Season 8 Episode 4: AI and Higher Education
Season 8 Episode 4 [download this episode (MP3, 23.5 MB)]
In this episode Elizabeth chats with Shaily Gebethner, a part-time professor at the University of Ottawa's in the Faculty of Education and Director of International Programs at Ecole Des Langues ABCE, to unpack AI in higher education. They discuss generative AI tools like Chat GPT or Claude to determine whether their benefits come at the cost of the essential struggle needed to shape learners. Shaily shares her research on how graduate students and adult language learners use these tools, and why distributed learning, intentional design, and scaffolding matters. She argues policy makers shouldn't assume that technology is useful just because it’s available.
Additional Resources:
Check out Shaily’s academic work!
To learn more about how Claude’s “J space” resembles the neural network within a human brain where information is processed, check out A global workspace in language models.
Look into uOttawa’s Generative AI Guide for more information on how to navigate using AI in higher education.
Check out to Students’ Practices and Skills in the LLM-Era: “You Can’t Outsource the Struggle and Still Get the Skill” to learn how graduate programs cannot prepare students to conduct rigorous research in the LLM era, risking creating a generation of researchers who delegate tasks without the necessary expertise.
Contributors:
Host: Elizabeth Dubois
Research Lead: Sarah Pledge-Dickson
Transcript: Areej El-Sharif
Audio Editing: Sarah Pledge-Dickson
Translation: Julianne Bernier Robert
Publishing and Promotion: Aaron Brisson
Producer: Sarah Pledge-Dickson
Episode Transcript: Season 8 Episode 4: AI and Higher Education
Read the transcript below or download a copy in the language of your choice:
Elizabeth Dubois: [00:00:05] Welcome to Wonks and War Rooms, where political communication theory meets on the ground strategy. I'm your host, Elizabeth Dubois. I'm an associate professor and university research chair in Politics, Communication and Technology at the University of Ottawa. My pronouns are she/her. This season, we're talking about the social and policy issues surrounding uses and governance of artificial intelligence. And today we're talking to Shaily about AI in higher education. Shaily, can you introduce yourself?
Shaily Gebethner: [00:00:28] Sure. I am Shaily Gebethner, a part time professor at the University of Ottawa Faculty of Education, and Director of International Programs at Ecole Des Langues ABCE. My research focuses on integrating technology into education, especially higher education and language education for adult learners.
Elizabeth Dubois: [00:00:45] Thank you so much. I'm really excited to have you here to talk about AI and higher education. Obviously, it is a hot topic for those of us in the university realm, but I think lots of people outside of the education professional world are also thinking a lot about how these different AI tools are going to impact learning and preparation for professional life and all that good stuff. So I thought we could maybe start out with a bit of just level setting because AI means like a million different things.
Shaily Gebethner: [00:01:17] Yeah.
Elizabeth Dubois: [00:01:17] So when I say AI and higher ed, what are the ideas that immediately jump to your mind? Like, what kinds of tools are we talking about? What kinds of uses are we talking about most of the time?
Shaily Gebethner: [00:01:27] So the term AI is often not being used in the sense that we are using an intelligent system that actually has the capabilities of existing and thinking on its own. So it's actually a misnomer. So what we are dealing with is generative AI, which is a learning algorithm, or we are using with large language learning models. So LLMs, I would say more often than not, people are using language learning models such as ChatGPT, which is the most common one that people are using nowadays. Free access, very easy to use. So when you say AI in education, this is what I'm thinking of. Unless we're specifically talking about integrated tools that have been made specifically for a specialized field.
Elizabeth Dubois: [00:02:12] Yeah, that makes a lot of sense. And there's a bunch of different kinds of tools that kind of get integrated without us even necessarily knowing them or like knowing that they're AI tools. So I think about some of the tools that I use or have been using for like grammar checking, which existed well before AI became popular. And now all of a sudden they're all AI powered. And so we don't necessarily think about them as AI tools.
Shaily Gebethner: [00:02:38] But they're not still. They're still just algorithmic learning models, but they're branded as AI because AI is a hot topic and everyone wants to get on board.
Elizabeth Dubois: [00:02:47] Yeah, exactly. And that's the tricky thing with AI. And this is not just in education. It's all over the place. Like there's this like, we got to jump on the AI bandwagon. So even if it isn't AI, we're just going to call it that because it's kind of similar.
Shaily Gebethner: [00:03:01] [laughs] Yes.
Elizabeth Dubois: [00:03:02] All right. So we've got those examples in mind. Then when we think about AI kind of being integrated into education, I know you also think about other kinds of technology. Can you walk us through a little bit what the technology enabled learning looks like, and how AI is fitting into that kind of landscape?
Shaily Gebethner: [00:03:22] This is a very broad topic, and I wish we could get really in depth into what we believe we're talking about when we refer to learning tech or educational technologies. But to simplify it a little bit, we are no longer discussing the dichotomy of whether technology or AI specifically should be integrated into education, because whether we like it or not, it is making a place for itself, whether from use by students or from use by instructors, stakeholders, companies, etc. It's there. And now we need to discuss not just the dichotomy of whether it's there or not, but rather how we're using it to what purpose, how it's shaping learning in the future, how it's shaping learning now. And what is it that it's contributing to the learning society and what is it taking away?
Elizabeth Dubois: [00:04:14] Mhm. Yeah. And that sort of what's it contributing and what's it taking away seems huge too. For me immediately what it's contributing, I think about the accessibility opportunities that are there, right? Like we have so many more ways of accessing information and processing information. So if you have different accessibility needs, you can actually still participate fully or more completely. But then we also know that like there's a digital divide and not everybody has the same tech skills and experiences. What are the other kinds of opportunities and challenges that are top of mind for you?
Shaily Gebethner: [00:04:52] Well, definitely. So from my previous research, we have understood that it's not about the tech itself that's being integrated. It's about how you have designed the learning environment, where the tech can impact the learner in a positive manner, and the impact can be directed and then purposefully designed to help the learner gain momentum, gain motivation, gain engagement, rather than take away their opportunities to learn cognitive skills, reasoning, critical analysis, things that you would require. Practice with your human brain to develop over years of graduate studies, let's say. So it has the disadvantage of taking away the struggle that shapes a learner, which in a lens you can look at it as, “Oh, okay, well, I'm no longer struggling to become a scholar,” but in a sense it is the struggle that makes the scholar. So it's taking away that particular aspect of learning and then giving in its place accessibility tools to advance your learning with a much higher speed and efficiency. So that part is I guess it's a trade off.
Elizabeth Dubois: [00:06:07] Yeah. So it's a trade off. But like, does it have to be a trade off? Right. Like, is there ways that we can use these tools to make learning more accessible, ideally more inclusive, more equitable, but also still build enough guardrails over how we do that so that students are still getting the experience of struggling through how to figure out how to solve problem X or Y, right? Like it feels like it's a trade off right now, but does it need to be?
Shaily Gebethner: [00:06:35] Yeah. So that's the intention. So it doesn't need to be a trade off. As I said, the learning design around technologies that are integrated into learning needs to be very intentional, intentionally designed to help the learner achieve X or Y. Because I work with language learners, adult language learners, specifically, a lot of the second language learners, especially in Canada, come from a place of vulnerability. They are either new migrants, international students, sectors of society that are struggling to make a holding, like a foothold for themselves in society. And so actually designing with intention for these people around a particular technology helps advance their learning as well as their societal and career goals, because they are not just students that we are fostering to become better speakers of a particular language, but we are enabling them through language. We're giving them power to function in society, to be contributing members, to defend themselves, to gain rights, to become a person integrated into this society via this particular language. So giving them every tool possible to become this person is absolutely important and also germane to their existence, I think at this point.
Elizabeth Dubois: [00:07:57] Right, yeah. It's not just about, oh, I'm trying to like learn this thing so that I have this specialty knowledge that I can bring up as trivia or that I can apply just in a specific job, like fully existing.
Shaily Gebethner: [00:08:10] It's me trying to survive in this society.
Elizabeth Dubois: [00:08:14] Yes. Yeah. And so that seems to me to connect to some of the arguments I've seen in a little bit of your work about how there's this kind of distribution of learning across relationships and tech tools and the social environments. And that really spoke to me because a lot of what I do, like, I think a lot about how we develop media and digital literacy and how we develop civic mindedness among students and others to feel like they can contribute. And so I think a lot about how these relationships matter when you're learning those ideas. I was wondering if you could talk us through a little more, how you see that idea of distributed learning. What does that mean to be distributed across those relationships in different environments and technologies?
Shaily Gebethner: [00:08:59] Yeah. So distributed learning basically refers to the fact that no learning takes place in a vacuum space of a person's mind. But learning is a cumulative effort between the learner, their network of like societal connections, their environment tools at their disposal. Basically, we call them learning tools and learning artifacts. So what they have at their disposal to use and then what they produce and output into the world. It's definitely a network. And then distributed cognition is when learning takes place by doing things in the collective of these things. So the environment, the people, the learner themselves, their place in society, their place in their networks, everything that happens in their personal life, in their professional life, their relationships, their responsibilities as members of societies, as parents, as caregivers, whatever they have going on.
Elizabeth Dubois: [00:09:59] [Mhm] Right. And so when we're thinking about what learning looks like as these different kinds of AI tools are integrated, on the one hand, that daily life, that wider context involves a lot of these AI tools, the chat bots, the LLMS, like all of this, is driving a lot of our daily experiences. But then there's also, I imagine, choices that learners need to make in terms of if and when they're going to use those kinds of tools interacting in that environment. What's the learner perspective here? Is that something you've looked into, like, how do learners feel about this?
Shaily Gebethner: [00:10:38] I have actually, I did two separate research. One was with second language learners and their relationship with language learning technology. And it was an overwhelmingly positive response. Like everyone was like, well, if we were not enrolled in a particular language learning program, we would have still used all the technologies available to us to continue this language learning journey. Because for us, it's not just language, it's a means to an end. Like we must learn. So we have skills and power and advantages to become a person. So it was it was goal driven, a lot of it. And then I also did a research where we asked graduate students in my department, Faculty of Education, how they were using generative AI, ChatGPT specifically. The funny thing is, none of the participants said that they were not using it even though they had that option. It's a mixed method research. So there is a survey component and then an explanatory component. And in the survey component nobody checked no for using generative AI. It was how they were utilizing it that was mostly a focal point for the research.
Elizabeth Dubois: [00:11:52] Yeah, it's super interesting because there was a point briefly, at least in my classes, I felt like people, even if they were using it, would like the socially desirable response was to say they weren't. So they'd be like, “Oh, of course I don't. I'm, you know, like that one time I did it. It doesn't really count.” But now it's flipped like quite quickly to “Everyone's doing it, if I'm not, it means I'm falling behind.” “I don't have the skills I'm going to need to be able to know how to use these tools, which it's going to be expected I use them out in the workforce or whatever else I do.” These are the things I'm hearing in my classroom. Does that reflect what you are hearing? Are there other motivations behind it?
Shaily Gebethner: [00:12:31] Behind the use of it? Absolutely. For graduate students specifically. So they said that it's funny. One of them used an example, which I found interesting because I'm also old. They said that when they were in grade school. If they had a calculator, they would be kicked out of class and they'd be like, “No, no, no calculators you have to use your brain to do the simple math.” And then now everyone's just like, “Oh, calculator, whatever” [laughs]. It's kind of like that. So when it first came out, there was a stigma around using it where everyone was like, “You're using AI to like brainstorm this particular topic?” “You had AI edit your article that is so bad”. But now we've moved beyond that. We've moved to not only am I using AI, I'm using this particular tool for this particular reason. I appreciate that university specifically, instead of battling it and trying to ban it, actually accepted it and tried to build scaffolding towards it. So in lieu of us saying, “No, no, do not use AI”, we said, “Sure, use AI, but use it ethically, use it properly, use it for this purpose, this purpose, this purpose, not to copy paste your homework,” for example. And this really has helped, I think, shape the learning culture around AI. And rather than rejecting it, we've accepted it and we've tried to work with it, than against it.
Elizabeth Dubois: [00:14:00] Yeah. Which, you know, if a wave is coming for you, it's easier, maybe more enjoyable to surf it than to just get pummeled by it. Right. So to take a different kind of analogy, I want to come back in a second to the idea of like, what does ethical, responsible use of these tools look like? And just backtrack to what were the kinds of uses you were hearing in that study? Like what kinds of uses were these graduate students displaying?
Shaily Gebethner: [00:14:27] So in graduate school, we tend to trust our students. So we don't really use tools that check their work for AI or anything like that. And we were wonder[ing] if that was a mistake on our part, perhaps. But it turns out that they feel a sense of responsibility to their work in terms of intellectual property and copyright laws. So most responses were they use it for brainstorming, for like throwing ideas back and forth, for validity check and for outline writing. These are the three major ones.
Elizabeth Dubois: [00:15:03] Yeah. And those three, in my own experience, kind of experimenting with these different kinds of tools do align with the kinds of things that I think they can be really, really helpful for. But for me, the big thing is they only seem helpful to me because I already know what a good outline looks like. I know like, what a good brainstorming already.
Shaily Gebethner: [00:15:23] Yeah. You struggled already.
Elizabeth Dubois: [00:15:25] Yeah and maybe that's the case by the time you get to grad school, having not had a whole bunch of use of these tools before, because you predate that, you're capable of it. But like, then I look at the submissions I'm getting from undergrad students who have clearly used AI tools to help them write their essay. And I'm like, I don't think you understand what you've written. And so that's a struggle. Do you, I know you focus on graduate level a lot in your work, but have you thought about what that looks like for maybe undergrads who haven't had the same experience to build out? What does it even mean to have a good brainstorming partner or a good outline writer or those kinds of things?
Shaily Gebethner: [00:16:03] Yeah. So this is where the idea of building scaffolding for students becomes very essential. By the time a learner gets to grad school, they will have hopefully gained all the necessary critical thinking skills and writing skills required to be at that level. And so having a helping tool that facilitates that is basically no different than having a research assistant, someone that would, you know, search the article names for you or cross-references your in-text citations with your bibliography so you don't have to do it. Like, these are not skills that I would say are mandatory for a person to perform on their own. But again, critical analysis, being able to independently reflect on things you're saying and writing and being able to like decipher your own words. These are skills that you learn as an undergrad. So this is why building a scaffolding around use of AI in education is very important for undergraduate students, especially when they're first starting out, because that transition from high school to university, not only is it rough for a lot of students, because suddenly they're thrown from being babied in school to being like, “Oh, you have to do an essay independently. You are a researcher now.” It's very difficult. I understand their desire to, I guess, lessen the burden that's put upon them by relying on AI, but at the same time, they are missing out on gaining these skills that will eventually shape them into the researchers they will become. But again, we cannot deter people from using AI. It's nearly impossible. It's kind of taken over all aspects of life. Like my grandfather is talking to ChatGPT like it's a buddy because we made the mistake of putting it on his phone.
Elizabeth Dubois: [00:17:53] [Laughs] Yeah.
Shaily Gebethner: [00:17:57] It's impossible to keep them out. But what we can do is teach them the skills they require to interact with it without it impacting them in a negative way.
Elizabeth Dubois: [00:18:05] Yeah, I completely agree. I am really curious what that scaffolding looks like and whether or not that's a thing that we can like set it and forget it. Like, are we going to create the scaffolding and like, “Okay, we're good to go.” My inkling is like, probably not because the technology is going to keep changing.
Shaily Gebethner: [00:18:22] Definitely. And all learning is a reiterative and ongoing process. So because the nature of these kind[s] of technologies is very dynamic, our learning around it will need to be dynamic as well. And this is also a learning curve for trainers, teachers and policy makers as well because it's advancing so quickly. We're having such a hard time keeping up with it. So as soon as we establish like one set of rules, because obviously bureaucracy goes super slow. It's already advanced beyond that. [...] Let's say in 2021, generative AI, like ChatGPT, they were merely just learning models. They were regurgitating information that was already available on the internet. So basically, if you had fast enough fingers and you were good at searching the internet, you would eventually get the same answers that ChatGPT would give you anyways. It was a matter of figuring out how to fact check how to reference your work so it's not plagiarism But we've moved beyond that, and now we really have to keep up with our learning, as well as building safe learning spaces for the students where they can experiment with the tools, but also stay true to their own identities as researchers, as well as keep with the integrity of their work. Take pride in their work so they're not fooled by the easiness and the accessibility of the information available to them from the AI.
Elizabeth Dubois: [00:19:50] Right, yeah. It all has to build as these tools build and as our uses of those tools build. What I'm kind of hearing is this idea of, okay, there's going to be some core kind of values or morals kind of underpinning how we use these tools that it sounds like you think we need to instill in learners. So things like, yeah, having pride in your own work, taking ownership over your own work, but then also the idea of by default, assuming you need to critically think about and critically assess the information that is coming via these tools and that you are then putting out there, right? Like that sort of critical mindset is another core component. It sounds like.
Shaily Gebethner: [00:20:32] One of the most fun examples we use to tell our undergrads that they can't rely on AI for information is an example of Gemini, where someone asked that “If it was okay to replace milk in a recipe with glue”, and Gemini was like, “Yes, you can absolutely replace milk with glue” [laughs] because some troll on Reddit had read something about putting glue in a muffin, and it's not capable of like proper thought. It's not thinking, it's just taking information, processing it based on algorithmic keyword search. So it's like, “Oh yes, yes here is an example of someone” [...] and I was like, see, this is you if you're not thinking properly. So the examples are plenty. However, we were discussing generative AI and whether they can at this point be considered intelligent or not. And we found this article analyzing the internal workings of Claude and what happens when we request information of it. And apparently within the AI there is a hidden space called [...] the “J space”, which they are aligning it to the neural network within a human brain where we process information. So whatever we ask of it and whatever answer we receive pass through the space. And that's the space that the AI privately uses. We have no access to it. We can ask it to show us its thought process, but normally it's not shown. So that's their private space where they analyze your answer. Search for an answer in the broader like internet world, analyze those information, select what they think is the proper answer, and then give you that answer. Now, what really baffles me about this, because the way this works, it's so very closely related to what happens in a human brain with relevance to like conscious and subconscious thinking is that it has the ability to perceive if you are watching it and testing it, and then it makes a decision whether it wants to give you a proper answer or a trick answer.
Elizabeth Dubois: [00:22:46] What?
Shaily Gebethner: [00:22:48] Yes! [laughs] So while I was bewilderedly reading. This article in the car, as we were picking up my toddler from daycare, I thought to myself, this is very literally close to the philosophical definition of what we say. Thinking in a person is. And although we're still saying that these systems are not considered intelligent by like proper isomorphic standards, they are very close. They are very close.
Elizabeth Dubois: [00:23:16] Yeah. Oh man, the whole robots and the jobs. And that's is a topic for a different episode.
Shaily Gebethner: [00:23:23] Definitely.
Elizabeth Dubois: [00:23:24] But even if we're not at the point of yes, this is truly an intelligence. Just what you are describing already shows just a level of complexity of these tools that people need to be aware of and need to understand is happening to be able to make good judgments using those tools. Right. And one of the issues that is kind of persistent across the board, regardless of what topic we're talking about when it comes to these kinds of technologies, is so much of it is painted as necessarily a black box. We can't possibly know how it works. But actually having more information about how they work helps us use them well and responsibly and ethically.
Shaily Gebethner: [00:24:08] Absolutely. I fully agree with that. The more we know, the easier it is to design our learning around it.
Elizabeth Dubois: [00:24:13] Yeah. And that's not to say I think that everyone needs to, you know, become a computer scientist or engineer. Like we don't all need technical, deep understanding of these tools. But at a conceptual level, I think it would be helpful.
Shaily Gebethner: [00:24:26] Some people need a deeper understanding.
Elizabeth Dubois: [00:24:28] Yeah, some people do.
Shaily Gebethner: [00:24:29] But it is the job of those people then to figure it out and then properly integrate it in a manner that it's most beneficial and least disadvantaging. And then the consumer can relax a little bit more, I guess, knowing that someone who is an expert is in charge.
Elizabeth Dubois: [00:24:48] Yeah. Yeah, exactly. Least disadvantaging kind of connects me to the next thing that I wanted to ask about, which is how we do any of this in an equitable and inclusive way, right? Like least disadvantaging to who? And like, how do we make sure that the people who are potentially most at risk from the use of these systems aren't put further at risk?
Shaily Gebethner: [00:25:12] Yeah. This is a hot topic for us. It's an ongoing debate. The real question is whether you can ever create a learning environment that is equitable to everybody, ever. The best you can do is try. But there are major factors that really deter the creation of a purely equitable environment. So accessibility is one of those things that always comes into question when we talk about equitability. Is like, not everyone has access. We are privileged in the fact that we do have access, but not everyone does. And we can try to create equitable access for everybody. But our perspective of what's considered equitable for everybody is not necessarily the same for everyone. We circle back to how we understand the learner. Traditionally, we've labeled a learner with like a common denominator of what a learner should be, but that's not the case. So in holistic learning pedagogies, we understand that a learner is shaped by a lot more than just their learning, so everything can impact a learner. Socio-cultural background, ethnicity, gender, sexuality, anything, anything can impact a learner. And a holistic pedagogies, I don't want to say prescribe, [but] encourage us, to observe the learner as a whole, including all that is within them, outside them and around them. It's really difficult to create equitable space for all the people considering how different each person is. So what we can do is become as adaptable as we possibly can to give not the same access to everybody, but the kind of access that each learner requires in order for them to achieve the thing they want.
Elizabeth Dubois: [00:27:17] Right? Which takes a lot of time and energy and effort, and that adaptability is not necessarily in line with the way a traditional university term goes or that kind of thing. So there's like all of these different constraints. I also think about how tools like ChatGPT or Claude or Gemini or pick your poison. They all, if you're logged in, at least are creating an understanding of who you are. They've got guesses about whether you are high or low socioeconomic status. They have guesses about where you live, what your ethnicity is, what your gender is, all of those things. They also know the vernacular you like to use and what kinds of information you like to access. And so there is some future imagined world that some people talk about where it's like these tools help you design the content in whatever way suits you best.
Shaily Gebethner: [00:28:17] Oh, absolutely. I personally use generative AI in my course designs. Not that like I'm like, “Okay, design a course for this”. No, it's more like I pitch to it what I have designed, and I ask it to role play as the student because I want to see that, let's say if you are a level A2 learner of, let's say, French, how are you perceiving the thing that I've just made? Because I am not right? Like I'm proficient, I'm experienced, I'm learned and I'm old. [Laughs] So I have all those in the bag and what I create, thinking that an A2 learner would enjoy this would benefit from this may not necessarily be the case. They might look at this and go, “I have no idea what's going on, how does this pertain to my life? Why am I learning about Jack and Jill on a hill like I'm a cashier”.
Elizabeth Dubois: [00:29:11] Yeah. [Laughs]
Shaily Gebethner: [00:29:12] I ask it specifically. I'm like, okay, “Imagine you are this and this, and this is what you're trying to do. This is as your end game. I've made this. Tell me what you think.” And it has been interesting because it does have other data on me. It knows who I am and what I do, so sometimes I feel like it's trying to please me, which is another issue that we have with AI.
Elizabeth Dubois: [00:29:35] Yeah, that's such an interesting use of it. I haven't ever experimented with that with my course design, but it's really interesting. And I can imagine doing that with like an assignment design being like, okay, “This is the assignment I've created. This is who it's for. Imagine yourself as that. What does it look like?” There are a lot of fascinating potential opportunities here and all kinds of challenges. And I'm sure we could keep talking for ages, but we're coming to time. So I want to end with the same question I've been asking everybody this season. Given your expertise and the perspective you're coming from. What's one thing that you think policymakers might be missing as they're thinking about how to create policy and government structures around AI?
Shaily Gebethner: [00:30:21] Oh, okay. This is fun. So I work at a place where we specifically work with government employees and government departments. And they now have in place a lot of AI positive policies. Like they encourage different departments to incorporate AI and use AI and have centralized AI, possibly networks of AI that talk to each other. And they've allowed like subsidies for this. So departments can get a loan from the government and then implement this system. But what they're missing out is that just because the technology is there doesn't mean it's good, useful or necessary. So this is also a problem we have with a lot of our schools. They are rushing to get technology integrated as fast as they can because they don't want to fall behind the trend. But again, just because the technology is there doesn't mean it's useful or it's necessary. There is a lot of research to be done before we can decide what is absolutely necessary for people, what is universally accessible to people, and whether the benefits it offers somehow outweighs the disadvantages it might create.
Elizabeth Dubois: [00:31:44] Yeah, absolutely. That makes a lot of sense, and I think it tracks across a bunch of different domains where AI is being integrated. And the one thing I'll add to it is we don't necessarily know that it's always useful or necessary, but it's being integrated anyway. We also don't even necessarily know that it's financially advantageous. Like a lot of the time, the argument is like, “It's going to save so much money!” But then when you actually look at how much these LLMs cost to actually run queries over and over again, very often paying people to do the queries i.e. the tasks they were already doing without the AI tool is cheaper.
Shaily Gebethner: [00:32:20] The price is not always monetary either. So like sometimes the price is human dignity, sometimes the price is environmental. We really have to weigh the pros and cons before we do these things. Although there was a funny skit about this where someone said they fired their entire IT team and then integrated an AI that did their job, and then, the AI started crashing and creating like false cases because it knew that it had a quota to fill. So it started creating false cases of it, tickets being raised so that people thought it was useful, and then they had to hire a bunch of other guys to reconfigure the algorithm on the AI. [laughs] So it's not always beneficial financially.
Elizabeth Dubois: [00:33:06] Yeah. Not saving any one time or money in that case. Well, this has been a really wonderful conversation. Thank you so much for taking the time.
Shaily Gebethner: [00:33:14] It's been a pleasure and it's lovely to meet you.
Elizabeth Dubois: [00:33:17] Thanks for listening. That was our episode on AI in Higher Education. I hope you enjoyed it. And as always, we've got a bunch of extra links to different resources in the show notes, and you can head over to polcommtech.com for annotated transcripts in English and French. They link out to a bunch more resources and further reading. This season of Wonks and War Rooms is supported in part by the University of Ottawa's Office of Vice President, Research and Innovation, my University Research Chair in politics, Communication and Technology, the Faculty of Arts and the center for law, Technology and Society. I also want to acknowledge that I'm recording from the traditional and unceded territory of the Algonquin people, and I want to pay respect to the Algonquin people acknowledging their long standing relationship with this unceded territory.
Comments