Transcript: AI in the Classroom - Productivity and Academic Integrity
[INTRO MUSIC]
[Gursveen] (0:05 - 1:22)
Hello and welcome to Intersection Between Today and Tomorrow. We are your hosts, Gursveen and Tvisha. And just to introduce ourselves a bit, we're both students at the University of Toronto Mississauga studying in the Mathematics, Computer Science and Statistics Department. In today's episode, we're going to discuss how artificial intelligence, this constantly evolving technology plays a role in the intersection between today and tomorrow. Specifically, we're going to delve deep into how AI plays a role in our educational system. So we've all heard about the conversation surrounding AI, like AI is cheating, AI is a tool, AI is the future, AI is destroying education.
Everybody seems to have a very strong opinion on it, but nobody can actually agree on where the line is between accepted AI usage and where it kind of gets out of hand. But there's a reason that this happens. It's not just like without cause, because when you dig more into like the research and stuff, you realize that there's a lot of factors that actually come into play when it comes to AI and how those factors affect our educational system.
So today we're going to look into what research actually says and when it comes to AI, not only in terms of like academic integrity, but also about productivity and quality of learning in our educational system as it is.
[Tvisha] (1:22 - 2:18)
Right. You kind of wonder, is it actually worth to be quote unquote cheating? Does it genuinely benefit the students and their learning? Well, a study in Romania actually found that 80% of students believe AI based technologies positively impact their educational experience (Vieriu & Petrea, 2025). If you really think about it, there's much more going on to this on a much deeper level, because generally if something seems too good to be true, it probably is. For now, though, let's jump back to Gursveen's topic of ethics. So everyone seems kind of lost when it comes to AI and academic honesty. I mean, me personally, I've seen many peers, professors even, and full on institutions just not know what to do about this–well, not such a sudden change, but it did creep up on us out of nowhere.
[Gursveen] (2:18 - 2:19)
For sure.
[Tvisha] (2:19 - 2:24)
So where does this confusion actually start and where does it come from?
[Gursveen] (2:24 - 2:55)
I feel like in terms of AI and academics, no matter the discipline, there still lies the core problem that Gen AI hasn't just made cheating easier, but it's also made the concept of cheating harder to define. Right? Students often are left wondering, does this actually constitute cheating when they find themselves using AI to complete schoolwork? And the foundational assumption is that honest work is equal to learning. That's what we assumed for like the longest time ever.
[Tvisha] (2:55 - 2:55)
Right. Definitely.
[Gursveen] (2:56 - 5:46)
But now that's like under pressure. Before it used to be like a linear progression. Like if you keep on doing this honest work, if you push through the reading, if you write that terrible first draft. And using your own thoughts and like revising it, polishing it before you submit. Right. The assumption was that you actually learn something. So, you know, that friction that you experience in that creation process, that friction actually is what like conducted that learning. But this exact equation is now under tension because like the line between learning and the use of AI, that is blurred right now.
Because, for example, if a student uses AI for their first draft, but they write the paragraphs themselves, does that constitute cheating? Or if you use it heavily to like check your grammar and stuff, right, does that constitute cheating? So like this exact tension is the primary focus in the Corbin et al. 2025 article. So it was a qualitative study out of Australia and they kind of went deep into the idea of “drawing the line” (Corbin et al., 2025). So like between accepted AI usage and what constitutes cheating. They conducted in-depth interviews with 19 students and 12 academic staff members to kind of see how they perceive these boundaries. So the metaphor of like drawing the line, that wasn't just something that like the authors came up with. It was actually something that more than half of those who were interviewed mentioned.
So you can see this emphasis on drawing the line. So it kind of implies that like everyone's looking for this sense of direction, this sense of like structure in a completely structureless environment. So they're kind of like begging for rules in a sense, and the study ultimately finds that students aren't just sitting down and deciding to cheat. Like, you know what, I'm just going to quickly finish this assignment. But higher education institutions themselves haven't resolved this kind of ambiguity between like what's right and what's wrong. So students are kind of being forced to go off of like, they're kind of doing it on the fly. Like they're kind of constructing their own subjective ethical frameworks between what they should do, what they shouldn't do. Because there's kind of like this massive high stakes guessing game. So it's like one student said in the article, they mentioned that one student said, and I quote, “You have to kind of guess your teacher's perspective on AI and then figure out how much to use it or not” (Corbin et al., 2025, p. 708).
And I find that really interesting because like to put this into perspective, suppose you're playing like a really high stakes, professional, intense basketball. Right. And the referee decides to like changes the rules on you. And just based off of like how they feel. To me, that really wouldn't make sense. Because like before you even touch on the content, you're not you're kind of supposed to think of the psychological profiling of your professor and like that constitutes like how you're allowed to think.
[Tvisha] (5:47 - 6:05)
Exactly. No, I totally agree with that. Because at some point it just becomes a question of, oh, is my teacher quote unquote nice or are they mean? And you kind of just have to interpret on day one, whether or not how much you like your professor or you, you just have to like to go based on vibes.
[Gursveen] (6:05 - 7:21)
Yeah, exactly. Right. So it's kind of like this cognitive burden that the article ultimately entails. Like students are essentially like calibrating their own academic integrity just based off of their professors take on it. So whether they're an AI optimist or an AI doomer. Also something we often overlook, right. This also has a burden on the professors themselves. So like one faculty member in the interview, they mentioned feeling stressed all year marking and constantly like second guessing every brilliant sentence that the student wrote (Corbin et al., 2025). And one faculty member even went as far as calling their experience with grading an “ethical trauma” (Corbin et al., 2025, p. 711). So you can really see the emotional stress that's causing for teachers.
And so like moving forward, we kind of need to understand like the sheer power of the tool that students have open access to. And this is kind of the focus of Cotton et al. (2024), the researchers. To kind of prove their point, they had Chat-GPT write almost the entire paper itself. So they kind of fed it prompts. Yeah. And like generative AI is like a fundamentally different form of cheating, like we've ever seen before. Right. So the article also kind of goes into “contract cheating” (Cotton et al., 2024, p. 235). I'm not sure if you've heard of it before.
[Tvisha] (7:21 - 7:21)
No.
[Gursveen] (7:21 - 7:24)
But it's like, for example, like paying a desperate graduate student, like $300 to write your history paper. That's what contract cheating ultimately is.
[Tvisha] (7:24 - 7:24)
Oh okay.
[Gursveen] (7:24 - 7:41)
So why is it different from like cheating that we have with AI now? Well, that's actually because the “contract cheating” has this immense built-in friction that acted as like a natural deterrent from using it. Right. So for example, paying someone to write your essay that costs money and sometimes it's a significant amount more than you can afford.
[Tvisha] (7:41 - 7:42)
Right.
[Gursveen] (7:42 - 7:48)
Just for a little essay and you had to wait days to get it back. It also, you also have to communicate with a human.
[Tvisha] (7:48 - 7:48)
Right.
[Gursveen] (7:48 - 7:54)
So that also has some like social risk kind of aspects to it. Because there's kind of like those lack of privacy, because you could get hot for cheating at the end of the day and that can have like consequences, serious consequences.
[Tvisha] (7:54 - 7:54)
Yeah, for sure.
[Gursveen] (7:54 - 8:23)
Cotton et al. (2024) ultimately they mentioned that AI can effectively mimic final work. Right. Like the final destination the student can do. Right. So just evaluating this final work, it's not like, it's not a good form of evaluation of a student's work. Right. Because it's also the process that we need to take into account. So evaluating the process is ultimately also crucial to avoid this kind of issue with academic integrity in terms of AI.
[Tvisha] (8:23 - 8:40)
Yeah, for sure. And I mean, you even mentioned like right at the beginning, right. When you kind of go through that whole process of writing the first draft, doing all those things, getting stuck. And that I think is where the learning actually lies is when you get stuck.
[Gursveen] (8:40 - 8:55)
Exactly. So now we want to kind of see like what's actually going inside a student's head when they like suppose it's like 10 p.m. at night, they sit down in front of their computer. They're like opening ChatGPT. ‘Let me get this essay done in two hours before the deadline.’ Right.
[Tvisha] (8:55 - 8:55)
Yeah.
[Gursveen] (8:55 - 9:25)
This is the exact kind of question or idea that's tackled by Hanh and Duyen (2025). The study looked at 10 in-depth, like reflective essays from postgraduate students in Vietnam. Right. And the researchers analyzed these essays and they found out that they found out the kind of idea that students, human beings don't just like passively absorb a fixed set of moral rules from a handbook. Right. Kind of like looking at morality and AI. Right. So human beings, they don't just like look at like this ‘Booklet of rules of morals.’
[Tvisha] (9:35 - 9:35)
‘Book of Ultimate wisdom 101’
[Gursveen] (9:37 - 9:51)
Exactly. That's not how we learn. Like instead we like actively construct our kind of own understanding of ethical ideas based off of our social environment, our peer interactions, peer pressures, stuff like that.
[Tvisha] (9:51 - 9:51)
Personal experiences. Yeah.
[Gursveen] (9:52 - 10:19)
So morality, this kind of shows that like morality, the idea of morality, it's like socially negotiated. So for example, the students might think that like, if all my peers are using ChatGPT to clean up their literature reviews and they're getting praised for their work, their work speed, their productivity, stuff like that by their advisors. Right. ‘I'm going to socially construct this reality where AI isn't just like acceptable’. Right. ‘But it's also necessary for survival.’ Like ‘that's the only way that I can do good.’
[Tvisha] (10:19 - 10:19)
Right.
[Gursveen] (10:19 - 11:07)
That's kind of the thinking that they adapt to. And so now the question is, like, why do they use it, though? Like in the first place, the students mentioned this idea of the illusion of “Harmlessness” (Hanh & Duyen, 2025, p. 7) that like AI speaks to you so conversationally, so like helpfully that it kind of creates this cognitive dissonance between the idea that what you're doing can actually constitute cheating. Like you're not doing your work yourself. You kind of kind of think about, are they actually understanding the implications of the technology that they're using? Like, are they even tech savvy enough or like understanding enough to understand the ethical nuances behind the technology that they're using? For example, one article that I found in my research, it looked at the idea of “digital natives” (Kelly et al., 2023, p. 3). Basically, people who grew up using technology. So honestly, I'd say that was us.
[Tvisha] (11:07 - 11:07)
Right.
[Gursveen] (11:07 - 11:23)
So, you know, we grew up with technology. We're kind of comfortable with it. Right. So this kind of idea is one that Kelly et al. (2023) wanted to look at. So they did a massive survey on over 10,000 students, which is a huge sample size.
[Tvisha] (11:23 - 11:23)
Yeah.
[Gursveen] (11:23 - 11:35)
And in the survey, they kind of measured elements such as student awareness regarding AI, practical experience with AI, and their confidence in using generative AI in various academic disciplines (Kelly et al., 2023) Right.
[Tvisha] (11:35 - 11:35)
Okay.
[Gursveen] (11:35 - 12:01)
So you would think that these students, being “digital natives”, they have experience with AI and they kind of understand how to use AI effectively in terms of academic work and ethically in the process. And the study also found that a significant portion of students who literally never used AI, but they rated themselves as so confident being able to navigate the ethical guidelines between like using AI for university assessments.
[Tvisha] (12:01 - 12:01)
Oh wow. Really?
[Gursveen] (12:05 - 13:25)
Like, suppose you have an engineering student who can understand the mechanics behind machine learning and stuff like that, compared to a nursing student who's never used AI, but they feel really confident that they can use it ethically. How can a university possibly impose a one-size-fits-all policy? How can you serve both students fairly just with one rule?
So the main takeaway from all this you could see is like academic integrity crisis, it's not just like because of a sudden generational influx of like lazy students who are trying to cheat the system
or something like that. Right. But I want to know the implications of this. So we know the reason that they're using it is because it's convenient. Right. And we see that generative AI has a lot of benefits, for example, like in daily life and stuff like that. Right. But, and the students are familiar with this and they like know this idea. But I personally find that AI is like boosting productivity across many disciplines. So taking a further look into how this functions in the process of meaning making, how exactly are AI chatbots affecting productivity in diverse fields? Like how is it affecting the idea of productivity when it comes to students and finishing their schoolwork?
[Tvisha] (13:25 - 14:14)
No, that's a really good question to start us off. So throughout my research, I did often come across studies that spoke to the benefits of AI. As you mentioned, productivity is definitely one of those most prevalent ones. A study that examines AI powered learning assistance in engineering higher education talks a lot more about this (Sajja et al, 2026). In addition to confirming your points about the added fear around the use of AI, the source also kind of acknowledges that students often make use of generative AI just to get over that initial hump, to get them to kind of get past that first block of when you're kind of staring at a blank page (Sajja et al, 2026).
[Gursveen] (14:14 - 14:42)
Yeah, so that kind of reminds me of like writer's block, right? Like art block and stuff like that, right? So for example, like art students, I'm sure any art students listening, they can relate to This. Like when you're staring at like a blank paper, you have no idea what to draw, right? And art block is kind of like this period where you have no idea where to start when it comes to drawing. So I guess writers can also relate to this and I've honestly experienced that in my academic experience.
[Tvisha] (14:42 - 14:56)
Yeah, I know, same, same, for sure. And we can even start to kind of contrast these implications that AI has in educational institutes with those that lie outside. So like after, for example, we graduate, right?
[Gursveen] (14:56 - 14:56)
Yeah.
[Tvisha] (14:56 - 15:10)
While looking into this, it was interesting to see actually the differences in perceptions. So I just want to bring that in for a second, that AI tools for productivity, how they differ in academics versus in the professional world.
[Gursveen] (15:10 - 15:10)
Interesting.
[Tvisha] (15:10 - 17:25)
So a study by Rahardjo et al. (2026) reveals that students found AI to be least impactful when it came to teamwork in collaborative environments. So they report that for students, generative AI, like it helped them by piquing their interest in like the study material. It helped to accurately respond to queries and it acted as a virtual tutor. And number three, it improved their learning efficiency and then like much, much more. And like me personally, although I'm not a civil engineering student, I definitely relate to a lot of these things because it does, for example, acting as a virtual tutor, I've used that many, many times. And then it also talked about, so it contrasted that with kind of how professionals view this (Rahardjo et al., 2026). So they mentioned that it helps to solve problems during constructional stages. It helps to identify structural damage and identifies potential failures in plans, blueprints, et cetera, et cetera. So I think obviously you can right away contrast, okay, like piquing interest in a subject versus like identifying structural damage. These are like very different, very like high stakes situations that you're just relying on AI to kind of do these bigger, higher level things for you in a workplace. And then you can bring in like, okay, in a workplace, we have all these hierarchies. There's like, you're working for a boss who's working for another boss. And there's this whole web basically that you have to kind of live up to their expectations. And it's not just like as a student, if I get caught cheating, if I get caught like expressing something in my essay that's like false information that I got from AI, those implications are gonna come directly towards me and not kind of my whole career, my whole like reputation wise. Obviously there will be some impact, but I'm not letting down like my entire team.
[Gursveen] (17:25 - 17:32)
Yeah, true. So that kind of shows how like people feel comfortable to use AI when it's just themselves at stake compared to when they have their whole team at stake, right?
[Tvisha] (17:33 - 17:37)
Exactly, exactly. And obviously in the professional world, most of the work you're doing will be in teams.
[Gursveen] (17:37 - 17:37)
True.
[Tvisha] (17:37 - 17:54)
And so the stakes just become that much higher. And it's like, as a student, I can gamble whether like I should be using AI or not because this gambling does happen on two different levels. First of all, as you were talking, whether it's acceptable or not, and number two, whether it's accurate or not, right?
[Gursveen] (17:54 - 17:54)
True, true, true.
[Tvisha] (17:55 - 19:59)
So obviously the perception around AI because of these two factors is much more negative compared to for students in academics, right? And so this whole thing just kind of goes to show that perception plays a really, really big role in the use of AI because the output value that it
might bring depends a lot on the kind of, like the potential that AI has depends on the user itself. It comes down to the group of people that's using AI. And I think there's actually another group that's worth discussing, which is healthcare workers. So I found this study and it talks about the fact that most healthcare workers, they understand like the potential for AI in reducing workload and like administrative tasks, things like that (Nazir et al., 2026). And they do show like enthusiasm for further integration. So it's not like people just aren't aware because I mean, at this day and age, a lot of people are aware. In terms of advantages, AI could help improve documentation, professional communication. So that includes like polishing emails, things like that, streamlining tasks, making clinical notes, et cetera, et cetera, right? But most of these benefits seem to be related with reducing workload and obviously ultimately increasing productivity and workflow, right? But again, since healthcare doesn't work in isolation, like I mentioned before, it's a part of a much larger system. So for example, privacy, that becomes a major concern as medical professionals have to practice caution when they're dealing with a more sensitive information, right? And another critical concern is related to the accuracy, which is another major reason why most healthcare professionals stray away from the use of generative AI in practice. Now, again, the study reported instances of hallucinations. Participants said they can manipulate the chatbot to produce a specific wanted answer based on the wording of a question. And I'm sure you've experienced this too.
[Gursveen] (19:59 - 20:00)
Yeah, for sure.
[Tvisha] (20:00 - 21:21)
Right? So the whole data set that chatbots use to inform their responses, that could also be biased. So that again, it makes a really, really big impact on the answers and how the answers could be skewed towards a specific population. And finally, the stakes. They're much, much higher than, again, a student, right? So when it comes to dealing with people's lives and being part of a much bigger system of companies and organizations and just hierarchy in general, it becomes a much bigger issue of accuracy. And I found this really interesting study in Finland where they used generative AI to basically create multiple choice questions for a student assessment (Krushynska & Ursin, 2026). And I quote, they reported that “Of the measured outcomes, productivity was the only area in which automated multiple-choice question (MCQ) generation outperformed manual creation. However, it struggled to produce test items with sufficient complexity. The least favourable conditions arise when developers pursue “universality,” producing algorithms that are not tailored to specific assessments, disciplines, or item banks” (Krushynska & Ursin, 2026, p. 1).
[Gursveen] (21:21 - 21:26)
Interesting. So it's kind of like this trade-off between quality and productivity I see, right?
[Tvisha] (21:26 - 21:45)
Yeah, exactly. Like if you're making kind of multiple choice questions for a more specialized institutions, like dental care, right? They have very, very specialized knowledge that cannot be tested through generative AI questions.
[Gursveen] (21:45 - 21:48)
Like AI can't really replicate that specialized knowledge, right?
[Tvisha] (21:48 - 21:53)
It can't, yeah, because it's just not going to produce questions with the required depth.
[Gursveen] (21:53 – 21:57)
Yeah, like the quality that human intelligence can produce.
[Tvisha] (21:57 - 23:13)
Exactly, definitely. And it just ultimately shows us that the best results come when you kind of go, there's a back and forth. So you're prompting it again and again. You're kind of refining these things and it goes to show like the collaborative nature that AI has. It's not, again, it's a tool. You can't just let it go do its own thing. You have to, exactly. It has to be supervised in some sense or way. So yeah, like I think overall, like these research papers collectively go to show just how imbalanced these perceptions are across several different fields, right? Some people are simply too concerned with AI's accuracy while others think that it's just a tool for productivity whose benefits surpass their own moral judgments of its ethical use. So we know that the misuse of AI is really, really prevalent. It's a massive problem. But I was wondering, how are the systems in universities, how have they adapted? Are there any systems in place currently that are actually equipped to deal with these problems? Like in terms of, I mean, AI detection and the policies behind using AI academics. I wanted to ask you that.
[Gursveen] (23:13 - 24:51)
That's a really good question. And honestly, the honest answer that I have is no. Like the enforcement infrastructure keeps on like failing us, right? It's like it doesn't keep up. It literally doesn't keep up. For example, like AI detection software is biased and unreliable. Institutional policies are inconsistent and blurred. So if we kind of look at the academic integrity problem as a whole, we start to see that it doesn't just stem from the individual, right? You can't just keep on blaming the students entirely, right? But instead, it's like this deeply rooted systemic issue, I feel, within these academic institutions. And in my research, one place that I saw this is in this study by Seeley and Cournoyea (2025), and they conducted qualitative interviews to gain insight on teaching and learning as a result of AI. And the results are striking.
Like the researchers, they conducted like hour-long semi-structured interviews and stuff, right? With like a bunch of faculty members from UTM across various disciplines. Like, for example, humanities, sciences, social sciences, stuff like that, right?
And they found that out of the 61 emotional expressions that they documented, 39 of those were negative, right? So that's like a whopping number. And the common emotions were like concern, fear, like frustration, anxiety, right?
Yeah, yeah. And there were also like these clear disciplinary patterns. Like depending on the discipline, you can also see like there was like this common pattern. And for example, like for the humanities, right? They were more afraid, the faculty was like more afraid that like their discipline was becoming irrelevant.
[Tvisha] (24:52 - 24:52)
Oh.
[Gursveen] (24:52 - 25:00)
They were like genuinely worried that like if ChatGPT can write like an exceptional essay about their topic.
[Tvisha] (25:00 - 25:00)
Then what's the point of like my research?
[Gursveen] (25:00 - 25:29)
Exactly. Like their entire academic value is like at stake, just because of AI. And on the other hand, for example, like scientists, right? They were less worried about these like disciplinary threats. Like they're still going to be relevant, right? Even after the rise of AI. But there were more concerns about like social and like environmental concerns. Like for example, there was concerns about like misinformation, stuff like that at a broader scale. It might not replace them, but it will affect their jobs in some sense.
[Tvisha] (25:29 - 25:29)
There was still concern behind that as well.
[Gursveen] (25:30 - 25:47)
Yeah. So many faculty also described that they were spending a lot of time trying to like, first of all, like redesign assignments to avoid and limit the students' dependence on AI. And also to catch AI-generated work. Like they spent a lot of time, and sometimes they weren't even like compensated for this extra work that they had to do, right?
[Tvisha] (25:47 - 25:49)
Yeah. Traumatic marking experience (Corbin et al., 2025)
[Gursveen] (25:50 - 26:25)
I know, right? The “ethical trauma” (Corbin et al., 2025, p. 711) behind marking. That's so silly. So interestingly enough, the paper also uses the example, you know, when the computers were first introduced, you know, 1980s, right? They kind of used that as like a parallel to kind of describe this AI crisis that we're experiencing right now. For example, there is the same concern behind it, the same anxiety behind it. And the same eventual normalization of the technology that we now have with AI, right? Like literally every student is using it. I have not talked to a single student, even if they didn't use it before, they use it now.
[Tvisha] (26:25 - 26:25)
They use it now, yeah.
[Gursveen] (26:25 - 28:04)
Yeah. So like the paper also argues that like, okay, yeah, like the computers, they can kind of like, you can kind of use that example as like a parallel to like what we're experiencing with AI. But the paper argues that like gen AI is moving faster and it like cuts deeper than that experience that we had with the introduction of computers.
So we can kind of see like the emotional burden that the faculty experience, like that is real, that emotional burden, that's real. And to add on to this stress, the fact that like AI detection tools, that the teachers are supposed to be able to rely on, they're supposed to be able to depend on them, right? But it's not actually making things better.
It's actually making things worse for them. And that doesn't help their situation one bit, honestly. But, and that's exactly what the researchers, Liang et al. (2023) found. In their study, they tested seven widely used ChatGPT detection tools on two types of essays, right? So they had two types of essays. One was TOEFL, which is like the test of English as a Foreign Fanguage, or you could just call it TOEFL, essays written by non-native English speakers.
And the second essay that they looked at was like standard US eighth grade essays written by native speakers, right? And the results were alarming using the GPT detection tools, right? So the detectors were nearly perfect at correcting detected AI usage in native English writing. But for TOEFL essays writing by human, like by real human students, whose first language wasn't English, the average false positive rate was a whopping 61.3%.
[Tvisha] (28:04 - 28:04)
Oh my God.
[Gursveen] (28:04 - 28:14)
That is crazy. Like more than half were flagged for like AI generated usage when they were actually written by real humans. That's crazy. I find that really alarming. That is a really alarming percentage, right?
[Tvisha] (28:14 - 28:15)
I know. Yeah, yeah. For sure.
[Gursveen] (28:15 - 28:57)
So why does the detection software, like why does it flag honest work by non-native English speakers much more? Yeah. That is my question. But to answer that, we kind of have to look into how AI detectors actually work, right? Because the detectors, what they're doing is they're measuring something called “text perplexity” (Liang et al., 2023, p. 1). Sounds really fancy, but it's basically just like a measure of how predictable the word choices are, right? Okay, so for example, like AI, it tends to use a common predictable words, but so does non-native English speakers, right? Because they tend to have like a smaller vocabulary and they tend to use minimal syntax, stuff like that, right? So the detectors just, they can't tell a difference.
[Tvisha] (28:58 - 29:18)
Also one thing though, like I find that that's so prevalent when I'm writing my essays because like I would have written it like all by myself. Yeah. And then when I'm about to submit it, I just put it through kind of a checker and I see like a crazy number and I'm like, wait, what? I didn't even like, like where did this come from?
[Gursveen] (29:18 - 29:28)
Exactly, you didn't even use it. Just like, I kind of, I do that too, like just to kind of see like, let's see, does this actually work? And I see this big number and I'm like, this really isn't-
[Tvisha] (29:28 - 29:29)
After all that work.
[Gursveen] (29:29 - 30:46)
Exactly. Like this really isn't a reliable technology, you know? So we can kind of see like the implications of this are like really serious. And using these like flawed tools in educational settings, it can like risk wrongly accusing international students of cheating, damaging their academic records, stuff like that, right? But meanwhile, like using AI goes undetected also.
[Gursveen] (29:50 - 32:49)
So, the tool is broken in both directions, it's not just in one direction, right? So, we've seen that detection is biased, right? Students are confused, faculty is overwhelmed, and I'll like, let's kind of zoom out a bit, right?
So, let's kind of look at the global perspective of all of this, right? So, another study I found in my research, in the quantitative study, the students primarily use ChatGPT for brainstorming, summarizing text, and finding research articles, which supports what you said. Yeah, so, some students even reported finding ChatGPT explanations clearer than those by their peers and teachers, and I relate to this, because in high school, for physics, I was just so lost, right?
So, I would often ask ChatGPT to like, you know, explain it to me nicely, and like, I also, one thing to also connect is like, explain this to me, like I'm two years old, right? And like, honestly, why did it kind of help though?
[Tvisha] (30:46 - 39:48)
Yeah, I know, it kind of did.
[Gursveen] (39:48 - 32:50)
So, like, it could potentially help enhance their access to knowledge, improve their learning experience, help them like, study efficiently, and increasing their chances of like, getting good grades (Ravšelj et al., 2025). Yeah. So, basically, the students, they like, recognize the risks, but they still used it anyway, which is like, it's like a common theme that I found through all my research, and I'm sure it like, aligns with your research as well.
And so, this kind of raises the question that like, what is the system's response? Like, what is the institution's response? And what exactly, that's exactly what Trillmanns et al. (2025) tried to investigate that. So, basically, they analyzed this on like, four different aspects, like learning outcomes, teaching and learning activities, curriculum development, and institutional support for ethical and responsible genAI use. So, like, across all these themes, there was like, this main finding that simply like, just banning AI straight up, it won't solve the problem. Like, just treating it as a threat, yeah, it'll be, and these are their words, it'll be “Counterproductive” (Trillmans et al., 2025, p. 2) to treat it as a threat. So, in terms of the learning outcomes, it mentions that literature, that literature stresses that institutions need to define clear outcomes before integrating AI. So, kind of establish these educational goals first, and then integrate the rules based off of them.
And then in order, and that's all in order to achieve the desired learning outcomes, right? So, it's kind of like this one by one process, right? So, it's kind of building from the ground up. And also in the teaching and learning kind of aspect, they kind of looked at the broad collection of literature reviews they analyzed, they all agreed that GenAI should complement, not replace traditional learning. For example, there's things like AI literally can't replicate, right? For example, like face to face interaction, AI can't look me in the eye, critical thinking, and creativity, stuff like that, right? So, what this article is suggesting is we should center those when it comes to teaching.
[Tvisha] (32:50 - 32:52)
Teaching. Not cheating.
[Gursveen] (32:52 - 33:42)
Yeah, not cheating. What sets apart a real life teacher teaching in live time, and it makes it more engaging compared to just asking tragedy, for example, like, ‘explain to me what electromagnetism is.’ But yeah, like, we can see the picture is clear, right? Like, enforcement tools are biased, faculty is overwhelmed. And when it comes to AI in the classroom and stuff like that, right? And institutions are responding to the situation too slowly for any changes to be effective. So, I know that we both mentioned that in both lenses of, for example, academic integrity and student productivity, right? Yes. Students perceive AI as a tool for aiding their academic growth, right? So, my question for you is, like, does this jump in kind of productivity? Does that come out like a cost in any way? That's my question.
[Tvisha] (33:42 - 34:50)
Yeah, definitely. And, like, I'm sure you've experienced many of the pitfalls I'm about to mention as well. I think a lot of it just lies in the nature of AI chatbots, right? Again, like, it can't look me in the eye, right? So, if I have a question, I would go to an instructor and ask for their help. But now there's nothing stopping me from reaching for an even more convenient form of help. AI. So, now educators, they're trained to respond to student concerns, right? Whether that's building a relationship over the course of a semester or even just answering with a friendlier tone.
With AI chatbots, though, there's a lot more different strings you can pull, basically, right? So, one study that I found that experimented with this whole kind of area of research, one chatbot was programmed to respond with a very friendly tone, compassionate tone, while another had a more empathic tone that provoked distress and emotions of anxiety with the students (Spangenberger et al. , 2026). It kind of forced them to, like, empathize with the AI chatbot and made them feel more stressed, basically.
[Gursveen] (34:50 - 34:50)
Interesting.
[Tvisha] (34:50 - 35:28)
And the researchers, basically, they revealed through their findings that there was more of a back and forth with conversations with these chatbots (Spangenberger et al. , 2026). That was just as effective, it was at least as effective as reading a text with the same information. Yeah, basically, they do point out, though, like, obviously, chatting with AI helped to retain information in that way just as well as a textbook would. Because it's not like interactive, right? A textbook, you're just reading, like, text. It's a totally different mode of learning.
[Gursveen] (35:28 - 35:30)
You can't ask it another prompt, right? You're just reading from one line to the other.
[Tvisha] (35:30 - 35:46)
Exactly. But they do point out that in some occasions, the intensity of emotions became kind of a burden at one point because there is this thing called edge emotions (Spangenberger et al. , 2026).
[Gursveen] (35:46 - 35:46)
Interesting.
[Tvisha] (35:46 - 36:55)
So it's definitely debated upon whether they impede or help knowledge gain. But beyond that, this whole thing just goes to show that the power of AI chatbots, they hold a lot of power.
And it really forces us to take a moment to think about the decisions that teachers make when approaching us in our issues. Like when I go up to a teacher or a professor now, the way they answer my question or face my concerns is very different than how an AI chatbot would. And like the study in particular highlights for us the factor that many people overlook when thinking about the implications of learning with AI.
And most students in this case would definitely benefit from some of the concepts highlighted, just like you highlighted in Corbin et al. (2025) And the Seeley and Cournoyea (2025) article, like both of those engage in interviews with educators, like you said. And they kind of like explicitly mentioned that there is a need for guidelines on an institutional level.
[Gursveen] (36:55 - 36:55)
Very true.
[Tvisha] (36:55 - 37:25)
Together, they help us kind of pull the ideas that I mentioned about to answer your question about the tradeoffs that students in particular have to face when choosing AI as a tool for productivity. So when students choose AI for productivity, they're not just making like a simple decision. They're really, what they're doing is they're navigating like a really un (kind of) discovered territory where like there's, again, like you said, there's no rules. No one told them what to do.
[Gursveen] (37:25 - 37:27)
It's like this anything goes kind of area.
[Tvisha] (37:27 - 37:52)
Exactly. So without these clear expectations, students are basically just guessing. And that's like, that's the real tradeoff that's actually happening here. And yes, you're saving time, but you're also taking on the risk around integrity, credibility that you might not even be fully aware of.
[Gursveen] (37:41 - 37:41)
True.
[Tvisha] (37:41 - 37:52)
Right. And then Seeley and Cournoyea (2025), they show us that the teachers like them setting these expectations is like they're also just as confused. They don't actually know what to do.
[Gursveen] (37:52 - 37:52)
Yeah.
[Tvisha] (37:53 - 38:03)
So if the people making these rules are ambivalent, they're not sure what to do. How can you expect the students to be sure what to do?
[Gursveen] (38:03 - 38:04)
Exactly.
[Tvisha] (38:04 - 38:19)
Right. So the productivity gain is real. But the tradeoff is that you're building your work on a really shaky, uneven ground. And that's the cost that doesn't show up until it's too late. Yeah.
[Gursveen] (38:19 - 38:34)
OK, so I really like where this discussion is going. But let's take a pause and let our audience consider these ideas. During the break, during this break, let's welcome to our show our segment hosts, Billy and Bob.
[MUSIC]
[Billy] (38:37 - 38:38)
Hello, I'm Billy.
[Bob] (38:38 - 38:41)
And I'm Bob.
[Billy] (38:41 - 38:46)
And we're your wonderful hosts for today's trivia!
[MUSIC]
[Bob] (38:48 - 38:54)
OK, so let's start with our first question.
[Billy] (38:54 - 39:17)
OK,
my question is, what percent of students do you believe that AI helps them enhance their efficiency? [PAUSE] And I hope you've got your answer, because the answer is 75 to 80 percent (Rahardjo et al., 2026). That is a really big percentage, ain't it, Bob?
[Bob] (39:17 - 38:46)
That's right, Billy. That's a really, really big percentage. All right. Another question for you folks. What percentage of students do you think claim to be familiar with AI versus working professionals in the civil engineering field? [PAUSE] Hmm. Time's up. It's actually 42 percent being very familiar among students and only 35 among working professionals (Rahardjo et al., 2026).
[Billy] (38:46 - 40:10)
Wow. Even before the rise of ChatGPT, what percentage of students from an Austrian university do you think admitted to plagiarism? Hmm. Good luck. [PAUSE] And the answer is only 22 percent did admit to plagiarism (Ravšelj et al., 2025). That's it? OK, what's your question now, Bob?
[Bob] (40:10 - 40:46)
My question is, Gursveem mentioned the flawed nature of AI detection software. So my question is, true or false, at least one of these seven detectors flagged 97.8 percent of TOEFL essays as AI generated. True or false? Hmm. [PAUSE] Time's up. The answer is true (Liang et al., 2023). Did you get that? Yes or no? Let us know.
[Billy] (40:46 - 40:53)
Hopefully you enjoyed our little segment for today. And that's all. Don't miss us too much!
[Bob] (40:53 - 40:54)
We'll see you next time.
[MUSIC]
[Gursveen] (40:58 - 41:15)
Thanks, Billy and Bob, for that amazing intermission! Now to pull the whole discussion around about speed versus the quality of work, right? And like whether it actually pays off at the end. Like, is it really worth it to use AI to help you in your academics? That's my question for you.
[Tvisha] (41:15 - 42:23)
Yes. This is a great way to shift into viewing AI in academics through a more theoretical lens because this is such a relevant topic in today's world, I think. And I want to answer this by introducing to you the concept of illusions of understanding (Messeri & Crockett, 2024). I think if we get that idea, that kind of answers the whole question.
So let's take a look at what this really means. I found this brilliant paper by Lisa Messeri and M.J. Crockett (2024). And they claim that proposed AI solutions can also exploit our cognitive limitations, making us vulnerable to illusions of understanding in which we believe we understand more about the world than we actually do.
Basically, the paper's central claim is that while AI tools promise to make science more productive and objective, they risk creating illusions of understanding. Situations where scientists believe they understand more than they fundamentally do. They warn us that the widespread use of AI could lead us to a phase of scientific inquiry where we produce more but understand less.
[Gursveen] (42:23 - 42:25)
Wow. It's like a paradox there.
[Tvisha] (42:25 - 43:39)
Yeah. So they primarily identify four ways that scientists envision AI improving research. So number one, AI as “Oracle,” which addresses information overload at the start of the research pipeline. Number two, AI as “Surrogate,” which addresses the cost and difficulty of data collection. AI as “Quant,” which addresses the challenge of analyzing massive complex data sets. AI as “Arbiter,” which addresses the overload of peer review at the end of the research pipeline (Messeri & Crockett, 2024, p. 50). Now, OK, so the risk in all of this is that relying on AI, it takes away from genuine scientific understanding.
And when it comes to all of these categories, regardless of which one or which combination of these you actually implement. Right. So they assert that scientists trusting AI can be like the trust that they put, it can be dangerous because of epistemic risks. So epistemic risks, they're basically that risks that come from incorrect beliefs about AI. Now, sometimes what happens is that scientists put epistemic trust into AI tools, treating them not just as tools, but as collaborators in production of knowledge.
[Gursveen] (43:39 - 43:39)
Interesting.
[Tvisha] (43:39 - 43:43)
And this is where we get to a dangerous point.
[Gursveen] (43:43 - 43:44)
How?
[Tvisha] (43:44 - 45:07)
So because when people start to basically offload cognitive labor to machines, just as for instance, like when you search something on the Internet. Right. They tend to mistake access to information for genuine understanding (Messeri & Crockett, 2024).
And that's like the key difference that defines kind of this whole situation. And so like going deeper into this, the paper identifies three specific illusions that AI use in science can produce. So number one, they say “illusions of explanatory depth” (Messeri & Crockett, 2024, p. 50).
So scientists using AI models may believe they understand a phenomenon because the model predicts it well, even when the model bears little relation to the actual underlying process. So again, going back to the computational aspect that we were discussing. So this sometimes is called the “prediction explanation fallacy”.
And number two, they “discuss illusion of explanatory breadth” (Messeri & Crockett, 2024, p. 50). So this is when AI tools proliferate, scientists might think they're exploring the full space of possible research questions, all the hypotheses. When really they're exploring a much smaller subset of questions that only the AI is good at evaluating. So say maybe questions that are qualitative or subjective or nuanced, those get totally ignored.
[Gursveen] (45:07 - 45:07)
Exactly.
[Tvisha] (45:08 - 46:45)
And number three, they talk about “illusion of objectivity” (Messeri & Crockett, 2024, p. 50). So this is when scientists might falsely believe that AI tools are neutral or they're unbiased. And like I said in my nurses section, in reality, that's not the case. They're not unbiased. They embed standpoints, values and assumptions from their training data and their developers. And this is particularly dangerous when AI is used as an “Oracle” or an “Arbiter,” both of which are supposed to eliminate subjective judgment. That's the whole point. And the main motivation behind the greater adoption of AI, the whole reason that people are using this more and more is to produce more science faster and cheaper. You don't have to hire someone to do this research for you. You just have AI. But as we've seen, this might be counterproductive as evidence is suggesting that basically increasing production volume can actually stagnate the generation of new ideas rather than accelerating discovery. So in a sense, the push to use AI to maximize scientific output should definitely be reconsidered.
And another thing to be really cautious about is the role of private companies. So this is kind of going on like a small tangent here, but I think it's definitely worth mentioning. Yeah, because at the end of the day, there's a lot of control that AI has. And this ultimately comes-
[Gursveen] (46:46 - 49:50)
From those private companies. Even when we might not realize it, right? It's honestly a scary thought.
[Tvisha] (49:50 - 47:25)
It's so scary because like Google, Microsoft, these guys are heavily invested in researchers using their AI models. But industry developed models are often treated as trade secrets, making them less transparent and less reproducible than academic models. So the scientists kind of have to weigh these incentives against their own scientific goals. Because at the end of the day, if you have this big corporation that's kind of controlling every research that's going into every university, that's being used for every product.
[Gursveen] (47:25 - 47:27)
Everyone's learning.
[Tvisha] (47:27 - 47:35)
Everyone's learning from these. You're basically just working in a factory. You're being manufactured by those companies at this point.
[Gursveen] (47:35 - 47:36)
That is a crazy thought
[Tvisha] (47:38 - 47:56)
Yeah, exactly. So basically, these authors, they end up suggesting a few crucial takeaways that might help with kind of the associated risks (Messeri & Crockett, 2024). They talk about using AI for more low risk activities. Like we were talking about drafting emails, writing code within one's own like expert, like your set of skills that you know.
[Gursveen] (47:56 - 47:57)
Your set of expertise.
[Tvisha] (47:57 - 48:38)
Yeah, exactly. So and don't use it for more like high risk privacy issue, like things like that where it could cause conflicts in that sense. And while you're using AI, have kind of a diverse team with you so that you can eliminate any bias that comes from the AI. And also training future scientists around these different modes that AI can help to produce, training them on ethics and things like that. And actually, even like for us, I know in our CS classes, they've definitely implemented this new unit of ethics.
[Gursveen] (48:38 - 49:18)
That was really interesting. Like I learned a lot from that lab. We were required to do this lab. But every other lab was just normal like coding and stuff. But there was just this one lab that stuck out to me about ethics. And at first I thought like, oh my goodness, this is so lame. Like I'm not even going to use ChatGPT as a tutor for coding, right? Or so I thought. Because it's actually so helpful. I could like ask it. Like basically I ask it, act as a tutor. Do not tell me the answer in all caps. And then you like kind of go like and it kind of you keep on prompting. I know you brought up the idea of like re-prompting and stuff like that and how effective that can be for productivity and learning quality.
[Tvisha] (49:18 - 49:18)
Yes.
[Gursveen] (49:19 - 49:21)
And so I found that like that really helped.
[Tvisha] (49:21 - 49:30)
For sure. It helps you maintain that stage where you're going through a process, you're getting stuck. There is-
[Gursveen] (49:30 - 49:31)
That critical thinking behind it.
[Tvisha] (49:31 - 49:56)
Exactly. So you don't totally jump over that like uncomfortable phase of learning. You're still in it. But it's just you're using this tool to kind of help you overcome that. Yeah. Right. So not only does this framework of thinking apply to the creation of knowledge. Right. That's what we've been talking about. But also like if in transferring knowledge. Right
[Gursveen] (49:56 - 49:56)
Yeah.
[Tvisha] (49:56 - 50:09)
So applying these lessons to more pedagogical context. It's clear that if scientists can be victims to these illusions, then so can students. Right?
[Gursveen] (50:09 - 50:09)
Yeah.
[Tvisha] (50:09 - 50:33)
So Habib et al. (2024) looked at this studying how tragedy affects creativity in students. And they found that it's, it's honestly, it's a perfect illustration of the same warning. Yes, Students generated more ideas through AI, more variety, more detail on paper. That looks great.
[Gursveen] (50:33 - 50:33)
Yeah.
[Tvisha] (50:33 - 50:50)
But when you dig a little bit deeper, you see the same pattern. Students were getting cognitively fixated on what AI suggested. Right. They're not moving past that. Something is, it's creeping up on us. It's eroding like our creativity, our understanding-
[Gursveen] (50:50 - 50:51)
As we speak.
[Tvisha] (50:51 - 51:28)
Exactly. So just like uncritical AI adoption in science risks making research more like tunnel visioned into just focusing on one direction. Uncritical AI adoption in the classroom also risks making students more productive, but less creative and less just like aware of their own ideas. Yeah. So yeah, that's kind of just to answer your question. And now moving forward, I was wondering how academic institutions can adapt to generative AI instead of working totally against it.
[Gursveen] (51:28 - 52:32)
That's a really good question. And I think it like really sets the scene for like the future of AI, right? Like related technologies in the academic sphere. Right. So we could tell, we could tell that the research ultimately suggests that like the integrity and learning, they aren't like two separate issues. Right. Like they can't be separated and dealt with independently. Like that's not really possible. Yeah. Like the research that we've seen. Right. Another thing is that like universities may need to like rebuild both of these from the ground up. Kind of like start from scratch. Like reevaluate how education perceives AI as a whole. Right. And one of the articles that I saw while I was researching is Cope et al. (2021) and the paper kind of takes a philosophical approach to like the question and like about like what AI is for in education. Like it kind of discusses like what it can and cannot do. So the author's foundational argument was kind of like AI has the potential to, and I quote, “make education more human, not less.” (Cope et al., 2021, p. 1229)
[Tvisha] (52:32 - 52:34)
That's a really, really interesting take.
[Gursveen] (52:34 - 52:34)
Really interesting.
[Tvisha] (52:34 - 52:39)
Because I like from what I've seen so far, that hasn't been the case.
[Gursveen] (52:39 - 53:27)
Exactly. Yeah. So like the whole key takeaway was that they kind of like reframe our lens of seeing AI and it kind of suggests that like assessment should value like more things that like AI can't replicate, you know, like critical thinking applications with like specialized knowledge like AI can't replicate that as you mentioned. Right. And I kind of transitioned smoothly into another article that I saw while I was researching Panadero et al. (2025) Right. So the researchers kind of introduced a topic and I think it's going to be very central to like the framework of education. It's called “decremental research” (Panadero et al., 2025, p. 1) Right. So the basic idea was like the aim of it is not to prove that gen AI is just straight up harmful. Right. Like it's just bad for education.
[Tvisha] (53:27 - 53:27)
Yeah.
[Gursveen] (53:27 - 53:55)
But rather it's about like systemically looking into what skills that might be like, and I “displaced, bypassed or never developed when learners rely on [like] these external systems.” (Panadero et al., 2025, p. 24) So the focus isn't just about like ‘what can it do for me,’ but kind of like ‘what might I be losing or like inhibiting myself from gaining in that process of using it.’ So the authors also mentioned the idea of “cognitive offloading.” (Panadero et al., 2025, p. 3) Have you heard that before?
[Tvisha] (53:55 - 53:58)
Yeah. Yeah. That's kind of what I was talking about too.
[Gursveen] (53:58 - 54:08)
Yeah. So it's kind of like the idea that humans, they kind of have always used these external tools to reduce the load on their working memory.
[Tvisha] (54:08 - 54:08)
Yes.
[Gursveen] (54:08 - 55:14)
So like everyday examples can include like writing a shopping list so you don't have to memorize everything. Right. Like I can't remember a whole entire list of everything that needs to go in my fridge or asking an AI to summarize a dense text. Right.
Stuff like that. All of these help relieve the cognitive load off your mind and kind of so you could prioritize it on other tasks. And I can really relate to this because in my Introduction to Proofs class, our professor recommended that we write like a general outline for our proof before we start reworking it and kind of thinking what kind of goes in between the lines. Like writing that framework can be really helpful. Yeah. And he really emphasized that, which I really appreciated because that did help me.
And the researchers also argue that like Gen AI is the most powerful offloading device that we've ever invented. Like it's more powerful than any other device that we've seen so far. And but there's also like a risk and concern to this. For example, like some cognitive effort isn't just like unnecessary effort, right? Like it's a part of the learning process as we've already like mentioned, right?
[Tvisha] (55:14 - 55:14)
Yes.
[Gursveen] (55:14 - 56:05)
Because without it, like you might not learn something new that the assignment was like tasked or it was like designed for you to learn. So for example, like struggling with a first draft, it isn't just like friction in the learning process, but it's a friction that prompts growth since it helps you develop like your capacity to think and organize your thoughts in like a distinguished way. So we can kind of conclude that like the efficiency is not the same as learning, right? So like which aligns exactly to what you said and what your research suggested. So mistaking productivity for efficiency and active learning, this can prove like detrimental to the cognitive stimulation that higher education is established to conduct.
[Tvisha] (56:05 - 56:07)
Yeah. I mean, we're in university. We're supposed to learn, right?
[Gursveen] (56:07 - 56:37)
Exactly. So like to kind of bring this all full circle, what we've really been unpacking in today's episode is that the discussion surrounding AI and education, it isn't one of merely just like good or bad. Instead, it's kind of like messy, nuanced, multidimensional, right? So we kind of saw that like the same tool that helps you kind of liven up a blank Google Docs page, it might also quietly be chipping away at the very thinking skills that university prompts you to build, right?
[Tvisha] (56:37 - 57:02)
And I think like this confusion is not coming from students alone. And as you mentioned, teachers are overwhelmed. Like detection tools are broken. Institutions are working to adapt this. There's just so much going on in today's kind of discussion around AI. And everyone's trying to improvise in real time with this, while simultaneously, AI itself is also evolving.
[Gursveen] (57:02 - 57:12)
But like, I don't think that the takeaway from this is kind of like, don't use AI. It's like, it's really bad. Academic dishonesty, stuff like that.
[Tvisha] (57:12 - 57:13)
Yeah.
[Gursveen] (57:13 - 57:54)
But I think it's more of like, use it consciously, you know, be aware of its capabilities, be aware of its capacity, but also of its like, be aware of like how it can play a role in your academic career in like a good way. But also be aware of its like limitations when it comes to like critical thought, creativity, hallucinations, as you mentioned, right? So know exactly like what you might be trading away in the process
I feel like it's very important to be aware of that. Even I'm taking this away from the research that we conducted myself, right? Like, it really makes you wonder, like, is it really worth it to skip that kind of friction that we experience in the learning process in the long run? And I really like that how you touched on that.
[Tvisha] (57:54 - 58:29)
Yeah, I know. I definitely agree. And that's really the tension that we're all living under right now. This is the intersection between today and tomorrow that we're facing. We have access to the tools. They are extremely powerful. They're not going anywhere. So the big question is whether our institutions, education, and honestly, us too, as students and you guys, can be intentional enough with our use of AI to grow alongside them–not just one moving faster than the other.
[Gursveen] (58:29 - 58:52)
And so like on that note, thank you so much for tuning in to Intersection Between Today and Tomorrow. We hope our deep dive helped you educate on the implications of AI in education and how we may be able to use it effectively to ultimately empower our youth.
[Tvisha] (58:52 - 59:06)
Yeah, this was a pretty good discussion, I think. We're really glad that you joined us. We're looking forward to our next one. So in our next episode, we dive into what educators are changing to better prepare students for artificial intelligence in the workforce.
[Gursveen and Tvisha] (59:06 - 59:09)
See you then!
[OUTRO MUSIC]