Tuesday, October 21, 2025 - How can AI be adopted in a way that turns more students into “explorers” rather than “passengers” in their learning? This week we bring you a conversation with the co-author of a book on student disengagement in school, Rebecca Winthrop, who is also researching the impact of AI on education. The episode is by one of Future U’s producers, Jeff Young, from his new podcast, Learning Curve.
Get notified about special content and events.
0:00 - Intro
4:19 - When the ‘Student Disengagement Crisis’ Started
7:25 - A Framework for Describing Levels of Student Engagement
15:18 - How AI Is Impacting Student Motivation
19:00 - Why ChatGPT’s ‘Study Mode’ Is Not the Answer
25:05 - Advice for Companies Making AI Tools for Education
29:32 - Tips for Students
34:42 - A High School Student’s Take on AI
48:30 - Advice For Teachers on Dealing with AI
51:35 - What Is the Purpose of School in the Age of Generative AI?
“The Disengaged Teen,”
by Rebecca Winthrop and Jenny Anderson
“Minnesota high school student weighs the benefits and pitfalls of AI,”
Minnesota Now
“I’m a High Schooler. AI Is Demolishing My Education,”
The Atlantic
Brookings Global Task Force on AI in Education
website
‘We Have to Really Rethink the Purpose of Education,’
The Ezra Klein Show
“Attention Please: Professors Struggle With Student Disengagement,”
EdSurge
“Playing the Grade Game,”
Bootstraps podcast series
Michael Horn
Michael Horn here.
Jeff and I are thrilled today to run a podcast from our producer, Jeff Young's learning curve podcast. You can check it out wherever you subscribe to podcasts or at learning curve.fm.
The episode that we've chosen to run in the future you feed is titled, ‘Will AI Bring More Student Disengagement?’ It features Rebecca Winthrop from the Brookings Institution, who, along with Jenny Anderson, co-authored a terrific book titled “The Disengaged Teen: Helping Kids Learn Better, Feel Better and Live Better.”
And in this conversation, Rebecca takes those insights, along with her more recent and ongoing research about AI's impact in education that she's helping lead at Brookings to flesh out just how could AI boost students' learning in what she and Jenny call Explorer Mode. But also how, if we don't make more fundamental changes to schooling, it could send even more students into quote, unquote Passenger Mode, coasting by essentially as they learn very little, because students are offloading the real work to AI.
There are some fun twists in this episode and a really interesting and frankly very nuanced look into how one high school student is seeing this play out in front of them real time in the second half of the show. So be sure to stay tuned for the whole conversation.
And with that, we'll send it over to Jeff Young for a set of conversations on whether AI will bring more student disengagement from his podcast, Learning Curve.
Jeff Young
Hello and welcome to Learning Curve, a look at what it means to teach and learn in the age of generative AI. I'm Jeff Young, a long-time journalist exploring the intersection of education and technology.
The idea for this episode started earlier this summer when I was listening to the Ezra Klein show. I love that one. And he had on a researcher from The Brookings Institution talking about how to keep kids motivated in school.
And it was really about the crisis in education these days about getting students interested in learning. The researcher that Ezra Klein was talking to found that most students just aren't that into school. And of course, when students are not that into school, it's probably going to be hard to get them to learn very much. It's pretty fundamental. I was riveted by this interview. Maybe you heard it, which stepped back and it took a big picture look at what the role of school should be now that AI is out there and can do some of the writing and basic information analysis that we'd long thought only humans could do.
It's kind of that, what are we doing here?
That’s the question that this podcast is really focused on.
After the interview, I was also left with more questions for their guest, and I really wanted to sit down with her and go deeper into these issues of how AI impacts student motivation to learn.
So I reached out to this researcher. Her name is Rebecca Winthrop. Her new book is called “The Disengaged Teen,” and it's full of interviews with high school students and with highlights from years of research into what has been called the ‘teen disengagement crisis.’
And Rebecca Winthrop was totally up for talking, which is cool because she is in high demand these days. She's been interviewed on the Drew Barrymore talk show, and Oprah Winfrey recently recommended her work, saying every parent with a teen should read this book. And Rebecca is also very much looking into AI these days. Specifically, she is leading something called the Brookings Global Task Force on AI and Learning.
So I'm going to quickly preview that this conversation took a couple of interesting turns that surprised me a bit, and it led me to bring in a high school student that you're going to hear from later in the episode.
I started out by asking Rebecca Winthrop, when did this student disengagement crisis start in schools?
Rebecca Winthrop
So kids have been disengaged for a really long time, if you look so we're not talking about this like just talking about started covid Only or smartphones only, right? Both of those things have heavily exacerbated the disengagement crisis.
But kids have been disengaged at least for the last two decades. If you look at U.S. census data, which asks a couple questions, not a huge battery of questions, about kids, you know, about two-thirds of them are disengaged from school, if you just say engaged, disengaged. So two-thirds, and it's been away for quite a while, and it's probably, if you look at the, you know, “The Breakfast Club” or “Ferris Bueller's Day Off.”
Jeff Young
I love those movies. Yeah, for a long time, you're right. It sort of glamorizes it in a way.
Rebecca Winthrop
It's this idea that high school is boring, and, you know, who needs it? That it's a waste of time. All this stuff has been around for a long time.
Now, the question is, what has changed?
And what has changed is that the cost of being a disengaged student are much higher today than they were a couple decades ago. Because when you enter the world, including a world with AI, you actually don't develop the skills and competencies you need by coasting your way through school, you are not going to be a collaborative problem solver or a deeply empathetic conflict negotiator or an agile learner who can navigate all the multiple new technologies or highly motivated and know yourself and know your interests and what you want To pursue, which is actually, I think, more and more what kids are gonna need in this AI world that can do all the tests and all the essays for
Jeff Young
Yeah,so in the movie students say ‘Save Ferris.’ For Ferris Bueller maybe that used to work. Like maybe Ferris did fine. But today, the star of that movie is not necessarily going to come out of high school and be just fine.
Rebecca Winthrop
I mean, it's a really interesting comparison because, and I'm sure we'll get to this, I think Ferris Bueller was a total resistor. He was highly creative, yes. And so I am less worried about Ferris and more worried about all the students in the class that are blowing the bubble gum and bored to tears when the teacher’s like Bueller.
From “Ferris Bueller’s Day Off”
Bueller. Bueller. Bueller.
Rebecca Winthrop
Just trying to survive, totally compliant. Ferris, at least, is exercising some agency over his life and deploying creative thinking and critical thinking. Personally, it's one of my favorite movies
of all time.
Jeff Young
Yeah, mine too,
Rebecca Winthrop
But it’s towards a different end than the school system wants him to.
Jeff Young
In “The Disengaged Teen,” she lays out a framework to describe the levels of student engagement, and her research shows that it's not just as simple as saying some students are good at being engaged and others aren't.
It's more nuanced than that.
She talks about four different modes of engagement and argues that it is possible for a student to toggle between one and another depending on the class or other things going on in their lives.
So these four modes of engagement to think about are Resistor Mode, Passenger Mode, Achiever Mode and Explorer Mode.
Quick definitions: Resistor Mode is that Ferris Bueller example — someone who is tardy or absent or maybe the class clown. They may have tons of potential, but they are checked out of the learning process and kind of acting out as a result. Passenger Mode is one that Winthrop says the majority of kids in her research showed. These students are not always seen as a problem by this system, but they're not pushed very much to learn in school.
Rebecca Winthrop
But these are the kids who are coasting. They are physically present. They might love school because they get to see their friends, but they've dropped out of learning. And frankly, a lot of kids we talked to who are in Passenger Mode were getting straight A's, right? So these are not bad students in any way. They might look fine on paper, on paper, they're just bored to tears because they're not challenged.
And then there's Achiever Mode, which is kind of a complicated one because these students are maybe too focused on grades and not always on the learning. They are the go-getters.
Jeff Young
I'm sensing extra-credit doing…
Rebecca Winthrop
Yep, they're doing the extra credit they're going after, you know, captain of the chess club and the football team, and they're running student government, and they, in many ways, are developing great skills — goal orientation, organization, executive function. They're ambitious.
So this one is really tricky. There's times when kids who are in Achiever Mode are doing great, getting good feedback from teachers and parents, and they're happy achievers, and they're feeling good. They're rocking it.
9:50
And we did find, unfortunately, that there is a dark underbelly and to Achiever Mode. And that kids can tip into Achiever Mode if they start taking on Achiever Mode as a personal identity. It's not just a mode that they deploy to get something done in a class and then maybe another class, they'll be in Passenger Mode, you know, because it's too easy on the coast.
They start thinking of themselves as an achiever. This is who I am through and through. And when that really happens, you see this dark underbelly come forward, and you see a lot of risk aversion and learning.
Jeff Young
You don't want to take a chance, because you might not get the A Plus.
Rebecca Winthrop
Exactly. We heard so many kids say, ‘Well, I totally want to write my essay saying this, but my teacher, I don't think, is going to give me a good grade on that, so I'm not going to respond that way. I'm going to write what I think my teacher wants.’
And, ‘You know, I'm not going to take the hardest class because I'm I heard it's really hard. I might not maintain my 4.0 or my whatever.’
So there's a lot of risk aversion.
They end up also being really fragile learners. So when they get a B plus or a B, oh my god, a B, which, you know, I have plenty of students who would be thrilled with a B. Some of them might be in my household. You know, it's like an existential meltdown.
We talked to one student who was really, you know, in this Achiever Mode, who, his first time that he took the SAT did not do very well, and was like, ‘Well, that's it. I guess I'm not going to college.’
If you're a long-time listener to my work you might have caught the Bootstraps podcast series I did a couple years ago. I talked about my own experience in high school when I struggled a little bit with some of this grade chasing. The episode was called “Playing the Grade Game,” and I will link to that in the show notes.
But yeah, this one hit close to home.
And finally, there's a. Explorer Mode, which is kind of the gold standard of learning, but it's also the most rare in today's schools.
Rebecca Winthrop
Explorer Mode, where we found, with our Brookings transcend work, that less than 4% of middle school and high school said they regularly get a chance to be in Explorer Mode in school.
Jeff Young
Less than 4% okay, so this is a rare mode.
It's rare. It's not necessarily a rare mode for how kids are naturally hardwired. It's a rare mode for the opportunity in school to be done. And it's really where kids’ curiosity meets their drive and they dig in.
It's where from the work of Johnmarshall Reeve, you have what he calls agentic engagement. It's where agency meets engagement. So kids in Explorer Mode influence the flow of instruction to make it more interesting and supportive to themselves. And it can be in tiny ways. It can be like, ‘Can I write my paper on that?’ Or, ‘Hey, I disagree with your, you know, I disagree with your phrase of the of the essay question, I think I want to write it on that. Is that okay?’ Or, ‘Hey, can I study with a buddy?’ Or, ‘I would really help me if I watch YouTube. Can I watch YouTube?’ Or whatever, you know, all the way to if they're given the opportunity, you know, co-constructing learning, project based learning, etc.
You know, there's a wide range of innovative pedagogies that if kids are given the opportunity, they will take if done well. But you don't have to be in a totally redesigned high school to be in Explorer Mode.
Two decades of [randomized control trials] show that if teachers just give kids a little bit of autonomy and choice and options and ask them for feedback and give them sort of a little bit of space, kids can get into Explorer Mode within existing schools and existing curriculum.
And it's really the agency piece, the proactivity piece, the initiative piece, because you are not done to which is how passengers feel and achievers feel done too. But they're going to crush it. They're going to rise to the occasion. You know, kids in Explorer Mode feel excited, motivated, energized.
And you know, the more engaged and motivated you are. Like, you attend, better you remember, better you make connections, better. You have more energy, you process, you dig in, like, and it's not a huge hard slog. It makes a big difference for kids' outcomes.
And Jenny and I would argue, in Explorer Mode, you're being prepared for a world of AI.
Jeff Young - 14:34
So one of the things that I am struck by in seeing your framework and just following AI and education, it seems like there's this contradiction, or kind of a tension with AI, at least its potential when it comes to learning, is that, on one hand, it seems like the ultimate tool for Explorer Mode, right? A student in Explorer Mode, because it's literally helping you with your agency. There are literally AI agents that you can, you know, like hire to go help you to achieve your goals.
And yet, AI is so often used by students as a crutch to outsource their assignments to ChatGPT instead of thinking at all. And so that seems like the ultimate Passenger Mode enabler.
So what do you make of — and from your research — how do you think about AI and how it is impacting education in the schools that we have today.
Rebecca Winthrop
I think what you just said, ‘the schools we have today,’ is actually the most important. It's the whole question.
So in today's world, I think AI is being used to create more passengers.
Jeff Young
Which is already the dominant group.
Rebecca Winthrop
Which is already the dominant group, right.
Now, I don't think it's a foregone conclusion that that is the only way Gen AI could influence learning.
I'm in the midst, as you know, of this big research project with the Brookings Global Task Force on AI and Education, which I'm leading. So we're in the throes of looking at the data, and I definitely see that there is a path. It's just a narrow path for Gen AI to really help transform education and schools.
So you know, we've long talked about, could some of the sort of architecture of schooling, whether it be seat time or just, you know, sort of the task completion be shifted to not just acquiring knowledge and content, which is remains incredibly important. We do not want to throw the baby out with the bath water. But marrying that with knowledge application in whichever. Or innovative pedagogy or format or school design you want to so, you know.
Could AI really push schools over that hump any way that you know, many schools on the margins are doing already, there's lots of innovative schools. It's just not at the center of the system. It's not the dominant, predominant approach.
Jeff Young
But that takes a little work to adopt AI responsibly. There's a default going on.
Rebecca Winthrop
Frankly, I think there's two big things that need to happen. I think we need big guardrails, commercial, general purpose, AI, tools and platforms. Here I'm just talking about generative AI that are absolutely in kids hands most of the time outside of the school.
Jeff Young
They are some of the biggest users by the numbers of chat. GPT, yep. School-age kids, college kids, college kids.
Rebecca Winthrop
Yep. So you know, I'm really focused on children — sort of 18 and under, just because legally and developmentally it's a little more clear. Let's leave higher education to the side. But you might have opinions.
But for children who have developing brains, and, you know, literally children's brains develop the way they're used. They should not be totally cognitively offloading all their thinking and learning to AI. That would be a very bad outcome if they for this generation, if they did
And you know, these, these products are, you know, these AI platforms and products are miraculous. They're incredible. I am a big fan. I use them all the time. I use them for recipes. Number one thing, what's in my fridge? I have, you know, tomatoes and an onion and pinto beans and a loaf of bread. What do I make? It comes up with great stuff in addition to all the other things, but they are so powerful and incredible that they're very seductive. And just because ChatGPT comes up with study mode side by side with normal mode, does not mean young people are gonna only use study mode.
Jeff Young
Yeah. There is this study mode option. It’s recent. I saw that announcement.
OpenAI, which is who makes ChatGPT released this study mode feature at the end of July, and the idea is that it's supposed to add guardrails to this AI chat bot so it won't just spit out an answer to a question or write an essay, but will guide students to figure things out for themselves, or at least that's how it's pitched.
Rebecca Winthrop
Honestly, I played with it right when it came out, it'll get better.
But actually, oh my god, I think I'm more worried about study mode. Thank God none of the kids are going to use it because why would they do that when they could get all their answers?
I added two half sentences in study mode, and it wrote me an essay and told me, great job.
There was a lot of back and forth. It was like, What do you think about that? Oh, okay, that's a good idea here. How about this five paragraphs? Ooh, how about I change that?
Five paragraphs? Oh, you did a great job. This is all your work. Rebecca, I think I put less than 20 words in. I put it on my LinkedIn newsletter.
Jeff Young
I’m going to pause things for a minute here because this was new to me. I had read about study mode, but the biggest critique I had previously heard was that students were just unlikely to use it. I had not seen reviews saying that it may not be pushing learning as advertised.
I went back and read her LinkedIn post, and here's what happened.
“Consider this,” she wrote: “What if your child toggles on ChatGPT’s study mode and asks for help with writing a short reflection on why parents need AI literacy. And their only original contributions to the conversation are, ‘I think they could help guide their kids not to cheat, and all kids will use AI if they can. So all parents should have AI literacy,’
ChatGPT then replies, nice work. You started from scratch and built a reflection you can own, and then the system produced the following polished essay:
“Parents should have AI literacy so they can guide their children's learning and help them use technology as a tool for growth, not just shortcuts. When parents understand AI, they can help their kids avoid using it to cheat and instead encourage them to use it to learn new things. AI can even be used to solve problems like helping write an essay on why parents need AI literacy.”
After that, Winthrop says that the study mode system asked her, ‘Do you want to keep it like this, short and direct? Or expand it slightly, add one more detail or example to make it a half page?’
And she typed, ‘expand it.’ And then the system added a sentence at the end. “For example, AI can support creativity by helping kids brainstorm ideas for school projects or practice new skills, but it works best when parents help their children use it wisely. All kids will use AI if they can. So all parents should have AI literacy to prepare their families for a future shaped by this technology.”
So that was the study mode's answer. In her post, she details more of the back and forth that got there, but the amount of guidance from this bot is huge, and it does seem like the student can easily fall into Passenger Mode here and just type a few words or phrases and have this bot turn it into an essay that some students might kind of feel like they wrote all by themselves.
Rebecca Winthrop
I think that's worse — the kids thinking that they do a tiny little bit of lifting and it's all their work, and they're great.
Jeff Young
I've heard some similar critiques about other products, right? Like, I think some advanced versions of Grammarly. And it’s basically all the products, not just ChatGPT.
It seems like there is this sense of, ‘You did great.’ And if you think about it, you kind of push the button. Yeah. So I have on this point, I'm very curious. So you know, you also have OpenAI and other companies putting — at the same time they're doing study mode — they're also making huge discounts for high school and college students to use their product to get them into the regular version, the advanced versions that can do a lot.
And I guess if you were, and maybe your report from Brookings will do this, but if you were advising these companies — if they invited you in tomorrow, and they said, ‘Help us’ — what would you have them do?
How would, how could they make their products better to get students to Explorer Mode?
Rebecca Winthrop
Yeah, they're not gonna like it because it will hurt their bottom line. Because the reason they're going after students is learning is a huge use case of gen AI.
Jeff Young
As far as what do you do with this thing, yeah.
Rebecca Winthrop
It's huge, huge.
And, you know, selling to institutions, getting individualized, you know, users, students on their platforms. You know, that breeds brand loyalty, right. And there's a huge play at the moment of who's going to own education.
You've got OpenAI, Microsoft, and Anthropic have partnered with the teacher unions, right? And I think that's great of the teacher unions. You are the ones who should be helping to design what the use cases for this is because you
Jeff Young
To be at the table?
Rebecca Winthrop
To be at the table. And it’s a partnership to train a ton of teachers on AI literacy.
But more than that, what are the use cases? How can the developer serve you teachers? To really figure out how to make this a product that would help kids get into Explorer Mode, help teachers, help scaffold learning, etc.
Jeff Young
Okay, so they wouldn't like your advice. What would it be, though?
Rebecca Winthrop
Okay so the advice would be: don't let kids use the unscaffled platform.
Jeff Young
So instead of giving them a discount, you’d block them. What age would you put?
Rebecca Winthrop
I mean, this is, this is a good question. I don't know at the moment. I have to think through that. Luckily, we're not done with the report yet.
Jeff Young
Yeah, but what would you say, just some range?
Rebecca Winthrop
I mean, at the moment, ChatGPT, age 13 to 18, you need a parent consent for a kid to sign up.
Jeff Young
But people can lie about their age, but that's the official policy now is that kids under 13 are blocked.
Rebecca Winthrop
That's my point. That's where it is today, enforce your current policies.
Like, if you have a parent who's like, ‘okay, you can use this and has the time to sit beside the kid and scaffold and support,’ that's great for kids who have that. There's an equity dimension I'm worried about, but let's even leave equity to the side.
Like my kids say ‘Mom, can I get ChatGPT? Okay, what do you want to use it for? At least, I know it's a conversation. I've been like, ‘Oh, he wants it.’
And then I'm like, ‘Okay, let's talk through how you should use it.’ Let’s maybe set time limits on it. Maybe I'll go in for Parental Controls and figure it out.
But at the moment, you can get ChatGPT anywhere. They get it through Snapchat, by the way.
Jeff Young
So okay, so at least, at the very least, enforce the age limit that is actually technically in effect.
Rebecca Winthrop
Yeah, but, but really, these tools are not optimized for kids, so I would absolutely, absolutely 100% take them off the market for under 18 honestly, and come up with something else that is more scaffolded, more contained.
Because it's not just that I give you all my answers. And you know, there's plenty of edtech folks who are working on this who are doing more-curated designs. You know, like socratic dialogs, etc, that are more appropriate for kids and learning. It's also that, you know, these general-purpose texts are so powerful, and they're powered by the entire freaking internet, so the types of conversations you can get into are totally inappropriate.
I'm not saying all kids are doing that. Plenty of kids are not, but they can, you know. There's no block.
I'm not gonna drop my 14 year old off in the red light district and 3 a.m. and say, ‘Good luck.’ ‘Have some fun over there.’
But that is what they have access to. There is tons of sexualized content. You can basically have human-like conversations that suck kids in, about self harm, about all sorts of things that we don't really want our kids having unmediated sort of access to.
Okay, so it's a tall order, but that would be a big help if the companies would, would, would do that.
The other piece, I would say, is to absolutely 100% change your incentive structure for AI friends and companions, because the incentive structure is the same as social media, which in social media, the company's incentive structure is quote, unquote, engagement. I say engagement because we talk about student engagement, but that’s very different engagement. It's to keep you on the platform. Engagement is how long you're spending on a particular platform or app.
And you know, the more minutes go by, it doesn't matter if you're learning to bake or learning to self harm, companies get the same amount of money. You know, it's more advertisements for you. Like they're not their incentive structure inside the companies are not to, you know, towards truth or safety.
For children, it's for engagement, at least the AI companions, is for engagement, and that's how they're financially, you know, set up, so you need to fundamentally change that incentive structure.
That’s another thing they're not going to like because it definitely will hit their bottom line. But it's what needs to happen.
Jeff Young
On the day I interviewed Rebecca Winthrop earlier that day, I read an op-ed in The Atlantic written by a frustrated high school student. The headline was, ‘I'm a High Schooler. AI is Demolishing My Education.”
It was by a senior at a public school in New York City, and she wrote that she doesn't use AI much for her work, but that she sees it all around her.
In the essay she wrote that “during a lesson on the Narrative of the Life by Frederick Douglass, I watched a classmate discreetly shift in their seat, prop their laptop up on a cross leg and highlight the entirety of the chapter under discussion in seconds, they had pulled up ChatGPT and dropped the text into the prompt box, which spat out an AI-generated annotation of the chapter. These annotations are used for discussion. We turn them in to our teacher at the end of the class, and many of them are graded as part of our class participation. What was meant to be a reflective, thought-provoking discussion on slavery and human resilience was flattened into copy-paste commentary.”
The student wrote that she even saw other students use chatbots during after-school activities, including during a debate tournament.
I described this article in The Atlantic to Rebecca Winthrop, and I read her the student's conclusion, which is pretty bleak. “If chatbots have made school easier to get through, they are also making school equally as hard to grow out of. The technology is producing a generation of eternal novices unable to think or perform for themselves.”
What would you tell the student?
Rebecca Winthrop
The message to the student is really tough. And so we heard this from a lot of students. The same story, you know, I take my essay, I break it into three, I run it through three different Gen AIs, and then I run it through an AI humanizer.
These tools are all out there.
And then they say my teacher has no clue.
And then we also hear kids saying — a new college student, freshman, saying “I'm getting a C. I'm a good student. I know how to use my brain, but I'll take my C. I think it was a C plus C plus, even though I know so many of my peers are just using AI to get through, but they haven't learned this stuff. But it is a bitter pill. I don't know how long I can do this for, right?
How do you motivate somebody who feels like that C student, motivated student, who's working hard and learning material, wants to learn the material, was there to learn, not for just for the grade and the gatekeeping and the task completion, which is not what the professors are there for, either, or educators. But somehow AI has helped make school one big hack, you know.
Jeff Young
So how do you keep that student engaged?
Rebecca Winthrop
Exactly. So, you know, that is a hard one to me. I was like, ‘Well, I'm not going to tell you not to use it. What if you want to go to grad school and you are going to be kept out because you, all of a sudden, are doing the real work, and everybody else is getting an A, and you're at the bottom of the curve, you know?
So, you know, my advice to that student was like, do what you need to do to get through and hack the current system until we figure it out. But learn outside on your own. Learn the content. Because it's an impossible situation for kids at the moment.
I mean, just pragmatically, right until we figure it out. Don't shoot yourself in the foot not get to grad school. It's literally don’t lose your scholarship. Do what you’ve got to do to play the game until the institutions figure out what the heck — but learn on your own.
Jeff Young
It's such a practically hard thing to do too, though, because you're really asking someone to do a really hard hit-your-head-on-the-wall thing and like, suffer and do terrible work, but like, get it done and then go hack and have ChatGPT do it for you and turn in this garbage or this thing that you know is bot-written and looks beautiful, it's not yours. And yet you get rewarded for it in the system.
Rebecca Winthrop
Super painful. Super hard.
Jeff Young
It doesn't even feel tenable, right?
Rebecca Winthrop
Untenable. And are kids really gonna do it?
Jeff Young
After a quick break? What advice does Rebecca Winthrop have for educators when it comes to AI. Stay with us.
[Music]
This is going to be the shout out to sponsors when we get that far. But right now, the show is entirely listener supported. If you haven't already, please rate or review the show, and if you want to pitch in and get bonus material, visit learningcurve.fm and click on the Support page to find out more.
Okay, back to the episode.
Jeff Young
I wanted to talk to that student who wrote The Atlantic article, but I was not able to reach that student.
But I happened to hear another high school student interviewed on the radio where I live here in Minnesota, about AI use by students, and that student was willing to give her perspective for the show.
The student I talked to is Elianah Dollar-Simmons, a high school senior in a suburb of Saint Paul who takes classes at the University of Minnesota as part of a dual-enrollment program in her high school.
She has done a lot of work with AI and she's interested in going into computer science. In fact, she has even helped build AI models as part of a summer program at MIT and at a different summer program she did at Carnegie Mellon University. I first wanted to get Elianah’s reaction to that advice that Rebecca Winthrop just shared. So I played the clip for her.
Elianah Dollar-Simmons
I guess I would say that my initial reaction is, I'd be curious to know if at this — I don't think it was a hypothetical. I think she said this was a real university in question.
I'd be curious to know what their current policies surrounding AI are in the sense that at least at the university I've been going to for two years now, the U of Minnesota, we have very strong academic dishonesty policies in place where if you are caught using AI, that puts you in very serious trouble with the academic department.
And of course, there are students obviously, who still use it, but I would never say it's gotten to the point where students who aren't using it but are comprehending the material are being pushed to the bottom of the grade stack. I guess I've never heard of that situation before. So it's actually really interesting to me to think that a student who is doing their best and comprehending the material could be on paper at least falling short to students who are just using AI to get around having to do the work.
And I think it's disappointing that's the best advice we can give to students, is just play the AI game until we figure out something better.
But I feel like that's how it's been for a lot of new innovations, whenever they first hit the education sphere. Compared to a lot of other industries — like transportation or medicine or law, which is lagging a little bit behind on AI right now, but can usually adapt pretty fast — education has always been in the, ‘Well, you'll just kind of have to play the middle until we actually figure out how we want to handle this.’
And I think that puts a lot of students at a disadvantage. And I wish there was a way for education to be able to see when new changes are oncoming, and then say, ‘How do we prepare for this before it becomes an issue?’
I think in literature, I take the AP Literature class, one of the things that we learned is that in a story, you should always introduce something before you think you'll need it, so that when it does pop up later, the characters are going to be like, ‘Why is Indiana Jones afraid of snakes?’
Indiana Jones
Snakes? Why do they have to be snakes?
Elianah Dollar-Simmons
Isn't he afraid of nothing? Why is he afraid of snakes? Well, we introduced that two or three books ago, and so when it does actually pop up, the readers are already on pace, and they're not like, ‘Where did this come from?’
I feel like it would be great if education could kind of get to the same approach of, let's figure something out before we need it. Because as the technology continues to develop, and we actually do need it now, we're putting a lot of students at a disadvantage because we weren't prepared. You know,
Jeff Young
Yeah, I love the Indiana Jones reference. I'm trying to think how you would have done it as a school when these things all have popped up. I guess it's been a couple years now. But if education had reacted quicker to these snakes appearing, then maybe we could have, as the education system, could have then been more ready for this moment when they're more widely available and used by people.
Elianah said she has seen some other students outsource their learning to chatbots.
Elianah Dollar-Simmons
I work as a teaching assistant at the U of M, and it's for the senior-level advanced history class on the history of computing. And there was a student during one of our lectures the other week. I tend to sit in the back, and I'm not sure if everybody knows I'm the TA, and I watched her take the article they were supposed to read, put it into Gemini, put it in the ChatGPT after that, and then put it into another AI model that was supposed to make it sound more human, like human translation, and then use that to do her assignment.
The whole time I was sitting behind her, knowing I was gonna have to grade it later. And so it's very much like, I don't witness it that often in the classes that I'm taking as a student, but I've definitely seen it in academic spaces.
Jeff Young
And what did so when you saw that, you know, at the University of Minnesota. Were you surprised? Or what was your reaction as you watched the student do this?
Elianah Dollar-Simmons
Probably a mixture of surprise and disappointment, in the sense that I have read the article, because I'm gonna have to grade so many perspectives on the article. And I thought it was really interesting, and I thought it was something that was easy to get through.
It wasn't super technical or anything. It talked a lot about movies, about these specific people who were working on, I believe this was Turing technology, so they were using it to crack German ciphers.
And so to see someone just not even read the article, just put it into ChatGPT, into Gemini, and then call it a day, was I was really disappointed, because I felt like she was really going to be missing out on learning about what I thought was something really fascinating.
Jeff Young
I mean, what seeing the experiences you've had where you've kind of seen somebody using it to just kind of not engage in the material, and knowing that you're using it to learn and go beyond what the curriculum is asking you. Like, do you think AI is gonna be helpful to the education systems of high schools or, or is it gonna be maybe more harm than good? Like, where do you, do you? Do you have, what are your feelings about how it could play a role in high school?
Elianah Dollar-Simmons
That's a tough one.
I think that it really depends on the attitudes that we have towards using it, in the sense that, like I said, for my AP classes, you're not outright banned from using it. It's not like if you use it and we find out you're going to be kicked out of all your classes. You're going to get Fs, all these sorts of things. It's still academic dishonesty to use it to do all of your work for you.
But at the same time, if that's what you're doing, you're not going to be ready for the exams and the tests where you don't have access to outside technology. And that doesn't really hurt anyone except for you.
So a lot of students, knowing and having that attitude, have been more engaged and more able to put a good foot forward. At the same time, I know a lot of students that use AI to make themselves more organized or help them keep their tasks straight.
Because I think I would say the two biggest misconceptions about teens right now would be that:
A: We don't do anything. We just go to school, go home and sit on the couch and play games or watch TV, and
B: that every single teenager is using AI to cheat on their homework and get out of having to do actual work.
A lot of the teens that I know are using AI, specifically ChatGPT, like a planner where it's ‘I have all these assignments to get done. Help me block out times in my day to make sure I can do that. Also go to my sport practices or help me figure out some basic ideas for this essay I need to write while I focus on this job that I'm working to help support my family.’ Those sorts of things.
I have a friend who actually I think, has chat GPT premium, and she uses it to put together to-do lists for her and saves those so that she can study for the ACT and also have time to do all of her homework, and also have time to do her extracurriculars and everything else that she wants to do.
Or it'll help her with college research and laying out different majors, different options. Help her compare financial-aid packages, things that definitely aren't taking a photo of your math homework and hoping Google knows the answer.
So I think it can really be a double-edged sword depending on how we use it. Because, of course, there are people who use it like that, and then there's people who become overly dependent on it. I think our attitude and how we not only approach it, but how we treat the people who feel as though they're reliant on using it can make a big difference in the
classroom.
How would you think the ideal would be to treat how we treat students in this way? What would you recommend?
I think probably one of the most important things is getting to the underlying reason that they feel like AI is the way that they want to approach getting their work done. Because for some students, it could be, ‘I have a lot of stuff going on at home. I have little siblings I need to take care of. I have to work a job to help with my parents. Low-income. I might have elderly family to take care of, so I don't have time to do my assignments the way that I would want them to, but I can have AI do them quickly for me, summarize the notes, and then I can take care of things that I think are more important.’
The teacher might have to take the approach of saying, ‘Well, why don't you consider your schoolwork to be higher up on your priorities?’ Or, ‘How can I change the assignments that I gave you so that either,
A: you're actually able to do them at home and you're not worried they'll be time consuming.
B: Is it something that maybe we help you get started on it during class, so that when you get home, it's not as daunting, because usually the hardest part is actually getting started, or
C: Do we set a time for you to go to a library or for you to come in after school?
And we figure out those things in a way that won't conflict with the work you do for your family and your responsibilities, but also ensures that you're actually getting something out of your education.
For other students, I think if it's not family issues, it's a lot of I don't want it to get a bad score. I want to get the best scores that I can. I want to do really well in all of my classes, and if I've bitten off more than I can chew, well, ChatGPT can take care of the rest.
And it's well saying, ‘How do we get you an actual planner that could help you with time management?’ ‘How do we encourage students to make time in their day so that they're not procrastinating and waiting until the last minute when they've taken on so many academic responsibilities, and now feel like the only way to stay on top of everything is through using AI.
And then, of course, you have the students who just don't understand what's happening in class, but don't want to be the person who says they don't understand. So you use AI, which we assume understands everything, and that will get the work done.
So that can be a matter of, ‘Do students need extra help comprehending the material?’ ‘Is there basic foundations of things that people might not have a grasp on before we proceeded to the next set of the curriculum?’ ‘What are ways that we can approach each of these different students who are all using the same solution, but are facing different problems, either at home, at school, or in their own time?’
I think that even though AI is like a blanket, coverall solution, using a coverall solution of just getting rid of it, I think actually does a disservice to a lot of students who are using it for very specific, different reasons. You know what I mean.
Jeff Young
So that is one high school student's perspective, and by the way, she has already worked on a class project to try to use AI to help people with Crohn's disease. As I understand it, she was trying to sort of cook up this idea that would help people with this disease manage their diets better, to find what foods could be triggers.
So she is thinking about real world aspects of AI outside the classroom, in the world that she'll enter when she's done with school.
[Music]
Jeff Young
So what can educators do in this challenging situation? In my conversation with Rebecca Winthrop, I asked her what advice she gives teachers.
Rebecca Winthrop
I'm out and about at the moment doing book talks to districts or schools, and this question comes up a lot, and I suggest, ‘Do an AI audit,’ even if you're in even if your institution isn't doing it, you as a teacher, do an AI audit of your instruction, your homework and assignments.
Jeff Young
And what does an AI audit look like?
Rebecca Winthrop
The AI audit is not necessarily how to use AI. A lot of sure teachers were like, oh, I need to figure out how to use that.
I was like, that's fine, too. You should figure that out.
But the first if you do nothing else, the first step of the AI audit is take everything in your syllabus for the year and go on Gemini, ChatGPT, Deepseek, whatever, and pretend to be a student and see if you can be hacked.
If it can be hacked, find a different way. It could be. You're okay. We're spending through, you know, we're spending the entire class, everyone with a pencil and paper, because we haven't equipped educators to do a better job. You're writing me an outline with a pencil right now, it'll be your paper right now, right? So you're turning it in, and I'm going to give you feedback, you know, like that's one thing, or you are going to.
Or there's oral presentations, or you can, you know, write me a brief draft, handwritten and then I want you to go and use AI, you know. I mean, educators are gonna have to figure this thing out for themselves.
Another thing that I've seen high school English teachers do is say, Okay, here's our five books of the year we're going to read them. And you are going, you can, you can use AI. That's fine. You can use AI. But what we're going to do is spend all year learning how to have active listening and deep critical dialogs around books and the themes. And like each class is you sit in groups of six, you have a little recording, and you have a theme to talk about, and you're going to turn in the recordings, and I'm going to give you feedback on your ability to have active listening and critical dialog and and your content, your substance.
Jeff Young
I want to get back to the Ezra Klein conversation you had, which, so those who haven't heard it, should check this Ezra Klein show out from a couple months ago, the headline they put on it was, “We Really Have to Rethink the Purpose of Education,’ which is a quote you said during the interview.
So you noted that since AI can write the essay now and ace the test and all these things we know as possible, that maybe schools do need to be rethought a little bit.
What do you see as the purpose of education in the age of generative AI? And then talk about how to get there.
Rebecca Winthrop
If I could have schools do four things.
I think it would be, get kids into Explorer Mode, so basically, really good learners, and you need to be motivated and engaged to be a good learner, because kids are going to have to be agile learners and learn a bunch of new things and navigate this whole thing.
The second thing I would have schools do is really connect with build the social, interpersonal relationship connectivity piece, because I'm really worried that kids are going to get less and less of that, especially with AI companions and friends out there in places where they do have access to AI and so a little bit of counter cultural in person, you know, which includes, you know, social, you know, ban on cell phones, usually, largely unless you have a pedagogical, real purpose to use it. But then it's a tight grip with educators using it for something very specific, and it goes away. You know, don't have cell phones at recess, etc. This is where kids are interacting — learning to interact, listen, communicate, be in conversation with each other. So that's the relational sort of purpose.
Jeff Young
Just to highlight that real quick, what I'm hearing there is, like, cell phone bands are we're finally having this conversation and experimentation about how useful that is in learning. But you're like, go beyond that in the world of AI to be like, extra attentive to let's have times where kids are really encouraged to develop these interpersonal relationships.
Rebecca Winthrop
They have to learn to talk to other people face-to-face.
I mean, so many kids we talked to talked about being afraid to talk to someone they were texting with at school who was a friend or maybe a romantic interest, and they would pass them in the hallway and they wouldn't say hi to them, because what if they responded? What are they going to say back? They don't have the time to think about their answer like they do in text. And now text is like, you know, screw that. I don't even want to bother texting. They might text something I don't want. I'll just talk to my AI friend who's totally sycophantic and always loves me.
So I'm worried about that piece of children's well being and upbringing. Yes, absolutely.
So the third thing, I think, is ethics, a really strong conversation with kids about: What's the world we want? What's the life we want? What's the community we want? What's right? What's not right?
Like you need an ethical orientation around this social connection as well as their learning abilities.
And then the last, to me, would be which is connected, but not quite the same at the civic purpose of school. Schools really are the main places in every community across this country where people have to get together, not from your neighbors or family, right specific dialog, like, honestly, when we are when the rest of the, you know, mechanisms where that used to happen is falling far, far away, and we're increasingly polarized.
I mean, I really, you know, do put great hope and faith into the institutions of schools. It puts a lot of pressure on educators, but, you know, it's one of the original purposes of a common school movement. Anyways, in the United States, that was the, one of the big arguments in the mid-1800s of why we need a common school movement anyways, why should American kids have a similar sort of experience?
And it was about, we're a new country. We need to hang together. We need to have some shared vision identity, some base of knowledge that we have a common set of understandings we're working from and dialoguing around.
AI seems to be providing a chance to step back and not just be in Passenger Mode ourselves when thinking about whether education systems are working.
But doing that will require some deep thinking by us humans and real engagement with the process. And I worry a little bit about AI's role in engagement.
I'm not sure if other people are feeling this way. I'd be curious. But when I'm writing articles these days or doing the knowledge work that I have done for years and years, when I hit a moment of writer's block, or when things are kind of hard, I now sometimes have this question pop into my head, ‘Could AI just do this for me almost as well?’
I am not outsourcing what I'm doing to AI. But even having this thought is a new cloud over my work sometimes. It certainly does not help motivate me to keep going.
And it seems like that could be happening to students more and more, at least in the education systems we have today.
If you have thoughts on the issues that were talked about in this episode. I would love to hear them. You could shoot me an email or send a voice memo to Jeff at learning curve.fm. Some of you have already been writing in that really makes my day. Thank you all. I'm working on a listener feedback segment or maybe even a whole episode soon. Here's a quick sample of one of those reader messages.
Listener comment
My main pushback comes and responds to your hypothesis that we professors only want to use Gen AI for the things we like least about teaching.
So stay tuned for that.
This has been Learning Curve, episode three. This is an independent podcast, and you can support the show by giving it a rating or a review or check out learningcurve.fm to find out how to be a paying supporter and get bonus content. Our music is by the amazingly talented musician who goes by comic who and the episode art is AI generated by Midjourney.
I will say I far prefer working with a human than a bot. I'll be back in two weeks with another episode.
Thank you for listening.
