students would outsource everything to the A I, right? And we've heard many times that, you know, um we can treat JN I as another education tool like cos or when two came out or Excel or sheet, we can move on to more complex stuff. But the danger is that if we allow that to happen, the student will outsource not only their writing but also the thinking process behind it. Right. So with calculator, the student have still think about the question, what numbers to input, what pressure to put similar with Excel shape and Google, etcetera. But with gen I, you can just copy everything into gen without even reading it and then you get the answers. Right. Right. Which is quite presentable. So that's the, you off the thinking. So, right. And then you use a sort of socratic method to prevent, you know, students from sort of leaping ahead to just having to do everything for them. Tim. I, I know it. Uh mmbz you've been using LL MS to help, uh you've created LLM that can help people improve their modern standard Arabic. I don't know if you can talk a little bit about that. And again, how, how do you make sure that it actually functions in this kind of tutoring role? Uh So as, as complex Mezua, I the Mohammed Bin Day at the University of Artificial Intelligence, we, we're based in Abu Dhabi. Uh Greetings from UE. Um uh We're watching very carefully what's happening in Singapore learning a lot of lessons. Uh So we're the world's first uh A I university or a graduate university masters and phd students, about 70 faculty. Um We're uh um unlike most uh universities, we're actually in the game of commercially relevant LLM model training. Uh So if you've done anything involving uh Arabic, if you've done anything involving Indonesian, uh if you've done anything involving A I, safety chances are you've, you've used uh LLM or you've used resources that have been developed at the university. Um So we, we very much sort of straddling the, the commercial academic divide uh as a university, a New Paradise University, if you like coming back to the actual question. Um uh um the use cases for, for LL MS and A I more generally in education are very real um uh in terms of research, you know, I think there's a, there's a bit of a, um, a, a tip of the iceberg effect that we're seeing in terms of G GP T. Uh And it's important to realize that there's a lot more that's happening in, in A I uh beyond just uh LL MS. Uh So, um we're also working in the space of personalized uh learning uh trying to increase both the, the access and the scalability of uh education. Um in, in part using LL MS. Yes. Uh but um very much trying to uh learn how a student is learning, uh um try to learn what are the points of confusion uh in terms of their learning process and, and cater the presentation of the curriculum and the generation of questions to uh what their weaknesses are and to try to complement their strengths uh um to uh um both increase the sort of personalized tutoring access. But if you come from a more wealthy background, you, you can, you perhaps assume uh to students throughout the world, not just in English uh in languages uh beyond English as well. And I think that it's very important that we don't fixate too much. Yes. As educators, we, we sort of think about plagiarism and, and through cheating, it's a very real uh issue uh as a, an educator, uh that I do think about. Uh, but there are, there are so many benefits, so many good things that can come out of the A I models. Well, beyond the, the sort of simplistic chat QVD, use cases that I think we need to be aware of. Yeah, absolutely. That's interesting. I want to come to questions from the audience in a little bit. So please think of your questions and I'll come to you in a minute. Um Cairo, I know that, you know, with panda, it's not just about their supporting the student, but you've also built uh aspects of features of the product that support teachers. Can you talk about that? Because I think that's sometimes overlooked. There's so much focus on what is the student doing and, and less on. Well, how can this technology actually support teachers? All right. So when we started in 2020 our product was designed to help students directly. So students will go to school like they normally do. But after school outside of school, instead of using paper based stuff they can use instead. And then um around a year and a half ago, teachers actually came up to us saying, hey, um is there something we can do for them? Right. So now we have a new module called Panai teacher where students can actually, teachers can actually teach there have their own ID and have special functions. That's what that for them. So I don't want to get into so much detail, but it it covers, you know, teaching, teaching planner, it covers monitoring the students, analysis of the students as well as self development content to develop themselves the teachers themselves rather than just to develop the students. So what we've seen is that there's um you know, a valid concern on the over reliance on A I interaction that diminish the vital interaction with the students and their teachers, right? So um it's it's important that no any tool that we develop actually enhances the interaction with teachers as well. And that's why, for example, in pan, we have the par model, the teachers model try to complete the ecosystem somewhat and, and and we've seen even with our chatbot when students are, um they um initially, yes, they were asking academic questions, but after months of, you know, using the product, students actually started to share about their personal life, sharing about their parents, sharing about their boyfriend, their friends, about the problems in school, their friends, right? And confiding into the chat board, right? So that's also a danger where you know, young students, especially because we were dealing with K 12 primary school students as well, treating the board as if like a friend, right? So replacing the vital interaction with his students and their friends and social interaction with teachers, et cetera. So that's the den. So um the, you know, the ethics behind it so interesting. I, maybe we'll talk a little bit more about that. This sort of the way these, yeah, maybe this, this morse from being a tutor to kind of being a companion chap bot and, and maybe there's a sort of tendency to, to go that way. Um, uh, so I wanted to ask you if you, do you feel like, uh, either the way teachers have to teach, needs to change or what they have to teach needs to change. Uh Because of gender of A I and because of the, you know, the popularity of these chef bots, um So I'll share the perspective, the legal perspective since I'm with the law faculty. So I guess projecting into the future, the now, maybe we know that human and machine will need to work together, right? There needs to be some interact, there needs to be interaction there and how the unit operates, leveraging the strength of each that will make it a AAA positive um combination. We also know that things will change extremely quickly and people will need to learn and unlearn skills and knowledge extremely quickly. So, taking that in mind and coming back to the curriculum that we teach at the law faculty, the question then is what do we need to change? And how might we need to change how students learn things? Obviously, there will need to be a law and tech element. Um And by this, this exposure to technology can come in various different forms. Students need to be aware of obviously the latest digital products out there so that they can think about legal issues that can arise from that there can be exposure in terms of how educational tools personalized learning because then they can see to what extent it is useful or not and the strengths and weaknesses of the tool. So playing around with that, taking personal responsibility around that um exposure to what's happening in the industry because that changes at a breakneck speed as well and learning how to deal with um or do the digital speak, I suppose in the in the foreseeable future where um students or lawyers need to speak to technologists, right? To build products that could help the industry. Um But apart from that, I think in terms of just thinking skills, um you want to have higher order thinking skills like analytical skills where students can look at complicated things, break it down into its component parts. Understand the relationship between the two, you need synthesis skills where they can look at different threads of thought and create new patterns. You need also the values that students are invited with because they need to understand why they are doing things. They need to understand why they are serving the community and those things aren't changing. Those are things are sort of always been there and they remain consistent, right? That's, but I think now they're more important more important than ever. Yeah. And if you don't kind of celebrate the humanness of it, I mean, that's what your combination and the human in the loop is meant to be. I want to get questions. I know there's at least one question from the audience. So if we can go to the gentleman up here in front, he had his hand up and then I also see that someone in the back. Um, but the gentleman up here, I saw him first. So, uh I don't know if we can get a mic to him, uh quickly, if not just stand up and, and say your name and where you're from and, and what your question is, should know how to use A I, whether it's the earlier predicted A I and A I to be ready for the A student has to learn that otherwise there could be universities which are doing it better than you and your students get left behind. So rather than saying plagiarism or let them use Google came in and people started using Google. So the whole curriculum has to change in terms of asking questions which are about critical thinking where they just can't copy paste from chat G BT, right? So, uh a student has to be well versed with all the latest stuff uh and change the curriculum, change the question and you're seeing very rightly that students should be able to work with technology. A law student should be able to work with technologies to come up with new A I models which help uh law. That's a good. Do you have a, do you have a question there? Is it just sort of a comment on, on that? I'm saying that the critical, the questions have to change in the curriculum has to change. I think that's echoing what Sunita was saying. Which is that, yeah, that maybe you change a little bit to emphasize the these skills even more. Um now because you have technology that can do a lot of the, the simpler things. That's what you were saying, Tim, do you, do you agree with that? I mean, I do and I think this is something that all universities are grappling with at the moment. I think even in the A I space for us as a university, our focus has been very much on education of, you know, the, the A I innovators, the people who are going to be driving the next wave of uh um A I innovation for in research. And absolutely, we need those sorts of people. Um But one question we've been asking ourselves is what would an undergraduate curriculum look like if we had AAA blank slate uh based on the existing technologies that are already here as well as what we know is coming? Uh What should that look like? Uh And the answer isn't necessarily that, you know, they need to understand linear algebra and they need to be able to, to uh train in Pytorch. Uh It's possibly there's an even greater need for A I super users. People who understand how these different models work. They understand the weaknesses of the models, they understand why there are going to be shortcomings in terms of um you know, whatever there was a lot of talk about through culture values uh um in city answer. But why is it that you get biases in models? Uh How is it? I can use the most successful, how can I buy them in different settings, whether it's uh legal settings or, or business settings? Um But not necessarily the, the sort of core skills we able to train LL MS uh you know, they, they, we don't need the world to be awash with LLM trainers. You get a couple of 1000 of those in the world and you're set really. Uh there, there's a lot more innovation to be had but uh really, it's the flow through effect. Uh and the, the middleware if you like and, and I think uh for a modern A I curriculum, even in computer science, I think these are the things we should be thinking about rather than, you know, the sort of classic, teach them algorithms and data structures. And, you know, have you taught the way I was taught as an undergraduate? No, I don't think that's the case at all. Interesting. I know there's one other question uh the woman has her hand up uh very high in the middle of that there if you Yeah, yes. If you can stand up and we'll get your mic to you very quickly. And if you could just say who you are and what your question is. OK, thank you so much. Uh This is one way from Singapore and really happy to be here to hear your perspective because like the definition of education has like widened with like technological cycles, getting smaller and smaller. So my one question for everyone here is that uh how do you, what is one practical tip that means not theory, practical tip that you have for everyone here because we are all here to learn, to decide what to outsource and what not to outsource. Yeah. Keen to learn from your 11 practical tip Cairo, they have one practical tip for what students can. So maybe rely on an A I tutor for and what they shouldn't rely on an A I tutor for II I think um on the student side, they definitely will try to cheat, right? Ie the fastest, the most efficient way to cheat, right? But um like this gentleman has said, um teachers have tried to change the way they do evaluation, right? For example, like um um now uh exams are being done on paper, essays are submitted with editing history now and then assignments are increasingly more personalized. So to force the students to answer and include their personal stories as well, right? Rather than something they can copy of the internet only. So I just want to highlight that it's important to highlight to the students. What is education in the first place? Right? If, if um the message is is we say you go to school, you go to college, you buy a big house, you buy an expensive car. Then if that's the case, the message is that this is just a transaction and obviously students are going to cheat. But on the other hand, if you say it is actually a way for us to for students to become an informed citizen, a better decision maker better, then maybe that's, you know, the way it is supposed to be. So we can ask for philosophical question to our kids, but it invites us to, you know, rethink how we think about education as well. So the mental of education. So I need to do you have one practical sort of tip or or thing that's about output what you should. So I second I rule with what it means to be human. Like I said, the human and machine interaction, right? So you need to leverage on your strengths and what are those strengths that will be important? Your critical thinking skills, your interpersonal skills, your emotional intelligence skills, I think you make sure you're not out sourcing those skills, right? Because you're just cheating your future self essentially and Tim, you have one tip for people. One tip agency and accountability are the two things that we have to hold on to. So whatever you're submitting as a student, it is your work. If you're using A I tools as part of that, you need to own it. So if it's code, say you have to be able to tell me what every line of code does that you're submitting, even if you've used code uh to generate it. Uh And you need to own it. And I think it's important to realize that a lot of these issues we're talking about aren't actually new issues. It's, it's a new guys uh because we're talking about A I tools. But, you know, contract, if a rising has been around, whether you like it or not for a number of years, we've been grappling with it as universities. The big difference here, as I, as I say is agency, I can use these tools to uh um enhance the work that I'm doing and embed them into my work flows. But I am ultimately the owner of what comes out of it and I have to be able to defend everything that's there. If I'm doing uh getting someone else to write my essay, obviously, II, I can't do that. This is me just with uh you know what, whatever uh um uh embellishments using A I tools. Great. Well, that's a great note to end it on. Um we're out of time. I'm afraid if we could go on a lot longer. But I want to thank my panelists and thank you all for listening. Thank you.