Episode Transcript
Speaker 1 00:00:10 This is the Owl Cast, the official podcast of ACS Athens. Listen to the exciting story of the American Community Schools of Athens. Check out what drives all the members of our international community of learners as we create the education of the future. Here's John.
Speaker 2 00:00:42 2023 will probably be the year when the impact of AI in everyday life really hit a milestone. The large language or language, predictive models such as chat, G P T, took everyone by storm with progress and updates happening in a breakneck speed in the future. The effect of AI will definitely mark 2023 as a keystone year. Last year, we had our first opportunity to discuss with our guest the impact of AI and science on education. Today, Dr. Adonis Carla Science and Technology faculty at a acs. Athens talks with us more in depth about this topic and we attempt to answer some more profound questions. And the timeliness of this discussion is very telling of the speed in which things change. We had this talk earlier in 2023, and it became inevitable that we realized today how things have changed since that discussion took place. Dr.
Speaker 2 00:01:41 Carambas is developing a K-12 artificial intelligence framework for teaching and learning, including learning goals, curriculum dimensions, topic areas, and a K12 curriculum progression chart. He's a co-author of about 70 articles in astrophysics and education, and has attended more than 25 conferences and workshops with several contributions of his own. He's a member of the International Astronomical Union, the American Institute of Aeronautics and Astronautics, the International Astro Statistics Association, and the HE Educational Society of stem. His interests include AI education, steam education, space education, educational innovation, teaching and learning analytics, game-based learning and education for the fourth industrial revolution. Today with Dr. Car, we discuss AI and the large language models such as chat, G P t, predicting rather than constructing content, the importance of AI in science and technology curriculum, how to use AI without giving up our thinking skills and the possible faulty effect of AI due to cultural and social biases and the human factor.
Speaker 2 00:03:06 I believe our last conversation last year hit a chord in our community, especially our discussion about artificial intelligence, which is the topic of today's discussion. It's always good to talk to you, Dr. Caramba, as welcome Back to the Outcast. You are leading the development of a K12 artificial intelligence framework for teaching and learning, including learning goals, curriculum dimensions, topic areas, and a K12 curriculum progression chart. There is a great faculty and student team behind this effort. This year we started the innovation lab, which among other exciting things, it develops the means and tools for our community to immerse into artificial intelligence through teaching and project-based learning. Since last year, AI has been pushing its way in leaps and bounds through schools, and there is no international conference in education without a strong emphasis in ai. I'd say let's unbundle what's going on? How has AI progressed in the last few months?
Speaker 3 00:04:10 Uh, thank you for having me, Mr. Papadak. It has, uh, really progressed a lot in the sense that what is now available to the public that was not available before was this large language model called cpt. And uh, it is a model that allows the user to ask questions to intelligent machine and get convincing answers.
Speaker 2 00:04:34 Human-like answers,
Speaker 3 00:04:35 Human-like answers. And the actual technology is called natural language processing, um, and natural language generation. And, um, it's only a few months from, uh, our previous podcast, but yet with, yes, today it is a different landscape and I think it's not just the technology itself that is, uh, was made possible, but also the fact that it was made available to the public because, uh, the d PT three engine, um, is around for at least a couple of years, but it was only available to those who would subscribe and use it as a product. Now everybody with an internet connection can really access this powerful output.
Speaker 2 00:05:19 You mentioned that it's the large language model. What does that mean?
Speaker 3 00:05:24 Uh, first of all, uh, what those models do is to understand the human language, uh, at least the one that's uploaded on the internet. So, so
Speaker 2 00:05:33 It understands the nuances, it understands what you expect to mean by it, and then they learn by the different ways of people talking. First
Speaker 3 00:05:42 Of all, uh, the, the attempt is to predict what is the next word that should be generated, uh, after certain, certain prompt and after the frequent connections between words are established and the pattern is identified, then this can be associated with some meaning because this output is not capable of, uh, returning human-like language, but also to understand at a good extent what the user is trying to to get. So it's both the structure of the language and the meaning of the language at some extent. Of course, it's not perfect. Mm-hmm.
Speaker 2 00:06:17 <affirmative>. Okay. Maybe this is a question many people have. Uh, if I do ask a question on this particular chatbot or however we want to call it, and I ask this question multiple times, am I getting different answers? Every time we
Speaker 3 00:06:33 Are getting different, different answers have every time, um, there's a button available that says regenerate the response because it's not a a linear, straightforward model. It has many possible answers and is itself trying to figure out what's the best way to, to deliver text. And that's one of the reasons that the, this company Open AI released to the public. So they have a lot of users data to, to improve it. Um, there are many options for this model to return, and it's supposed to be fine tuned over the time to it better.
Speaker 2 00:07:06 I think my next question will be following us for the years to come. Why is artificial intelligence so important to the science curriculum nowadays? What is the benefit of using ai?
Speaker 3 00:07:18 Uh, it's important to, for the entire q crew of K through 12 and higher education, uh, it can apply to, uh, all courses because it is the most powerful and important technology of our time and supposed to, uh, to be, to remain the most powerful one for the next, at least one decade. So we are going to use AI in a professional life in our social life anyway. Students are going to use ai, uh, in so many different ways. Uh, therefore it is important just to be incorporated in the curriculum, not only knowing about ai, but also how to use ai. It goes way beyond the science curriculum. It stems from computer science and technology statistics, and then it applies to all aspects of society.
Speaker 2 00:08:07 And, and how about the challenges to the learning community?
Speaker 3 00:08:11 Delegating decision making and learning to machines is something that we need to think about more carefully. We don't want our students or even adults to give out or to give up skills, thinking skills. Uh, it is important for this technology to be used in a way that is releasing us from repetitive maybe tasks or mundane or dangerous tasks, and, uh, allow us to have the time to become more creative. But there are really many, many limitations. And it's not only on the way we use it, it's also about the access to ai. Not everybody's going to have the same access to ai. And yes, it is true that ZBT is one step towards democratizing ai, but to speak the truth is not gonna be the case, at least for it's gonna be a transition until everybody, uh, can use it.
Speaker 2 00:09:02 Well, as you said, chat GT is the latest and probably the most popular chatbot. I'm pretty sure there are other coming up. Uh, you type a question. This tool returns a factual answer in a human voice. I believe this is an excellent example of technology where the intelligence of the input dictates the intelligence of the output, or as developers say, garbage in, garbage out. Um, is it logical to say that it needs much more thought to create the question that just Googling it?
Speaker 3 00:09:31 I think it is, it is quite a case. <inaudible> is able to generate, um, convincing and competent language, therefore the prompt has to be at the same level at least. And, uh, it's not only the thinking part, the human factor, uh, that is requesting something, but also the human factor building the model. So it is the technology that is augmenting, uh, the humans, uh, that has become very powerful, uh, right now and might be preventing us from see the human factor that is always there, uh, with us at the level that this technology is becoming so intuitive that we might sometimes fail to feel it's something we that we have made. Like it's not straightforward thing. We can see it, it's software, it's not hardware. Um, it's sometimes it feels quite, um, humane. Um, and that might blur the, uh, the limits, uh, between the biological and the digital, uh, sometimes, but the human factories in both sides, I think.
Speaker 2 00:10:32 Are you concerned by this? Are you scared? Are you, what, what is the feeling that someone might get, uh, like someone in your position, you are a teacher, you're a scientist, you're a researcher, you, you know about ai. Is there something that rubs you the wrong way with the way that this thing has come out?
Speaker 3 00:10:51 Uh, practical positivity is usually the way I, I think and move. I am optimistic, but I, I see no way for somebody to process holistically and arrive to utopia or dystopias. There are limitations. There is a matter of accessing information. There is algorithmic bias, as you said, garba in garbas out. But data might be, might be reflecting biases, might be reflecting inequalities and fairness because society is unfair. And plus we process this data which are are unfair already. So, um, bias, um, magnifies, um, there are many, there are many, uh, limitations. The ethical use of technology, if, if it's, but can write an essay for me, do I know what's the right thing to do or do I have this internal drive to stop myself from abusing technology or for, you know, being dishonest. But I mainly see the, the poster because that's revolutionary. It's not only, uh, about efficiency and you know, being more productive and all that. It's about having this available to us, this intelligence. We never had that chance to, as a general public, to, to discuss with intelligent machine before. And I'm happy I'm living in this era actually, and I'm, I'm curious to see what are the, what people are going to make up with, uh, technology.
Speaker 2 00:12:12 Curiosity feeds creativity.
Speaker 1 00:12:20 You are listening to the OWL Cast, the official podcast of ACS Athens.
Speaker 2 00:12:37 The impact of AI in education, the challenges it poses and the potential it offers to everyday human activity is our topic of discussion today with DR and Donis, Carabas science and technology faculty at acs. Athens. Stay with us as we discuss the challenges schools face with AI chatbots and the question of banning or embracing ai, higher order thinking, the guiding principle for ethical use of AI in education, the concept of personalized learning, physical, virtual, augmented and mixed spaces of human activity and the importance of fundamental literacies. What is your opinion about the decision of some school districts in the us and I am sure many more around the world are considering it to ban chat G P T or similar, uh, software from their students. First of all, can it be banned? And second, what's the point?
Speaker 3 00:13:41 There is no way it can be banned. And I don't see much point. I understand some districts might, um, have wanted to act immediately to prevent, I dunno ing, but I mean, looking at the educational perspective that is more holistic. I see no reason to ban it. Our students are going to become, adults are going to be using AI anyway, and this is just the, you know, the, uh, the tip of the iceberg that's gonna be available in a few years. I don't agree with this, with the banning. I understand the fear sometimes and the concern, but what about we use this technology to augment our education system and make us better learners, uh, more noticeable people.
Speaker 2 00:14:24 Well, um, you know, in your latest article on Medium, you mentioned that schools instead of banning judge G P T, they should still try to harness the power. Its predictive language model and human-like responses. So you are in the classroom, okay, you have an assigned activity for the students. Up to now, students immerse themselves in the knowledge they gained used critical skills to synthesize and evaluate their response and share it with the rest of the classroom. How can you, as an instructor invest time and effort to help students use the new AI technologies in an ethical, as you mentioned, an honest way for their own learning benefit? What are the steps? What are you telling them?
Speaker 3 00:15:05 Well, first of all, we need to discuss with them about what's the importance of tech or technology, what's the meaning of coming to school? First of all, if we realize why we come to school, what's this, the education about? Because
Speaker 2 00:15:16 That's another question. Would we need schools in the future? That's
Speaker 3 00:15:19 A big question. And yes, we do need schools. The question of course, uh, schools need to modernize their, their offerings. Always this is the case. Um, so I would, I would discuss with my students about ethical decision making, first of all, then this is a tool that is going to help them be more creative. This is how it works, use it and, uh, create something, uh, that requires higher order thinking than I would expect before. And this is the, the real human machine collaboration era that we're getting into. Um, I would like to see the human creation. I would like to see the humans last machine creation as well. Using this technology doesn't mean it gets you to solving problems. And there's a variety of assessment and, uh, we don't have closed questions. It's open questions that we ask them to reply to and build upon. Then let's use this technology and see what they come up with.
Speaker 2 00:16:14 Can you give me an example of a question that you would most, uh, regularly make, uh, to the students? And then how would you rephrase, how would you rethink that question knowing that the students might be using such a technology?
Speaker 3 00:16:30 Um,
Speaker 2 00:16:31 A simple example. Don't get me to hire physics now,
Speaker 3 00:16:34 <laugh>, of course. And, uh, well, I would, uh, reframe myself from asking them to complete an essay home. I would like to them to discuss with me about the, the process of thinking that led them to a certain solution or suggestion or any, any problem solving situation. I would design assessments maybe in a better way just to ask so that there's enough variety in the assessments so I know where they are in the learning. And what I would try to do in the near future is to actually use it to, to do things that I would not be able to do myself, like personalized learning. Imagine I'm teaching a concept and the students are practicing after, uh, this small lecture where about some students, um, um, are behind and they need personalized support, one-on-one tutoring, where about I have this system that can really be used by the student to, you know, fill some gaps without me having to learn myself in a certain amount of time. The personalized education, this is something that is going to become a bit more feasible in the years to come. And that's something I, I will never be, uh, able to do myself in the classroom on my own.
Speaker 2 00:17:46 Uh, I'm gonna take you to a different, uh, tangent now because I know you are a person of, uh, science and STEM and ai, but there is a big discussion about the, the mental, uh, health capacity of students, especially those that according to the latest research, uh, since 2012, students are spending 11 more hours a week alone at home. That probably translates 11 hours per week online. So having a machine, having a model of artificial intelligence coming back to them and giving them feedback about their lives or about their schooling or about their relationships, how does that make the physical space of school important for the mental health of the students? I mean, I'm pretty sure you've seen students struggling, especially after the <laugh> pandemic, but I'm pretty sure this is going to be a trend in the next years.
Speaker 3 00:18:51 The screen time is skyrocketing, and if we don't try to deal with this problem at the young age, then it might be very difficult for, for a teacher to deal with the sort attention span of the students or a mental state that might not be, you know, prone to, to learning. One question I ask myself many times is not to reduce the screen time, especially from middle school and high school students, that we cannot really ban the use of technology, but how can we have a different mix of use of technology? How can I steal a bit from my entertainment and leisure to replace with, uh, creative, uh, problem solving or creative work? Another remark I would like to make is that we are now using this chat bots by typing. It's going to, it's a matter of type until we are going to be using our voice.
Speaker 3 00:19:43 Because really voice recognition is not such a big, uh, easy to solve nowadays, quite trivial. I think that as technology advances and we use more of our senses, screen time is going to decrease to decrease. Absolutely. So I'm, I'm quite positive for that. But still being immersed inside technological spaces, digital spaces, metaverses is something that we need to be very careful about. And, you know, be mindful when we use physical spaces that we have to, and how much time we do be around nature be discussed with other people. That's very important just by discussing with another person, is, is a physical space that cannot be replicated online.
Speaker 2 00:20:27 Uh, Conrad Hughes, a well-known and distinguished educator from the International School of Geneva recently wrote that using artificial intelligence to accelerate processes and find solutions should not be the problem, but it will be our fault if we let artificial intelligence carve out the end of our thinking and not the beginning. What are your thoughts on this?
Speaker 3 00:20:49 I, I'm closer to say that it's not about deciding whether it's the beginning on the end, the technology is there available 24 7. We cannot augment any part of our thinking. The question for me is what are we giving up? Uh, what part of our thinking are we delegating to the machine? And this could be at any point, whether is the end of thinking or the beginning of the thinking. If the thinking is structured and is written in, uh, in tools, if I know how to problem solve, if I know how to think, uh, creatively, I can use my imagination. There is no way it's going to be such an obstacle beginning or the end of my thing is going to make me less of a thinking person. But I get the point. And the point, uh, I suppose here is that we have to be purposeful and focused on ourselves and our, our humanity.
Speaker 2 00:21:44 So what would you answer to parents or even students who ask you whether the value of education in schools is or will be challenged by ai? I think you mentioned a little bit, but I think it's important to stress the, the importance on this. And
Speaker 3 00:21:59 I think it's most important, more important than ever to realize how the fundamental literacies are important. Imagine somebody who does not go to school to follow a structure program to become data literate, technology literate, AI literate to collaborate with other people, live with other people, uh, try to be empathetic through relationships, having mentors, teachers, mentors to guide them through processes. This cannot be replicated. I think it's more important than ever to, for a student to be knowledgeable and flexible in their learning and as a screen and start what cannot make you flexible. Those bots are translating information in a way that is, uh, is human like, and this is, this is, they are think beyond knowledge and think beyond skills even, uh, has to do with values and dispositions and how we approach technology and the society. Whether we do our best to solve global problems, I adopt this can be, uh, really replicated by any otherwise, we'll just Google it and find a solution.
Speaker 2 00:23:07 Google is another whole discussion whether Google is going out or whether it's going to be adopting to the new reality. Um, I, I think we're living in very interesting times and, uh, we see that around us every day, especially those of us who are in school environments. And, uh, we have a lot to think and a lot to learn from using this technology. Dr. Caram, thank you so much for being with us again. See you on the other side.
Speaker 3 00:23:35 Thank you for having me.
Speaker 1 00:23:40 You are listening to the Owl Cast, the official podcast of ACS Athens. Make sure you subscribe to the outcast on Google Podcast, Spotify and Apple Podcasts. This has been a production of the acs, Athens Media Studio.
