Subscribe
Share
Share
Embed
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.
Alright, Rob. It is time for a coffee drop, and I wanted to do a quick coffee drop on a story that I have seen all over social media for the last, couple weeks, and I wanna give a bit of a response to it. And I say this with all, you know, I'm not the expert on this. I just know the discussions that you and I have had about generative AI in education. And, when this story came out, I knew that it would be one of those that the anti generative AI constituents, if you will, would put in their basket of stories to pull from to say that generative AI is ruining education.
Jared:And you and I have both said that, you know, it's not necessarily the solution to all of our problems. Generative AI is certainly not going to do that. But it's also not the cause of all of our problems because if students are going to cheat they're going to use whatever methodology they can get their hands on to cheat regardless if it's generative AI or Quizlet or passing notes in class or anything like that or copying other people's responses. So all that to say, there is a story that's come out from Brown University for an economics professor, I'm not gonna give his name, who believes that a large portion of his class has used generative AI to cheat.
Rob:Probably.
Jared:Unsurprisingly, unsurprisingly, the discussion has been, well, here's proof that AI is ruining education. But as I was reading these stories, there were a lot of questions that I felt were unanswered. And I thought it was one of those jumping off points. Instead of having a dialogue about what's actually happening, it's become, well see, here's it's the I feel like it was more of the confirmation bias for what everyone wanted to see for that. But anyway, more of the story.
Jared:This was a take- home midterm in an economics course that was usually an in- person exam but students wanted to do take- home. They didn't feel comfortable about going to a campus. Brown just had a shooting there and so they felt uncomfortable about it. So the professor made, he made some modifications to allow them to take the midterm at home. The results were the students performed really well, an average around 96 with dozens of perfect scores.
Jared:He became suspicious because his exam, his midterm especially, he made it difficult. Usually makes it very difficult. And for decades, the score was not that high. It was usually, I think it was like around 70% or 65%. Because he was doing a take home, he even made it more difficult, he says, because they could do it at home.
Jared:I'm not sure if it's, this is one of the questions that I have. I'm not sure if whether or not it was closed book, open book, open note, anything like that. We don't know many of the details of it. But anyway, he's like, this seems strange that it was so incredibly high. These scores were, he became suspicious.
Jared:And he noticed also noticed that the solutions, basically the responses that the students gave were similar to ChatGPT outputs. They resembled ChatGPT outputs. Yeah. So he decided to change the final to an in person exam and told students that it would hopefully validate the midterm. But as a result, once he said it was in person, many students dropped the course and the average on the final was a huge just drop off dramatically.
Jared:The grade difference was astronomical. So the professor was like, well, clearly they used generative AI on the midterm. But we don't really have proof that the students used AI. We have suspicions, we have assumptions, we have evidence that could possibly lead to it, but all we really have are patterns the professor found suspicious, a dramatic difference in the midterm and final, and that many students not necessarily had identical work, but had the same ideas that would resemble ChatGPT responses. So the jump has been, oh, it's clearly AI, so it's destroying education.
Jared:And this became, "oh, students can't don't know how to learn and kids don't know how to read." All the comments I've read are just absolutely like, "get your pitchforks, let's go" kind of a scenario, kind of a thing.
Rob:Storm the castle.
Jared:But so this is where like, don't and I'm not I know students are using generative AI to cheat, but we don't have all the details of this. We're running off of assumptions for this story. And my other thought was like, you're bragging about how difficult your midterm is, but then if a student used generative AI and they could get a 100% on your assessment, what exactly is your assessment measuring in the first place? I mean students are still responsible for their academic integrity. It's but, this is where I struggle because I don't wanna, just jump to this conclusion, this assumption that it's an immediate cheating, students use generative AI, blah, blah, blah.
Jared:And the evidence is that there are repeated patterns, It was a take home. And then once we did the final students drop the course because they didn't because they didn't wanna be they would didn't wanna fail the there's a lot more details I'm leaving out. But like, know, if if he said if the if the final grade distribution was the same as the midterm, he would keep the midterm. But if it was different, he dropped the midterm and only take the final as the major percentage of the course grade, essentially setting students up to fail if they actually did cheat and then like they didn't do well on the final exam. I mean, that's another factor that students may have dropped also because they were uncomfortable with the shooting on campus.
Jared:Maybe that's why they dropped when he said it was in person, but the immediate jump to me has been, oh, they use AI.
Rob:Well, think, you know, if you can, let's just go with that. Let's just assume the vast majority of the students use ChatGPT. What do you do from where do you go from there? I think that's the better conversation. Like, why are they using it?
Jared:Yep.
Rob:What is it? I think you were on the right path. Like, what what's the purpose of the measurement in the first place? In terms of the assessment, a midterm and a final, what is it actually assessing? And if they did choose to use ChatGPT to pass that assessment, what are the value propositions that they're accepting or that they're identifying that would lead to that behavior because there's no behavior that doesn't come from some sort of value judgment.
Jared:Right.
Rob:At least consciously, I would say, I would argue. And students are making a conscience a conscious choice like, oh, they're seeing an easy path. Why does that surprise anyone? Yeah. Like, why would you be surprised by that?
Rob:Yeah. I mean, they've been doing it just because I think the only thing that you could say here is that ChatGPT has facilitated this so much faster and easier for others but that's not what's quote unquote killing education.
Jared:Right.
Rob:It's not about whether they actually cheated or not it's about the value judgment that they've already made. Mhmm. And if they've already made that value judgment, regardless of whether they were successful in cheating or not, that's the problem.
Jared:Right.
Rob:But yet we're getting up in arms about, oh, look at all these people cheating We're blaming technology as if somehow technology is the thing that's done this. When you and I have talked about this so many times, technology is just a tool.
Jared:Right.
Rob:And it's the people behind the technology and how they use the technology that makes the difference. And in this case, students have found a way to use the technology to take a shortcut, get ahead, and it's because they've made a value judgment that all they need is a piece of paper, not an education because life has told them something different than what we're trying to tell them as educators. We're saying, "well, you have to know all these things." Well, then they turn around and they look at the world and they look at how people succeed in life and they go, "the two don't match up."
Jared:Yeah. Right.
Rob:Right?
Jared:Mhmm.
Rob:And so we're kinda at a disadvantage at that point, especially if if it's a, you know, a worldview in the secular realm that doesn't see your actions as reflection of the God in whose image you were created. So Yep. I mean, it it yeah. I get where you're coming from, and I think you're exactly right. It's like, why are we getting upset about this?
Rob:This seems like it seems like much ado about nothing in some respects, but in another respect, it isn't. Like, in another respect, it's opening up this thing. And I think what it seems like you're saying is what you're frustrated about is you're drawing the wrong conclusions.
Jared:Right. Yep.
Rob:And we're not dealing --we're dealing with the symptom. We're not dealing with the root of the problem.
Jared:I'll be curious to see where this ends up. I'm also curious to hear from our listeners if they're willing to contribute and what do you think about this story? Does it convince you that AI is the problem or does it suggest that assessment seem to evolve and education needs to evolve or just like you said Rob, with the whole, you know, they've already made a determination of what they value. So maybe that's where the bigger question is for that. But I just wanted to do a coffee drop on this and rant a little bit.
Jared:So thanks for letting me do that.
Rob:Not a problem.