MIT has long been known for its quirky student culture, where collaborating on problem sets is a frequent form of socializing. But a new internal report warns that AI is changing all that, as more students turn to chatbots rather than classmates when they’re stuck. The report, which is sparking discussion across higher ed and even in some media reports, argues that this esteemed university needs to reboot all aspects of teaching and assessment now that AI can complete just about any assignment a professor can hand out. Can AI be part of the answer to preserving social learning at universities?
A new internal report at MIT warns that AI is making learning less social, as more students turn to chatbots rather than classmates when they’re stuck. The report, which is sparking discussion across higher ed and even in some media reports, argues that this esteemed university needs to reboot all aspects of teaching and assessment now that AI can complete just about any assignment a professor can hand out. Can AI be part of the answer to preserving social learning at universities?
Read the report from MIT’s Ad Hoc Committee on AI Use in Teaching, Learning, and Research Training.
"Happy 60th Birthday to the Word “Hack," in MIT Alumni.
See the student MIT prank that landed a firetruck on the university's Great Dome.
Sponsor:
This episode is brought to you by Studiosity. Real learning happens through real thinking and real experiences. It's something that takes effort, not shortcuts.
For over 20 years, Studiosity has partnered with educators to protect the thinking, the essential work that the student must do for themselves, not detection, not an us versus them culture in teaching and learning. Just the real journey behind every degree that is worth holding. They're doing it for the institutions that trust their students, for the institutions that back their teachers to teach, not police, and for the institutions that stand behind their degrees. Keep the learning real with Studiosity.
Get the walkthrough at studiosity.com/learningcurve.
Jeff Young:
MIT has long been known for its unique student culture.
Of course, you've heard the term hackers. Well, the word was first coined to describe the unusual high-tech pranks that students at MIT have done over the years.
Here's an example of one of those pranks. In 2006, students managed to place a large fake fire truck on top of a prominent campus building to honor the fifth anniversary of the September 11th attacks. The 25 foot long mock truck, it seemed to magically appear overnight on top of the great dome on campus, thanks to a team of anonymous students.
Other similar stunts over the years included putting a lunar lander on top of that same dome, a giant statue of Pac-Man, and back when a new Harry Potter movie was coming out, students marked the dome with a giant lightning bolt scar.
As part of the tradition, the student pranksters always leave instructions for how they achieve the feat, as well as a step-by-step guide to safely removing whatever they'd done. Now, this has been going on for a long time at MIT.
Apparently, it was back in the 1950s when someone first described these pranks as hacks. Of course, MIT is a place where many of the biggest tech innovations have emerged. I found a page on the university website that lists about 50 of these, including the transistor radio, Gillette razors, the spreadsheet, the World Wide Web, and not surprisingly, MIT has played a key role over the years in the development of artificial intelligence.
So, at some point, someone first described a computer break-in as a hack, and when you think about it, hacking is kind of an interesting term because it can be used to describe something harmful, like a bad guy breaking into a computer network, but it's also sometimes used to describe a clever solution, like a hack to solve a problem in a surprising way.
But these days leaders at MIT are worried that student culture around learning is changing, and the reason is generative AI. Last month, MIT released a 13,000 word report that is getting a lot of attention within higher education, and even making headlines in the national press. It's all about how ChatGPT and other AI tools are impacting teaching and research.
And it made some points that surprised me, even though I spend a lot of my time looking into this stuff. An article in the Washington Post about the report was headlined, "AI Can Now Credibly Complete Most Undergraduate Assignments. MIT warns.” The report says AI has caused a “long-term tectonic disruption” to college teaching.
Now, what struck me the most was what the report says about how AI threatens student culture. Though it doesn't explicitly talk about student pranks, it says that since ChatGPT and other AI tools have emerged. Leaders are hearing that informal study groups are starting to fade.
But as stark as its warnings are, the report is unusual in that it also offers some optimism that the folks at MIT and other colleges can overcome this new challenge.
There's a sense that professors and students might just come up with a hack to get through this.
Hello, and welcome to Learning Curve, where we explore how education is adapting to the rise of generative AI.
I'm Jeff Young, a longtime education journalist.
For this episode, I'm diving into this MIT report with a focus on how AI is changing the culture of colleges when it comes to learning.
I'll admit this is the first time on Learning Curve that I've focused a whole episode on a report from some committee, which could sound maybe like a dull starting point, but this document is unusual, and I think it's actually the perfect way to kick off the second season of this podcast. One unusual aspect is that the report says it follows two guiding principles. The first one is be humble, meaning the folks at this esteemed institution do not pretend to have all the answers here, and the second is “be bold,” suggesting that this moment requires big changes to how teaching is done.
So, in this episode, you are going to hear from the co-chair of the committee that put this report together. For the record, the group is called the Ad Hoc Committee on AI Use in Teaching, Learning, and Research Training.
So, I found out that this committee included not only professors and MIT leaders, but it had four students, two undergrads and two graduate students. And since student culture is where I'm focusing today, I naturally wanted to hear from one of these students.
Daniella DiPaola:
So my name is Daniela DePala. I am currently a postdoc at. MIT Media Lab. I just finished my PhD also at the MIT Media Lab this past spring, and I've been at MIT for the past seven years as a master's student and a PhD student. So I've been here for a bit. It
Jeff Young:
turns out that Daniela's research is actually about using AI in education, and so she has a pretty nuanced view here, but she has also been in a position to see how the arrival of ChatGPT changed undergraduate life here.
Daniella DiPaola:
So I am a RA and live in the undergrad dorms where I did, and I had been doing that for five years.
And so for me, I could get the graduate perspective from my working with my peers, and then also, I was seeing how this was impacting undergraduates who were going through similar changes.
Jeff Young:
It boldly admits that we are in a complex moment here with no clear answers, which is really what this podcast is all about. Again, the aspect of this report that struck me the most was how AI was changing the social dynamics of learning at MIT. The report also gets into how AI is changing learning. We're going to get into that later in the episode, but student culture is a key piece, and it turns out that is what struck Daniela as well as she helped put it together.
Daniella DiPaola:
To me, that was the most important thing that needed to be addressed in the report, and I had done an engineering degree at another institution, and engineering degrees are difficult. And you have really difficult problem sets, and you're in the library long hours. When I got to MIT and I became an RA, I was so impressed by the idea that they call it “P set culture.”
So you have a problem set. Most students are sitting in circles, working around a table on these problem sets, and I was so encouraged by like this collaboration that inherently happened through classes, and it happens a lot in the first and second year courses. Given these were, you know, students were just getting used to MIT and maybe hadn't taken the classes before they could learn from one another.
What is also really cool about the learning community or the living community at MIT is that it's all four years are living on the same floor. So often, also that was an opportunity for older students to mentor younger students.
So all this to say, it was like collaboration. I think was one of the most exciting parts about being in a residence hall at MIT, watching students actually working with one another. This past year, I and that was happening also right after COVID, when I think a lot of people were saying, ‘You know, students aren't engaging with one another enough. We need to bring back sociability and the dorms.’ And we were really thinking about ways to bring students outside of their rooms and more engaging with one another after this time that we had been very isolated.
So fast forward to last year around this time, I was talking with a freshman who lived on my floor, and I asked. We were just talking about how things were going, and I had said, "Like, how's studying going? How are the classes?’
And she was saying, "You know, the classes are a bit more difficult than high school.”
And I said, "Do you have a good study strategy? Do you have a study group?”
And she said to me, "No, I just prefer to study alone in my room.”
And you know, this is one person, but it was pretty alarming to me that you know that someone was having this kind of freshman experience when I had seen a lot of students go through this process of forming study groups and meeting one another, and then working with one another.
So I brought this to my … I hosted little food gatherings once a week, so I brought this to my students, all years, and I said to them, "Do you feel like there's a difference from your first year to this year?
And one of the seniors really put this, I think, in a really important way. He said to me, "I think that in my first few years, you needed one another to get a problem set done.” Like it required collaboration between humans to get a problem set done, and then there's all these other benefits of working together too, of making friends and you know learning about different perspectives on how to solve a problem, for example. But sure, in order to get your assignment done, that's what you needed.
And he said that right now, “You could very easily use generative AI to support you through that process. You didn't need other people to do that.”
And so it was pretty alarming to me that this culture of working together and collaborating wasn't needed to get to the objective of what learning in school is right now.
And there's so many other benefits again to socializing, forming study groups, making friends. That I thought it was worth addressing.
Jeff Young:
Now, this is not just at MIT. Anecdotally, I have heard this happening at other campuses as well. These informal study groups are declining now that each student can just. Ask AI. Daniella feels like this means students are losing a core communication skill in the process.
Daniella DiPaola:
I think students, when leaving MIT, they benefited from this idea of collaborating with one another, of working with other people to solve problems, to getting a problem and saying, "Okay, I want to talk this out with somebody near me, and I do worry about that skill being, yeah, deteriorating over time.
Jeff Young:
So, what does the report offer? I know you said you don't have all the answers, but so as you talked with your colleagues on this committee, what were some ideas that came up to try to counteract this new reality?
Daniella DiPaola:
Yeah, I think I mean there's I think this particular issue can come from many different parts of the university. So you could think about it from the classroom: is a professor providing opportunities for students to do work together in class? Can they practice doing it in class or in office hours. Are they forming study groups in the sessions that they're in, like in person forming study groups so that they feel more connected to those study groups long term? That's one piece.
I think what's also really exciting about the report is it acknowledges the importance of residential life at a university in taking a role and making sure we're showing the value of residential education.
I also think it's interesting right now at the same time that we're saying that students aren't engaging with one another enough, or that AI potentially can deteriorate this interaction. We're closing study spaces like our libraries on campus.
Jeff Young:
Apparently, tight budgets at MIT in the past couple years have led it to close a couple libraries and reduce staffing and hours at others.
Daniella DiPaola:
So there becomes less and less social spaces for students to work together with one another, and so I think that's another piece of thinking about where physically can people interact with one another. Why is there a value in residential education, and so the report also gets at how can we provide more opportunities for people to physically come together and work with one another, and hopefully that will spark them working together as well.
Jeff Young:
And what are some of the examples, just or even just broadly, like what kind of things do you think of when you think of trying to encourage that?
Daniella DiPaola:
I think about dedicated study times where students are encouraged to meet with one another and work on their P sets. Can you create other community events where students are sharing ideas?
I think one very easy and relevant one to this moment is creating communities of practice. So opportunities for students, faculty, other members of the community to come together and share how their experiences with AI are changing and be open about it.
That's also what I do in my research as well. I work on some of those questions. Yeah, interesting. So in a way, using AI explicitly to as a talking point
Jeff Young:
as a way to get people to think about it, and because I think is the idea that a lot of students do have this sense, sixth sense, if you will, that AI is not necessarily helping their learning.
Daniella DiPaola:
I mean, again, these are students that have really high levels of metacognition.
Jeff Young:
You mean thinking about how they learn, sure.
Daniella DiPaola:
Exactly, and so when talking with them, it was pretty clear that they were saying I prefer to use it in this case or not in this case, and they could really think through, like, how it was harming what they were learning and doing.
However, I think there's also like there's other outside pressures from these students of if I can get one more class, I can get a double major, and so maybe it's going to be faster for me to do multiple courses in one semester, so I can get that double major. I can graduate a semester early and save on tuition.
So there's a lot of like external factors that are bringing students, and they're having to grapple with these pretty big challenges of, do I use this? Maybe I don't learn as much in this course, but I get these other benefits.
Jeff Young:
I just wanted to pause and highlight this point. These high-achieving students at MIT are essentially using AI to take shortcuts, even though they know that they are learning less. But the way the system works, the reward structure means that you can get more credentials, like a double major, by racing through more classes, or maybe save money by finishing faster. It's a hack.
As I mentioned earlier, I also talked with the co-chair of the MIT committee that came up with this report, Eric Klopfer. He's a professor of science education and educational technology at MIT, who has been at the university for more than 25 years.
Eric Klopfer:
That's really that's really changed the way that students behave, and you know, our students. You know, I think students. It's affected students everywhere.
Our students. We call them relentless optimizers. You know, they're trying to fit a lot of things into the time that they have here, and it seems efficient to do this because, like, going down the hall and talking to my friend, it would take a while. I'd have to walk down the hall, and then we're probably going to talk about some other things. It's not going to be really effective.
Then eventually we get to the answer, and it's going to take us a little bit longer because my friend doesn't really know it that much better than I do, so like it'd be much more efficient if I just asked ChatGPT and then I did it that way, and it doesn't feel like cheating in that way, and it isn't it isn't overtly cheating, but but it's shortcutting learning to the extent that it's just not going to be effective for for their for their long term learning.
Jeff Young:
To be clear, the report is about more than just the social component.
In fact, it states that quote AI demands a broader, holistic assessment of the nature, scope, and purpose of higher education today.
So, in other words, it's very big in its scope. So, I was curious to hear from Eric Klopfer how the report says professors can pull off this radical reset of college teaching to respond to AI.
Eric Klopfer:
I think a lot of it is going to come from experiments where you know not everybody's going to be sort of prepared or willing or able to do like a new form, totally new form of assessment or rethink what their entire class is like.
But some people will, and some people will if we can provide them the support that they need.
So we will try to bring on some new staff that will help support some of this educational transformation, the exact form of which is still to be determined. But it's something we're working on right now.
And if we can sort of say, you know, I talked to one of the deans and like, you know, if we can come up with some new models, like we have some there's some very well entrenched models for ways particular kinds of classes are taught. If we can come up with some new models for how those classes are taught.
Not everybody needs to develop that model from scratch. One person or a couple people can develop that model, and then we can pass that model along to other people. So I think that's what's going to have to happen. Is that?
Jeff Young:
Do you have any of those yet? I love it. I want. I mean, what's the new model? Is there anybody yet?
Eric Klopfer:
I mean, no, no. I mean, because we're you know we're really we're really just at the sort of the the the dawn of this, you know,
Jeff Young:
Any glimmer of what that might look like.
Eric Klopfer:
I'll speculate. You know, one model is you know in many of our humanities and social sciences classes, there's sort of a model of you know people read a book, maybe there's a lecture or two that sort of like summarize some things.
There's a discussion section. Students write a paper, individual papers, all that at home, and they come back and they sort of do the next book that you're going to now discuss.
That's a well-worn model. You know, not just at MIT. I mean, many did that.
Jeff Young:
Great.
Eric Klopfer:
And you know that model is just fragile at every stage. You know, students are summarizing the book with ChatGPT, and they're coming to classes with the questions like they know they're going to have to ask a question. ChatGPT has told them the question to ask in class, so it seems like they read the thing. The essay that they write is largely supported by ChatGPT. You know, there's so many different stages of the lecture. The lecture they're tuning out and recording, and maybe listening to a summary of later. You know, every stage of it is able to shortcut it.
So the question is, how do you actively engage the students along all in the way?
And so maybe it's about more collaborative projects. Maybe it's about sort of using new kinds of media that they can express their ideas with. Maybe it's about new kinds of interpretations of the of the texts that they're working on. You know, maybe it's about supporting active reading with collaborative reading, so that students are sort of mutually accountable for what they're doing. You know, things like that that might sort of help support students in doing that.
I was talking to somebody earlier. There was a nice piece somebody wrote in the spring. I think it was like, who taught a class like this, and they were like, "I know it's not going to work. So they worked with their students to like, "What should the assignment that we do is? And they, so the students helped develop it. They helped develop the rubric. They wind up writing a collaborative essay and figuring out what the rubric was. This was not here. This was elsewhere. And I was like, "That's the kind of thing that should be happening. It's engaging the students. You know, it's it's you know making them mutually accountable to each other. It's being collaborative.
You know, we didn't talk about it some here, but at some place we we talk a lot about other skills that our students need to learn. The the fact that they're not learning socially is not just a change in culture. Like this is what this is what they need when they leave here.
Yes, they need physics and they need some, you know, electrical engineering when they leave. But they also need to kind of communicate and work with other people, and you know, you know, solve problems, and you know, communicate with people who they disagree with. Like, there's a whole bunch of skills that they need that, like, they're not going to learn sitting by themselves.
So we need to have designed experiences where they can be learning those things as well.
Jeff Young:
After the break, how video games could hold one key to adapting to AI in education.
And we hear more from Daniella about the complex feelings students have toward AI these days.
Stay with us.
Sponsor:
This episode was brought to you by Studiosity.
Real learning needs a struggle, but for educators, assessment security just means more detective work, and that's not what students came for, and it's not what the teachers came for. True learning and authorship isn't found in a probability score or a surveillance log. It's found in the journey.
Studiosity makes assessment security easy by shifting the burden away from every educator and into a proactive institution-wide infrastructure. From Viva-style inputs to clear binary validity reporting, studiosity is the multi-layered assessment security process, turning institutional policy into practice. Just a clear signal to the teacher, owned by the student, that I did this thinking. De-risk your assessment flow and reduce workload. Give your teachers back the time to teach, not police, and give your students the agency to show their own growth. Keep the learning real with Studiosity.
Get the walkthrough at studiosity.com/learningcurve. That's Studiosity.com/learningcurve.
I want to remind you about another project that I am working on these days. It's my new newsletter about AI on campus that I'm doing for The Chronicle of Higher Education. It's called Jagged Intelligence. It's named after a term used to capture that AI is surprisingly effective at some tasks, but hallucinates and fails at others. What do you do with a tech like that? I hope you'll check it out and participate in our regular flash polls to share your hopes and concerns about AI. Sign up at Chronicle.com/newsletters. Okay, now back to the episode. Now MIT, of course, is a major research university where many of the professors are actually working on AI development in various ways. So, would it make sense for this university to build its own large language model that would be more conducive to learning?
Eric Klopfer:
I don't. I don't think we should be in the business of training models. What we can do is building tools around the models. And so you know we talk about you know models and harnesses now in the AI space, and I think higher education can be in the business of building harnesses, and building the tools around them that make them effective for students and make them and make them useful for students.
So not just effective. I think we now know like just giving students the answer is probably not the thing to do most of the time, and we want to we want to you know sort of include the productive struggle, but you know when the answer engine is right next to it, why would I want to use the one that has productive struggle when the answer engine is right there?
So we need to make ones that feel also useful to them, because
Eric Klopfer:
I think the ones that make them struggle right now don't feel useful, and I think part of that is about you know better understanding of like sometimes we actually probably should give a student an answer because they're struggling in a way that they're going to disengage and go to the other tool if I don't give them the answer right now. Can we build models that are more intelligent and more adaptive around that kind of thing so that that you know it's providing the the support that a student needs to stay engaged, to stay learning and to stay, you know, we in education we both the zone of proximal development that sweet spot between too difficult and too easy that keeps a student engaged and learning. How do we keep them in that space there?
You know, my background, I don't want to work in games. You know, that's the place where games. I was just about to …
Jeff Young:
It's like the gaming training inside the game.
Eric Klopfer:
So when you're in a game, like you're in that sweet spot, like it's you're just testing the edge of your expertise. You know, it feels good because you're like you're making progress. You know, if we can build that same kind of thing into education, where students feel like they're making progress because they've always kept the edge of their expertise. I mean, there's a long history of research in this area.
Jeff Young:
If you've ever played a complex video game, you can probably relate to this feeling that keeps you going. It's hard, but you're always about to solve the next puzzle or beat the next boss. Of course, there are plenty of examples of ed tech making products that are game-like, but that do not live up to this vision. So it's going to be challenging to harness AI in this way.
Eric Klopfer:
It should be like real good games. It shouldn't be gamification. It shouldn't be just for giving people badges and rewards for doing things. It should be the sort of like you know we talk about you know a good game is hard fun. You know it's a productive struggle. It's a good challenge, and the reward comes from overcoming the challenge, not from sort of someone giving me points in the aftermath.
And education should be the same thing. I should feel like I should be should be rewarded by the productive struggle of doing something difficult and overcoming the challenge, as opposed to the fact that I got a grade at the end that sort of tells me how I did.
I think again the sort of the straw that breaks the camel's back. We're seeing a breakdown in the sort of the value of the information that's coming. You know, we see it coming into MIT and coming out. You know, what does what do grades and essays mean when they're coming now to us? When many of that may, much of that may be assisted by AI.
You know, what is the signal that we're getting in terms of our students? And then when our students leave here, what is the signal that they have to an employer that they really have the skills that they need, and so maybe that's not grades. You know, maybe it has to be a portfolio of work. Maybe it has to be some other way that they're sort of communicating that they've really learned something here and able to do real things that are that are meeting the challenges of industry as well as you know, life outside of work, but but it seems right now that like that signal is is is weak because so much of it is being disrupted by AI.
So I think that sort of is an opportunity to think about the way that we you know maybe it is about portfolios or other kinds of ways of communicating the value of the work that people have done here, because that's that's what's needed, and in turn that changes the whole dialog around what they're doing as part of that work. Because the grade the transcript is no longer the thing that they're working towards. It's the showing I really have value and able to do the kind of work that's needed outside of school.
Jeff Young:
It's not just trying to keep MIT the way it was before AI came along. It suggests some reforms that were talked about even before AI came along, like changing the grading system to make it less like just gamification.
I was curious about what Daniella DePaolo thought of the suggestions that her committee made when it came to teaching culture at MIT. What about the classroom of the future? Like, what do you do?
You know, because it was interesting. Like, it seems like even a lecture class, which you know people have complained about the lecture for a long time, but it seems like there's a lot of ways in which like this is like you know AI really kind of cuts into that effectiveness.
Daniella DiPaola:
I remember one of the professors on the committee spoke about like the beauty of a lecture, and when I first heard that I was like, “What?”
But there is some kind of like really amazing aspect of there's somebody in front of you, who's thought of exactly what they want to present to you, you are sitting there and listening to it, and you're not distracted by other things. This is your time to like take it in, and so I think one, it's reminding students potentially of the beauty of the lecture, reminding them what they're getting, and that comes to this metacognition aspect of like, what are the skills I'm getting during these things that are very easy to replace during AI with AI?
Jeff Young:
Right, have AI summarize the transcript and just just get the bullet points later.
Daniella DiPaola:
Exactly.
I also think it definitely requires innovation, and a lot of conversations have also been let's use blue books, let's do all of our assignments orally. Let's, you know, there were all these different ideas about ways that we could assess it that wouldn't require AI, but would kind of keep the learning objectives similarly.
I think we need to do all of it, and the report kind of says that. Like, let's kind of let's try out a bunch of things and not like just sit here and not try them, but let's evaluate them and reassess how we want it to be used in the classroom.
And I think at a micro level, that's something that professors can be doing is providing a blue book assessment, providing an oral exam, having a P-set, having a traditional exam, and really having direct conversations with their students about the impacts that they feel like those types of exams are having on them, how they're using AI to support their learning in all these different situations, so that when a student in the future needs to make a decision on how to use AI to support their learning, they could be better informed, and they've had that practice they can go back and look at.
I think about this one example that I use in my research is if you think about a high school student, it's late at night. They just had sports practice. Now they have to finish all of their assignments. It will be very easy to finish one of the assignments using AI. So the student has to like grapple with this idea: ‘Should I be using AI or not?’
And if that student has had opportunities in the classroom to use AI to support their learning and their teachers guiding them, or they're using it on their own, then they're reflecting on it with their peers. Then, when they get to that moment, they can say, "Okay, what is the best way that I can utilize this?” And maybe it's not utilizing it because I know I'm behind in this course, and I know that when I did this in class, it. Didn't help me.
Or I knew I had a another example. Is do I go to my office hours for my professor, or do I do the assignment using ChatGPT as a tutor? And having those experiences and being able to recall them, I think is very important to making better decisions the next time.
Jeff Young:
Interesting. So basically, modeling like the right way, if you will, to to use AI to help you learn instead of just have it do it for you.
Daniella DiPaola:
It will definitely require more conversations about metacognition and how students learn, and making sure that educators and students are having open, regular conversations about this.
Jeff Young:
Yeah, so it's such a temptation that you really have to have this. Feels like there should be almost like on cigarettes as a warning.
It's like seems like these tools should just have a big warning about “This could hurt your learning.”
Daniella DiPaola:
Yeah, I think that's that's one big piece of it, and I do I do want to give students and educators credit where they understand how it's impacting their learning, but I think we need to continually be yeah pushing back on and maybe finding better ways to use it that's not impeding on our learning and supporting and supporting it. Do you see it as having potential if used correctly in helping people learn better. I think there are really exciting opportunities for different types of like ways to get information or different types of learning. I think one really exciting part of one exciting tool that I've seen used a lot, especially in high schools, is Notebook LM, and students basically taking all of the information they've learned, synthesizing it themselves, and then using that to get different mediums for studying. So, as a study, have it do
Jeff Young:
A podcast study guide or flashcards.
Daniella DiPaola:
Exactly, and I think that using AI as a study tool is an incredibly powerful tool. And also, this act of choosing one information to put into Notebook LM is also a really important one for reflection.
So I do see it there. I think one aspect of using AI that is often seen as a good use, but I actually would caution against is using AI for brainstorming or coming up with new ideas. I think that we know from research that having confidence in one's own ideas and like creative self-efficacy leads to one better creative outcomes, but to like more motivation to create and make new things.
And so, by not exercising that muscle of thinking through and creating ideas on your own, I think also that
Jeff Young:
Interesting. So it might be a bad habit to be like, "What should I do a topic on for my you know study to ask ChatGPT is not the way you want to really spark that beginning journey of curiosity.
Daniella DiPaola:
Exactly.
Jeff Young:
Yeah. How hopeful are you or optimistic that this will, you know, that some of these trends that you're concerned about will get reversed or can be addressed?
Daniella DiPaola:
One thing that's been really encouraging is the response to this report. I think a lot of people are reading it and saying, "Yeah, you're putting into words a lot of the things that we've been thinking and not trying to prescribe the solution, and so I am optimistic that people are going to come together to to continually revisit this topic.
It's clearly an a topic that everyone feels is important enough to continue to have conversations about, and I also really love that MIT is thinking about this from a multi-stakeholder approach. So, what does residential life have to have to do with the impact of AI? What is academic life, and how do we bring those together? What has surprised you most in those listening sessions and the research for this report?
The biggest, and I mentioned this already, but the biggest surprise from talking with students was the that they had this awareness of how AI could support or deteriorate their learning, and they were having to make those micro decisions constantly about how they wanted to use it every day.
Jeff Young:
Do you have a sense that students wish it would just go away? Or I guess there's no one answer to this, but do you have a sense that a lot of students kind of just wish this wasn't a temptation at all right now while they're in college?
Daniella DiPaola:
I think there are some students that feel that way.
I hear that a lot from my peers who went through school, I was an undergrad 10 years ago, so my peers are saying I'm so glad we didn't have to deal with this when we were there.
But the listening session that I particularly led is was about AI and the humanities, so I had a lot more students, yeah, that were focused on some of the harms in the humanities. And the harms.
Jeff Young:
Let's talk about that for a second, then, because I do. What were some of the biggest harms people were concerned about in the in that listening session?
Daniella DiPaola:
Human authenticity and creativity, like just the the concern of not being able to practice writing one's own ideas and creating, coming up with new ideas, I think there were concerns of is my professor just going to run my essay through ChatGPT, and I think that's an exciting opportunity for both parties in the classroom to have a conversation about their individual uses.
Jeff Young:
They did not like that idea that they just get graded by a bot.
Daniella DiPaola:
Exactly, and so then you have students who are like, Yeah, there's this like everyone's talking about this right now.
One of the students creating something with AI and the professor is creating it with AI. Most of the attention these days is about how students are cheating with AI, but for students, there is a concern that it's the professors who are starting to phone in their teaching with this technology. I remember one student saying to me, "I could clearly tell my professor was responding to my email with AI, and that didn't make me feel good because I felt like that we weren't having a connection, or I wasn't building any rapport with them. So I think you know these are the again these like small moments of human connection that happen in the classroom that are potentially deteriorating, right?
Jeff Young:
And I think a lot of people just looking at this report from the outside might be kind of shocked because MIT is obviously people. It's hard to get into. People are very good students to get there, and you know it's expensive. All the things like this is, and so that people in this environment would be, you know, having this experience with AI might shock some people. I guess.
Daniella DiPaola:
Yeah, I definitely think that, and that was it. Was something shocking when I came into MIT was that how collaborative it was and how much it required working on a team. I think that's the engineering process, though. I worked on a robotics team, and you need many people to work together to solve a problem. And so, yeah, it's kind of the beauty of engineering, and I think that it is something that we're we're definitely losing.
Jeff Young:
And so that's the I guess that that's why this is also like a. It feels like the report is also trying to have a flashing red light of like something needs to change. Is that fair?
Daniella DiPaola:
Yeah, I would say that.
Jeff Young:
There is one other really unusual aspect of this MIT report about AI that I wanted to talk about, which is that the authors are already planning a sequel. Actually, lots of sequels.
Daniella DiPaola:
The idea is to revisit the policy regularly, so there will be. I, I often also with policy, you have a list of rules. You see what the rules, and we you don't put the resources into reevaluating them and seeing how students are engaging with them or what it looks like in practice. And so I think it's really great that we've acknowledged that we need this needs to be like a living, breathing community, and regularly checking in about AI use.
I think another example: New York City Public Schools did a great job of saying that as well. Let's pilot something for a year and look back next summer and see how it's gone, and then reevaluate from there.
Jeff Young:
So it's not to say like we're smart. We figured this out. Here's the answer. Exactly.
Daniella DiPaola:
And I do think that is the skill that will need to be developed moving forward for most students and educators is understanding what you don't know about the piece, the technology, regularly checking in and communicating with the other humans around you about how it's going, and then re-evaluating.
Jeff Young:
So that's where we are at the start of the new semester, the start of a new season of Learning Curve.
This MIT report has all these warnings about the potential downsides of AI in the classroom and for student culture, but it also suggests that adapting is possible.
It concludes, “our work as a committee convinced us that the people of MIT are intensely interested in meeting this challenge, and uniquely equipped to do it.”
I don't know; it does sound hard.
Anyway, diving into this report made me realize that the stakes are really high here. It's not just about how well students learn, which is super important. It's about the future of creative innovations that have long come out of universities like MIT, and these innovations have changed all of our lives.
The spirit of pranks and games, and late night study groups are part of what has made that possible. And now the folks at MIT are fighting to update that spirit for an AI age.
###
This has been Learning Curve.
Every episode, we go deep into questions about how education is adapting to AI?
We have some exciting stuff coming in season two, including interviews with leaders of major AI labs and trips to more campuses around the country. If you haven't already, you should follow the show on your podcast app so you don't miss an episode.
And I'm always curious to hear what you think. Shoot me an email with your thoughts, reactions, criticisms, story ideas to Jeff at LearningCurve.fm. I might end up calling you up and getting you on a future episode.
This episode was reported and put together by me, Jeff Young. You can find out more about my journalism and speaking and about the show at LearningCurve.fm.
The episode art was generated by Midjourney, and the music was composed by a human artist who goes by the name Kamakoo. I also want to thank this episode's sponsor, Studiosity.
We'll be back in two weeks with another episode.
Until then, thanks for listening.