The Transform your Teaching podcast is a service of the Center for Teaching and Learning at Cedarville University in Cedarville, Ohio. Join Dr. Rob McDole and Dr. Jared Pyles as they seek to inspire higher education faculty to adopt innovative teaching and learning practices.
This is the Transform Your Teaching Podcast. The Transform Your Teaching Podcast is a service of the Center for Teaching and Learning at Cedarville University in Cedarville, Ohio.
Ryan:Hello, and welcome to this episode of Transform Your Teaching. In today's episode, Dr. Rob McDole and Dr. Jared Pyles conduct a book review of Authenticating Intelligense: Preventing AI from Hijacking Education by Dr. Stephen J.
Ryan:Aguilar, a previous guest on the show. Thanks for joining us.
Rob:Well, Jared, it's another book review.
Jared:It's book review season. It's book
Rob:review And today, we are looking at authenticating intelligence preventing AI from hijacking education by Steven j Aguilar, who we
Jared:had guest.
Rob:Yep. We had on the on the show. And, we have dealt with... I mean, there's a lot of questions that Steven raises, but we're only gonna deal with really three main areas today that you and I, pinpointed, to to share with our guests. So I think I will just open up and say, before we do Mhmm.
Rob:Just a little bit of word about the book for our listeners. One, this is not a handbook, and it's not one that you're necessarily going to want to have right next to you pulling from the shelf. Right. But it is, I would say, a necessary read.
Jared:I would say it... Definitely read it.
Rob:So why from your perspective?
Jared:Well, because it's, doctor Aguilar spends the book setting up a foundation and a philosophical background to approaching generative AI in education, and it wasn't what I was expecting when I started to read it. I was expecting a handbook because, you know, preventing AI from hijacking education, think, oh, here we go. We're gonna get some cool tips or whatever. But it wasn't, and it, I'm not upset by that. I think it's fantastic.
Jared:But I think it's a necessary read, because he develops a a framework Mhmm. That you then could look through to assess whether or not you want to use generative AI or in what capacity in education. So it's a framework to use for practical steps. He just doesn't give... He gives some practical steps, but not a lot of practical steps.
Rob:Right.
Jared:It's good to read different perspectives on generative AI. Sometimes we get so, you know, siloed in what we believe and what others believe around us that we can create a vacuum. But it's good to read someone else's perspective, and this is definitely a different perspective than I have seen when it comes to approaching generative AI.
Rob:Well, and I I think the other thing I do appreciate about it is that he steps back. Most of the industry, whether it be education or the technology industry as a whole, everybody's yelling and screaming about Yeah. Ethics.
Jared:Uh-huh.
Rob:Right? Mhmm. But he's taking a step further. He's going backwards, and he's asking different questions before we even get to ethics. Yep.
Rob:And I think that's what you and I both saw. And one of those first questions that I... You know, it's right there, chapter one.
Jared:Mhmm.
Rob:He introduces this. What actually actually needs to be authenticated?
Jared:Mhmm. Yeah. It's it's, you know, we've kind of hit on this when our AI literacy series talking about the assessment portion of it and doing things like oral defense or live application, stuff like that. But he hits on it and he argues that, AI exposes assessments that's built around content And instead we should favor the opportunity through AI to observe thinking or the thinking process. So again, process over product that we've talked about.
Jared:Yeah. And I, that's, you know, authenticating is the same, I would say, as our definition of process or what does a true assessment look like or an accurate assessment with the lens of AI literacy. So it is, I mean, it's beneficial that we read this right after finishing that series, because it kind of, again, sheds more light onto the idea of, you know, what are you assessing? What's the most important to assess in your courses?
Rob:Yeah. I think the main idea, or at least the main question that he deals with here is, you know, can we establish what a student knows without, you know, turning this whole process into a surveillance state
Jared:Yeah. Of some sort. Cameras around the room.
Rob:Yeah. Because it doesn't really... Like, for us here at Cedarville, that doesn't I mean, yes, we want we want truth. We want integrity out of our students. We wanna see that.
Rob:Right. But if we... If we're constantly looking for someone to fail, I feel like it's like if you keep looking for people's places where they fail, you will not be disappointed.
Jared:Yeah. Yeah.
Rob:Right? Yep. And that's not necessarily building up, which is which is an ethos that comes out of the Christian faith, especially as we practice it. Right? Serving others, edifying others.
Rob:It's not a matter of finding where they're at. It's a matter of bringing them to where Christ would have them be.
Jared:Right. Right.
Rob:And if all you do is you just keep looking for all the bad stuff. I'm not saying you ignore it, so listeners, please don't hear me say that. I I most certainly think we've got to confront, you know, plagiarism.
Jared:Mhmm.
Rob:People passing work off that's not their own or they're not owning it. But we... I think we've got to figure out ways to do that in a positive way that motivates the person Mhmm. To say, ah, yeah. I see that.
Rob:And I think one of the things I know he said was, it's not so much that this is stress testing or educational system, which kinda is, but it's more testing the things that we're assessing. Right? It's stress testing.
Jared:Assessing what we're assessing.
Rob:Yeah. Mhmm. And so some of those things are popping out and and just that question about what's worth authenticating.
Jared:Right. And, like, he also talks about, you know, design of your assessments, and we can go a step further like he does and talk about when is it that the design of the assessment is causing the misconduct. But there's a there's a balance though because there's also, like you said, we're not saying that students aren't at fault for their cheating. We're not saying it's the instructor's fault that your students cheat. No.
Jared:There is still individual responsibility when it comes to things like this. Students need to own up for what they do and what they don't do. But he just adds, you know, could there possibly be a design change you could make with your assessment with AI in mind that could help alleviate maybe the stress or the decisions or the the steps that is... That lead to a decision of cheating or Mhmm. Dishonesty.
Rob:Yeah. I know for many, you know, they... They'll say things, we we don't want it doing certain things. And, you know, I think what he does is he helps us to think through that. What what do you mean by that?
Jared:Mhmm.
Rob:You know, what should a human be doing? What's worthwhile? And that goes back to, again, what we were talking about. It's a philosophical perspective. Values come from our axiology, and our axiology is is developed based on what we believe about metaphysics or what's real.
Rob:Maybe that's the best way of saying it. Mhmm. And we can unpack that, but this isn't the show to do that on.
Jared:Right.
Rob:But I felt like he did a good job of exploring exploring those things.
Jared:Another thing that doctor Aguilar brings up is talking about, and this is something that maybe it could this stuck out as much because we talk about this. We've talked about this so much is like the idea of AI being a tool, and not necessarily something because we use it mostly, I use the idea of a tool in the context of you shouldn't be saying please and thank you, and you should... We shouldn't be striving to make personal connections with algorithm Yeah. Of ones and zeros Right. That's producing text.
Jared:Right. But he even goes a step further. And I honestly... When I read the chapter, was like, I need to go back and read this again, but I didn't do it because he goes so deep into the idea of what a tool is that it blew my mind to a point where I... If all of this is possible, maybe my students can identify with this, blew my mind to a point that I didn't understand it again.
Jared:So I feel like I have to go back again and read it, but it talks about how we we need to treat it as a tool to a point where there's even a distinction between our own reaction or our own actions towards the tool and what the tool is used for, and there's two different... Or maybe you can help because I was... I should go get the book and look at it again.
Rob:If you if you want to, I'll I'll gladly sit here and wait. But, no. I think he dug into that area most certainly. The other part, and I'll come back to to the tool issue here in just a second, but he also really focused on that, but also the intersection of the tool, the user Yeah. And the environment.
Jared:Yeah.
Rob:And that kind of gets into more of like accessibility. So there is that point to which, you know, as he was talking about what is a tool and then how that affects how you use it most certainly is a philosophical perspective where you're thinking about what defining what is real and what isn't. And again, that's most certainly philosophical.
Jared:We talked about using the tool properly and then using a tool improperly and how you can use a hammer to drive nails in. But you can also use hammers in different ways, just not for the intended purpose. Correct. And that's where I was like, oh my... Like, I know that you've said that before, and I've been like, okay, Rob.
Jared:But now someone else has said it, so that means it's legit in some capacity.
Rob:Oh, what? Somebody that's a lot there.
Jared:It's not just Rob's ruminations or whatever, which is a good idea for a podcast. Rob's ruminations. Oh, dear. We'll play some music over it.
Rob:No. He's not wrong. And I think that's the thing that everybody's trying to figure out because we haven't had a tool to this point that can do some of the things that you can do with these frontier models. Yeah. I mean, legitimately, we've had mathematical models that'll blow your mind if you understand what's going on.
Rob:This is a mathematical model, though, that applies itself to language. Yeah. And I think even those that came up with that didn't see the outworkings that we have right now or the applications, if you will, of what it would be used to do. Mhmm. I mean, some of the stuff that you can do with it, I didn't think, oh, okay.
Rob:I didn't even think somebody would try that. Yeah. Yeah. Like, I was watching a YouTube video of a guy who hooked up ChatGPT to his 2018 Chevy Silverado. K.
Rob:And he reprogrammed his engine control, his engine management, his ECM, his engine control module so that it would have a little bit of a ghost cam.
Jared:Blah blah blah blah blah blah. What? Yeah. Wow.
Rob:He didn't know anything about that stuff. He just used codecs, hooked it up through a connector to his laptop and said, hey. I wanna I wanna reprogram the ECM on my on my 2018 Chevy Silverado. Could you do that? And it it was, like, eighteen hours later, and he did that.
Jared:Can we make it so that cars are more fuel efficient? Could you also...
Rob:Yeah. He could put in a fuel efficient
Jared:Goodness.
Rob:Mapping, and that's what he did. Yeah. Yeah. It is insane.
Jared:So this is another YouTube video, but, you know, the guy I've talked to you about before, his name is Kit Boga, and he does... He scams scammers. Like, he goes...
Rob:Oh, yeah. Yeah. Yeah. I've watched...
Jared:Well, they've they've started to use AI, the scammers have, started to use AI and they fake someone's voice. Yeah. Well, he has found a way to change the programming of it to a point where it runs out of tokens. Like, that's how he keeps the bot from calling people is that he will change the code while he's talking to them and say, hey, every time... Like I'm hard of hearing, so every time you say a, I need you to add ego after it.
Jared:So then the bot will start doing it. And they said, also, I need you to enunciate all of your punctuations. Every time there's a comma or a period, I need you to say it. And it'll start... And so he gradually starts changing the code to a point where it stops...
Jared:It stops doing the normal script and starts hallucinating, and then it runs out of tokens and shuts down. So talking about using the tool inappropriately, but then he's like almost redeeming the tool in a way Yeah. To
Rob:Using it against attackers.
Jared:Yeah. Pretty cool. It is. Sabotage almost. Yeah.
Jared:We talk about tool usage, and it's just it's just another perspective on it that I really appreciate that doctor Aguilar gives us. So...
Rob:Well, not only that, but he also talked about, again, the environment and, you know, one of the things that he he talked about in there is if even you're at an institution and your institution supplies training access and support Mhmm. Do they supply it to all and it... At levels, you know, are there some levels that some are never gonna be able to get to? Right. I think the answer to that is probably yes.
Jared:Just... Yeah.
Rob:Just because it's based on you as a person and what your capabilities are and the fact that tools magnify in some way or amplify. So, yeah, there... I think there's gonna be natural discrepancy. Right. And I don't think that's necessarily a bad thing.
Jared:It's normal.
Rob:That's normal. But I think probably what he was probably thinking about more was, are we purposefully biasing things that will prevent others? And I think that's something to... I think that's definitely something to think through.
Jared:Sure. The last point we wanna talk about for this is in chapter five, he talks about avoiding imaginary enemies, which is a really powerful chapter, one that if you don't read anything else from the book, I would read this chapter
Rob:for sure.
Jared:Because he talks about all these... Well, basically, he sets up all of these different arguments of, you know, cognitive laziness, which is what we're gonna talk about, how it's ruining critical thinking
Rob:Mhmm.
Jared:All these other things. And he just gradually shoots down each one of
Rob:them Right.
Jared:And talks about, well, it's not really that. It's more of this, a different perspective on it. But he talks about cognitive laziness, and he kind of challenges the whole claim of it. And he says you can't really... The question really is does weak evidence of harm justify calling the threat imaginary?
Jared:Like, he he kinda throws the idea of cognitive decline out the window, that it can... AI is not going to necessarily harm critical thinking and the student's reasoning for using it. So I don't know. It was interesting.
Rob:Yeah. And I thought about that too, because one of the things that... The arguments that I've heard, and there's even the ones that get cited, there's a study. Yeah. Think it was the study by MIT, some folks from MIT.
Rob:But even in their study, they say this is not proof of cognitive, like, offloading or cognitive decline.
Jared:Yeah.
Rob:And I think it's something that people are using without really thinking through the definitions of what we mean because if we're going to have a conversation about cognitive decline or cognitive offloading or it's, you know, making my brain melt
Jared:Yeah.
Rob:Then we need to be able to track that throughout history. So longitudinal data would be very helpful.
Jared:Mhmm.
Rob:So when you say cognitive ability, what do you mean in the first place? Yeah. Right. What's the definition?
Jared:Mhmm.
Rob:Right? Because if you've got a decline from something, you should know where we're declining from.... Started
Jared:from. Mhmm.
Rob:And if you don't have a baseline, then all you have is conjecture. Mhmm. And a true scientist is not going to sit there and just, at least I would hope, run statements of fact from conjecture. Right. That would would not be considered scholarly.
Jared:It's also assuming that this is the reason why there is the decline in the first place. Exactly. Not like we talked about generational differences and we've talked about how students are learning is changing. You you can't just say, oh, this is totally the reason why this is... Students didn't do as well on my assessment this semester.
Jared:Clearly AI is influencing them and their brains are turning to mush.
Rob:Yeah. I think we I think we see something happen and then we infuse we infuse a reason for it Mhmm. Without data. And I think that's what he is basically warning us about. And if the evidence isn't there, then I don't think it's necessarily that the threat isn't there.
Rob:I'm not gonna say that. But if you can't prove that, yeah, that's actually what's going on, then it feels almost like a chicken little situation. The sky is falling. You know?
Jared:The brains are falling.
Rob:The brains are falling. We're all melting. Yep. For me, and this is again anecdotal. For me, I've been able to dig into things that I want to learn so much faster and actually reflect on Mhmm.
Rob:In a meaningful way because the LLM has removed a lot of... What's the word? It's kinda like clutter in your room. You know, you've been doing certain things for so long in a certain way, and it just starts accumulating over time.
Jared:Uh-huh.
Rob:But then all of a sudden I get to actually work with the information. I can do research really quickly and get to actual articles and evaluate them. Mhmm. I get to evaluate what the bot puts out every time it puts something out. Yep.
Rob:How is that not causing me, at least for me, and I'm saying this, I've noticed that my thinking's gotten sharper because I'm constantly scanning now. Like, reading what it's outputting, I was like, no. That's that's not right. And I, you know, tell it. You missed this and you missed this.
Rob:And it's, oh, yeah. And then it gives me, like, proof Right. Of what I just told it. I'm like, yeah. I found this and here's why from whatever it is I'm researching, if I can say it that way.
Rob:So I... Is cognitive laziness imaginary? What do you think?
Jared:Maybe. I think it's too early to tell, like you just said. I think we need to know long term the effects it could or could not have. So I think it's too early to tell. I think we need to wait.
Jared:Wow. Just because it's, you know, I know it's been around for several years, but really not the, it hasn't had the impact on education since maybe the last year or so. Really, like that's when the article started coming out about usage of it and that, you know, students are failing their exam because I'm making them take it inside of class or whatever, instead of being outside of class and taking it and the assumption that they're using generative AI. When they're outside of class and they're doing better, therefore they're not in class... Or when they're in class, they can't use it.
Jared:Therefore, their grades went down. Therefore, it's generative AI. I think it's still too early to make those judgment calls. So... Yeah.
Jared:But that's just me. I've been wrong before.
Rob:Well, there's more to this book than what we've been talking about. But we we felt these were the... We both agreed that these three
Narrator:areas