Loading summary
Rebecca Winthrop
Given that kids can access generative AI almost anywhere outside of school, it means that educators are going to really have to shift their teaching and learning approaches inside of school. My main recommendation to educators is pretend you're a student in your class, assign yourself your assignment and go try to do it with AI. If you can easily outsource it, change your assignment.
Vanessa
Hi, Cara.
Cara
Hi, Vanessa.
Vanessa
It is such a privilege to have back today Rebecca Winthrop, who was here not all that long ago with her co author, Jenny Anderson about their wonderful book Disengaged Teen. And it's pretty special to have you back, Rebecca, because you are here to talk today about a really, really complicated topic which we're going to dive into. But for listeners who don't remember Rebecca from her earlier appearance, she is the debut, right for her podcast debut on the show. And now she's going to get like a blue velvet jacket, like snl.
Cara
I'm wearing a blue velvet jacket.
Vanessa
Oh, my God. I actually have a really good blue velvet jacket. Wait, stop interrupting my very important introduction of you because Rebecca is an incredibly impressive person. I can't remember what Car said before we started recording. She's like, rebecca, you're just awesome. Is that what you said?
Rebecca Winthrop
You're just amazing?
Cara
Pretty much. If I didn't.
Rebecca Winthrop
That's what I meant was something along that line.
Vanessa
It was something along those lines. So Rebecca's the director of the center for Universal Education at Brookings, and her work of decades of work, including the book she wrote with Jenny, and her work at Brookings and beyond, is centered on developing and advocating evidence based strategies that bring people together, families, educators, policymakers, to help children maximize their potential. I mean, if there is no greater purpose on earth, I do not know what it is. And today Rebecca is here to talk to us about a big, complicated and a little bit scary topic, which is education and AI. So welcome, Rebecca.
Rebecca Winthrop
It is so nice to be back on your wonderful, wonderful podcast, which I love. Send to all my friends. Adolescence.
Cara
We love that. We love that. I mean, let's sort of start where Vanessa left off, which is because I don't want this to be a fearful episode, but we have to say that not so quiet part out loud in the beginning. This is the rapidity with which AI has entered all of our lives is astounding. And it generates panic among some, excitement among some. I will say for my two cents, I hold both of those things. I think many people do. But we have to deal with the fact that it will fundamentally change the, the way kids get educated and At Brookings, you embarked on a study that unlike most studies that go slowly and look at treasure troves of data over years, you were like, guys, this is moving so fast. We have to grab as much data as we can over a quick snapshot in time and, and get a pre. Mortem. That was your word, and it's a perfect word. You got to get a pre mortem out there in order to start moving the bow of the ship a little bit. If we are to have, if we the collective, we are to have any control over how AI is changing education and how education needs to change with AI. So can we start with your description of this study and what you were looking for and how you did it?
Rebecca Winthrop
Yes, absolutely. So you did a great job. Capturing was really three things. One was we wanted to do a pre mortem, which is an actual research methodology, which is basically, basically move the autopsy forward, take evidence and use your imagination.
Vanessa
Is the just do an autopsy on a body that's not in front of you yet,
Cara
live body.
Rebecca Winthrop
Basically. We weren't doing a pre mortem to see how, how to move the bow of the ship, although that's the result we were seeing. Do we need to move the ship? We just wanted to get the lay of the land. Where are we at? Super objective. And part of the reason we took this pre mortem methodology was exactly what you said, that when you do really intensive, good, high quality longitudinal research, it could be five years until it's done and it's out in peer reviewed journals. 5 years is way too late to sit on the sidelines as educators, as parents, as school leaders, as coaches, as therapists, anybody who works with children and students, we cannot sit this one out. So that was number two. Our number two reasoning was we know from social media that the people who work and love and care for kids on a daily basis were not at the table when social media was rolled out. And we need to be at the table this time. And what better way than to actually hear the voices of those people? So we did focus groups, consultations, interviews across 50 countries, including the US with students, parents, teachers, education leaders, technologists, policymakers, academics, funders, like the whole kit and caboodle. The second thing we did is we looked at every study we could get our hands on. They come out fast and furious, they're of variable quality. So we ended up really focusing on 450 studies that we looked at. So this is a year, it's a year long sprint. And then the third thing we did was a Delphi panel which is a very particular methodology where you gather a group of really top thought leaders on a particular issue and really start. It's a sort of complicated methodology where you're posing them, sharing back at them pieces, sort of the data picture that's emerging and they're weighing in. And it's a really good methodology when a field is uncertain and that you don't really know what the trend lines are, you don't have established frameworks. Those are the three things we did. And the goal was to see are we headed in the right direction with AI and education and if not, what should we do? So it was really three research questions. What are the risks? What are the potential risks? Generative AI. It's not any AI because AI has been around for a couple decades, used in education, generative AI and children K12 education. Really, how do you mitigate those risks and what are the benefits and how can you harness them? Those were our three big questions. Risks, mitigations, benefits and Rebecca, in order
Vanessa
to ground this conversation in what you found, I'd love for you to talk about your view of the role education and schools plays in kids lives. Because we can't have one part of the conversation without having that part of the conversation.
Rebecca Winthrop
Absolutely. Okay. We started this research investigation with a couple of core premises that exist not because we made them up out of thin air, but because they're in the research over the last couple decades. And I want to tell you what those are. But first I want to give you the corollary example of why we started with these premises. Because when social media was rolled out over a decade ago or so, we knew that social comparisons for adolescents can be quite dangerous. We knew that before social media existed, we did not need to learn that through social media. So you know, the idea is other people have been at the table when social media was rolled out. Perhaps we could have shaped it differently so that, you know, the positive pieces were really the things that were put
Vanessa
out in the world.
Cara
Rebecca, I just want to ask you really quickly before you get into your version. Do you feel that it was a missed opportunity to be at the table before, or do you think also we've learned a lot of lessons about not forcing ourselves to that table. And so that is an element of what is happening now, 100%.
Rebecca Winthrop
So I think it was a huge missed opportunity. In reality, do any of us have a crystal ball and if we had been at the table, would things have turned out differently given all sorts of financial incentive structures? I'm not sure. However, I really feel we're in a different place this time around. I think all the activism around technology, children's mental health, has really done a great job at bringing the public to be wide awake to these issues. And particularly, I would say, shout out to every single mom, dad, grandma, auntie. The parents have been a huge driver trying to demand accountability and change. And I think we're at a very different moment in time. I think when social media was rolled out, people weren't really thinking about it. They didn't think it would really change much. There was a lot of good pos. I mean, if you look back at the discourse, the discourse was it's gonna connect everybody. It's gonna unleash children's creativity, it's going to create world peace. I mean, that was almost the narrative, like a Miss America pageant, because we're gonna be able to connect and know each other. And of course, it has done some of that in some pieces. Like, you can I meet kids all the time who've learned an incredible skill to bake something or cybersecurity this or that, because they've gone on a deep dive on YouTube because they were motivated and excited, had access. But the problem is that the negatives are so intertwined with everything, and they're really, really negative.
Vanessa
So having learned some lessons about what happens when stakeholders and thought leaders who are really thinking specifically about kids, how do we think about kids? Right. If we go back to the initial question, which is like, what role does education in schools? What do they do? What role do they play? Another way of asking that question is, what do kids need from schools and education? And not just, like, going to college or whatever, but what are the underlying fundamentals? The premise by which you guys conducted this study as, like, the core value of education.
Rebecca Winthrop
So thank you for bringing me back to your original question, Vanessa.
Vanessa
That's what I'm here for.
Rebecca Winthrop
That's what I'm here. So there were three fundamental premises that we held. One is learning is social. This is true. No one can deny it. We have neurobiologically been wired over millennia to learn in relation with others. Children are not little mini scientists alone in their room running labs. They are not little mini machines. We are not little mini robots. So learning is social. And children's brains develop the way they are used. So whatever your kids are doing now while they're with you for the first 18 years of life, that is fundamentally shaping how their brain develops. And so we have to hold that in more mind when we think about what we want AI to do or not, and we'll get to this in a second. But I often find there is this narrative of adultism around AI and education. I run into it all the time in my field, interestingly, not with educators, but with, like, peer leaders who are a little bit adjacent to education, not inside. And they say, but I use AI all the time. It's amazing. I'm doing this, I'm doing that, I'm doing this. And I said, you know what? Me too. But I didn't use it when I was young for X, Y and Z purposes and hence develop critical thinking skills. So that's, number one, how learning happens. It's social. Children's brains develop the way they're used. They need some experiences to really fully develop their critical thinking, judgment, agency, et cetera. Number two, schools are about more than just academic learning. We learned that in Covid. We knew that in education Forever, but the public learned that in Covid. It's around socialization. It's around learning to live together. It's around young people, for the first time ever, really getting to know people outside of their family and their direct neighbors. So there's an entire sort of social, civic purpose behind school that we can't forget about. And then the third thing is really around how we know technology influences education. It's usually never 100% about the technology design, and it's about how technology is absorbed into systems. So we know that if you sort of take technology and sprinkle it around on the top of a classroom or a school without really thoughtfully, with good pedagogical practice, good teaching and learning practice, embedding it into the heart of teaching and learning, it's not going to do much and it might distract and make worse outcomes. We've seen this, we've seen this over decades that some of the education systems who've invested the most, this is a while ago in, like, IT and computer labs and all this stuff, and tech in schools have actually not been the best performers. What's happened is it's teachers are spending a lot of time futzing around with the tech that isn't really integrated into their lessons, and then they get less teaching time. Right?
Vanessa
And you cite research that shows that at the stage where there's been more and more technology integrated in the classroom, all the measures of success, Success, Right. What we can. The quantifiable stuff, reading, math, science scores have gone down. Right? Those scores have gone down. Plus student sense of belonging has gone down. So it's not just that being in school is important, being surrounded by others, but the role technology plays even when in school can affect that social experience as well as the skill building experience.
Rebecca Winthrop
Yeah, and I would just say for that, Vanessa, that it's basically this great quandary. There's, you know, last couple decades we've seen a ton of tech innovation. It's pretty amazing. There are incredible use cases of technology. I've seen incredible use cases in education and technology when done really well.
Cara
We'll be right back, but first a word from our sponsors. I have had photo books on the top of my to do list for years, literally, and it bums me out that I cannot seem to find the time to do this because there is literally nothing I enjoy more than looking back at old photos of my kids. So it's why I'm thrilled to have discovered journeyprint. Journeyprint is easy to use. It takes seconds to auto arrange photos into a beautiful book and it'll even add location maps and dates. Plus, journeyprint uses sustainably sourced paper and for every order placed they plant a tree. The photo book has moved from a slightly dreaded task that will take so much time to quick and easy. Now the thing that's going to eat up my time is flipping through all those incredible memories. If you're looking to have a memento of your family's amazing moments, or maybe some other awkward ones, give Journey Print a try. Go to journeyapp.com podcast awkward and use the code awkward to get 30% off your next purchase. That's journey spelled J O U R n I n app.com podcast Awkward and use the code Awkward. I've always been about quality over quantity, especially in my closet. If it's not well made and versatile, it's just not worth it to me. That's why I love Quince. Quince uses premium fabrics like 100% European and linen, 100% silk and organic cotton poplin. They're all perfect for the changing seasons, but they're also built to hold up. And Quince works directly with safe ethical factories, cutting out the middlemen so you're not paying for brand markup or fancy retail stores. Just for quality clothing, I just got the Italian leather dual compartment toiletry bag in cognac and I cannot wait to use it as I hit the road for my spring speaking travels right now, go to quince.com awkward for free shipping and 365 day returns. That's a full year to wear it and love it. And you will. Now available in Canada too. Don't keep settling for Clothes that don't last. Go to Q u I n c.com awkward for free shipping and 365 day returns. Quint.com awkward
Vanessa
take our curriculum. We have a completely digital health and sex ed curriculum, k through 12, which not would not be possible and scalable at the level it is without leveraging all different kinds of technology. So we are huge believers in thoughtful, integrated, pedagogically sound use of technology in education. We're so proud of what we've done and yet we're also sensitive to what different schools tell us about how much technology they are choosing to or want to use in their classrooms. And I'm just curious what you hear from educators, Rebecca, on this sort of broader topic before we get started specific into AI.
Rebecca Winthrop
You know, there's a big hot debate. There actually has been a debate. It's not new, but it's getting more attention I would say in the last couple years around the role of techno student facing technology in classrooms. So your curriculum, I think, having seen it, I believe, is teacher facing.
Cara
Yeah, yeah, it's teacher le.
Rebecca Winthrop
So there's a lot of things that you can do, what I call back office technology that are incredibly valuable for schools. It's around efficiencies, bus schedules, all sorts of things and also teacher facing supports that is qualitatively different from a student facing technology interface. And the student facing technology interface can be really tricky. So there is the issue of phones in school, which we all know and agree and the US is actually, frankly thanks to Jonathan Haidt, now on board with and lots of other other countries have been on board for years around this idea of, you know, I'm sitting here trying to teach you, what are you doing looking on your phone? Now there are exceptions to that in really remote areas that don't have science labs that, you know, teachers need kids to use their phone. But it's very pedagogically controlled. So I talk about, I'm in favor of a bell to bell phone ban with a pedagogical exception. That means a teacher who might need it, who might not have access to other forms of information, can, you know, has control over the phone for a particular teaching task and kids can use it. There's lots of other interfaces that kids have in technology. And this is the debate that has long existed in the ed space but is now getting more public attention, which is there's also all sorts of laptops and iPads and this and that. And you find, I find some teachers who have really careful, judicious Use Love it, love it. Think it's great. They are tracking all their assignments. Kids are interacting, kids are. And then mainly kids are using it to do things they couldn't do, like make videos and do podcasts and weave stuff together and do deep research. And that use case seems to be one where kids are benefiting, educators are benefiting, and the kids are not in front of a laptop the whole time.
Cara
Mm.
Rebecca Winthrop
It comes out at a particular point in time. There's another use case which is one to one. Kids are doing everything mediated through a screen. And I have to say, I've walked into many a classroom, not all, not all, but many a classroom when half the kids have got their screens up and they're playing Minecraft unblocked or Zelda unblocked, because you can get any video game unblocked on a Chromebook.
Vanessa
Or gambling or shopping or porn, right?
Rebecca Winthrop
Yeah, or porn. Yeah. I haven't actually seen that in a classroom myself, but it certainly happens. Or going to the bathroom to take a bathroom break and watching porn in there, like that happens too. So that's not good for learning.
Cara
So I just want to introduce the third leg of the stool.
Vanessa
Simply, simply said that's not good for learning.
Cara
But I want to reintroduce the third leg of the stool, which you wove into your answer earlier, which is that there's this sort of teacher time component. Like teachers, educators only have so much time to learn and explore the technology themselves. And I have to imagine that there is a very big. I don't want to blame it on age because I don't know that it's about age anymore, except to say that there's a very big spectrum, wide spectrum of educators. On one end being the, the ones who are extraordinarily comfortable with the use of these various technologies and speak the language and use the apps as fluently as the kids do. And on the other end of the spectrum are the other extreme sort of Luddite educators who have done often this kind of teaching for a very long time. They know kids, they know what works and they know the benefits of the non technological approaches. And it strikes me that one of the things that makes the introduction of student facing and then that sort of code for AI driven, I think a little bit education. One of the things that makes that sort of most fraught is that you've got educators who no way do they have the time, inclination or bandwidth to, to get themselves to the place where they need to get to, to master the way these tools are working to be able to use them sort of safely and effectively. And I put myself in that bucket. I am, I am a Luddite and I'm trying and I educate myself on AI all the time. And, you know, I'm doing the thing where 20 minutes a day, I'm trying the different platforms and I'm learning what they do. And I, you know, all the, all the advice to the old folks is what I'm following. And yet when I read your report, I was sort of gobsmacked by the notion that if I was standing in front of a classroom and AI was a big part of that classroom, I don't know how my students would fare because of where I come from in life. And I feel like. Was that an element of it?
Rebecca Winthrop
I have a. Absolutely. I have a great, great deal of sympathy for teachers because none of us want from one day to the next. Our entire job that we've trained for, really good at, worked hard at, totally upended by a commercial product release, which is basically what happened when ChatGPT released. So educators are going to have to totally change their teaching and learning practice. Now, when I say totally, I mean there's some educators who are already doing the type of pedagogy in teaching and learning that AI is not a problem for. But most of the way that we do school, especially in the United States, the majority, particularly sort of the traditional high school, frankly, be a public charter or private. It's more the approach to learning is task assignment and task completion. That is actually kind of the core of the lifeblood of most teaching and learning in classrooms, which is read this, write an essay, or here's we're going to teach you about this math problem set, do your homework or study for this test and you know, come take it, or you know, what, what have you. And given that kids can access generative AI almost anywhere through social media, through like it's, it's everywhere and they can't even avoid it outside of school, it means that educators are going to really have to shift their teaching and learning approaches inside of school. My main recommendation to educators is right now, do an AI audit, pretend you're a student in your class, assign yourself your assignment, and go try to do it with AI. If you can, easily outsource it, change your assignment.
Cara
Such good advice.
Vanessa
So, Rebecca, before we get into. And that's a fascinating example. And again, teachers who have been doing this work for, I mean, my kids have teachers who have been teaching for decades who have honed and crafted and evolved, and now they've got to deal with this. And I feel incredible sympathy. I'm wondering, you talked about a sort of like the US based approach to, to education, right? Task oriented, task outcome. Can you talk for a second about philosophically, what another avenue would be? And so that it's, it's not just about how do we handle AI, but also like, is there an opportunity here for educational approach in general to, to shift?
Rebecca Winthrop
I actually think that's the great. The greatest potential of AI is to shift education in a direction that is more aligned with the learning sciences and that we've wanted, we, the Royal education community, we has wanted to shift for a long time. So not shifting to adapt, not just to make a change, just to accommodate AI, but AI could be the fuel, the wind at the back, to make shifts that we've wanted. So the core of that shift is pairing knowledge acquisition. That is my academic learning, that is my content. We cannot throw the baby out with the bathwater. Kids have to learn content. The other thing I would tell every parent, make sure they learn their facts. Like it's become even more important. Because how else do you discern on deepfakes and, you know, wrong information, which is proliferating. So that's knowledge acquisition. That has been the biggest focus of education systems and schooling. But we have to pair it with knowledge application, which is, I'm taking biology and I'm also taking some math. And now one of the projects I want to do is replant some trees in my local park and maybe I do it with my team and maybe I make a grid and I figure out where to plant them and I research. And you could use AI here. I research what are the native species. It will be much more efficient to research that with a deep research generative AI tool. And it's okay. You learn how to do deep research with generative AI because you're using that as an input to applying your knowledge. And then you're going to check and you're going to discern is it giving me the right information? So anything that marries knowledge acquisition with knowledge application is the direction that we need to go. There are many, many examples of individual schools that do this. Some districts that do this. It's just that the US is ginormous and very decentralized and that's not the norm at the moment. But it would be great if that sort of wind of AI could push it to be the norm. Because kids learn better that way. They remember, it's more exciting, they're more engaged. It makes it real, It Makes it meaningful. There's so many benefits to doing it that way.
Cara
Okay, so that sort of is a nice segue to getting to the results that you found. And I think, you know what would be helpful is starting with the big headline, the all caps headline, and then kind of breaking it down a little bit into, you know, as you said the top the study was not you. You didn't make a hypothesis and set out to prove it or disprove it.
Rebecca Winthrop
Right.
Cara
You basically went out asking an open ended question.
Rebecca Winthrop
Right.
Cara
So can you tell listeners what did you find? And then how do you break it down a little bit?
Rebecca Winthrop
So we found that currently with the current deployment of AI in education, the risks of AI are overshadowing the benefits. That's the big takeaway. And it's not that we didn't find good beneficial uses of AI in education. We did. It's that the type of risks are really fundamental to children's learning and kind of undermine the very benefits themselves. So shall I walk you through sort
Cara
of some of our.
Vanessa
Yeah. Because Rebecca, this is such a layered and multifaceted conversation and if it's okay with you, if you're not going to lose your train of thought as you walk through what you found, Cara and I may sort of pepper in. Well, what about. And how do you handle this? Because there's so many permutations to each application of this. So yes, tell us if we're throwing you off your game and you need us to stop interrupting you.
Rebecca Winthrop
It's more fun to be dynamic otherwise me talking. So let's first talk about the context of AI use for kids because that really sets the scene. So number one, kids are accessing AI a lot. We know that at least 90% of kids say they use in the US say they use AI in their personal life. Recent studies really recently that are forthcoming say that 85% of kids surveyed big national surveys use it for school related work. So kids are using it. Number two, they're accessing it everywhere. It is not an app they have to intentionally download necessarily. It is not a device that their parents have to buy them. It is software. It is showing up in every single digital domain that your kid is already accessing. I had a conversation with one student, high school student who said, well, my school banned ChatGPT, but you know, we all just use Snap, you know, Snapchat AI or meta AI. So they go onto Snapchat. They in Snapchat, they're, you know, getting their streaks and they have my AI friend, which is an AI friend or companion that they talk to for fun. But then, you know, they take a picture of their math homework and upload it and say, please give me my math answers or write my essay for me. Because my AA friend has built. Built on ChatGPT. So it's everywhere. So that, so that's the context. So in school, out of school.
Vanessa
Rebecca, do we have an average age of like first use or. Because so much of this is so porn parallel that it's wild to me. And I'm curious if you have any data on ages.
Rebecca Winthrop
I have not in part, I think, because a lot of the surveys are looking at 13 to 18. Because that's the legal. Right, right. And it's a little awkward. I will. So those are the big nationally representative surveys of which there's multiple. I will tell you from our qualitative research, all our focus groups, you know, interviews, consultations, kids are definitely using it before the age of 13. Not a surprise.
Cara
Well, and, and I mean all you need to do is a Google search and Google AI is now the top third of the PA and it's. Right. So like it's in synergy.
Rebecca Winthrop
It's everywhere.
Cara
Everywhere. Yeah, it's everywhere. Yeah, yeah.
Rebecca Winthrop
So that's the context. And this idea of blurred lines I think is really different. So you know, Roblox is a massive gaming platform. Massive. I think there's more users than people on the planet. Ants with people on multiple accounts. They have started a massive learning hub for like literacy and numeracy and science. So like in your big gaming platform you have learning games and then, you know, on social media you can.
Cara
Which on the surface sounds amazing.
Rebecca Winthrop
Right. Like, I'm not saying it's all bad, it's just that like the, the lines are blurring and then, you know, in social media you're doing your homework in addition to talking to a AI companion about how annoying your parents are. Right. Or getting dating advice. Like it's all blurred, which is one of our big findings. Like it's. Kids are living in a sort of 360 stew and AI is in every interface, basically. So what does that mean? What did we find? Well, let's start with the benefits. So we did find, I mean, basically really, we found students have learning experiences that are enriched by AI, sort of AI enriched learning experiences. And these experiences usually happen when AI is bounded to a learning activity. It's bounded to vetted content. Not just the whole huge world, you know, World Wide Web, which is what AI is built off. It's bounded and integrated into A high quality pedagogical which is sort of teaching and learning practice that makes sense. So what happens with these AI enriched learning experiences? We have kids who are neurodivergent or living with disabilities being able to access learning in new ways never seen before. One of the most extreme and like moving examples is kids with real communication problems with aphasia. I believe I'm pronouncing that right, Am I? I'm like, I only read it. I actually don't work in this area. Is digital copies of their voice so they can all of a sudden communicate in the classroom with their educators. I mean, in moving and groundbreaking for the families of these kids. Right. Like, I just almost tear up every time I go through that section of the report. So incredibly great. There we have learning experiences that are really helping kids learn because they are making interactive static content. You know, textbooks, worksheets, any sort of digital content where you can, that's sort of bounded to the actual vetted content where kids are reading along and they say they don't understand a vocabulary or say they don't quite get a concept. But you can say, hey, I, I just don't get this. I've read it three times. Can you help me? Or, you know, could you give me an example with sharks? I love sharks. Like, and it's interactive, which is amazing. And I think we're going to see that be incredibly powerful. Early studies have shown a lot of them are by the tech companies themselves. So we need to. But I've seen some of these products and talked to kids who have been like, it's amazing. I like, I feel like I'm talking, you know, to it almost like a teacher. So I think areas around getting feedback are going to be amazing. Kids being able to get feedback quicker. There's. I'm on the board of this great organization called Junior Achievement. And part of what they do is entrepreneurship training and they use real entrepreneurs to give feedback. Now they have, they're developing this pitch simulator where they go, the kids like, do their pitches, get AI feedback to prep them, like you know, to go. And that's, that's great. So those are all.
Vanessa
And what about studying? I mean, I've heard from a lot of young people that they find the most helpful use of it is when they're prepping for exams and they've, you know, done the problem set or they've done all the homework and they want similar questions to practice on. They can get that. Now that's what kids tell me about. I'm sure there's all sorts of other.
Rebecca Winthrop
A hundred percent. And the caveat I would put is. Motivated kids.
Cara
We'll be right back. But first, a word from our sponsors. Your planet is now marked for death.
Rebecca Winthrop
Marvel Studios the Fantastic Four First Steps is now streaming on Disney. We will protect you as a family. Light em up, Johnny. Marvel's first family is certified fresh on Rotten Tomatoes.
Cara
That is fantastic.
Rebecca Winthrop
And critics say it's one of the best superhero movies of all time. Marvel Studios the Fantastic Four first steps now streaming on Disney. Rated PG 13.
Vanessa
What time has it been?
Rebecca Winthrop
It's clubber time. So the motivated kids are rocking like nobody has seen. And there the problem is, there is a huge gulf between the motivated kids and the unmotivated kids, as Jenny and I have shown in our book the Disengaged Teen. So the highly engaged, motivated kids, kids, one girl told me, I take all my class notes, I put them in AI, make a podcast. I listen to it three times while I'm like, walking the dog and cleaning my room and I'm ready for my test.
Cara
Like, wait, I'm sorry, say that again. That's amazing.
Rebecca Winthrop
Isn't that great?
Cara
Takes all of her notes, turn them
Rebecca Winthrop
into a podcast because she loves listening to stuff, right?
Vanessa
She's an auditory learner.
Rebecca Winthrop
And that was three times while she's like on a run, while she's cleaning her room, like, she's helping make dinner, whatever it is. And she's like, it's great. It's working great.
Cara
I'm totally trying that.
Rebecca Winthrop
Yeah, try it. Flashcards, flashcards, all sorts of cool stuff.
Cara
So that's amazing.
Rebecca Winthrop
The motivated kids. We already see a sort of inequity between, you know, engaged and disengaged kids. But I think we're going to start seeing a huge gap between motivated kids harnessing AI to. Because they're motivated to learn, they're insatiable and they're just flying forth. And then the unmotivated kids who are using it to outsource their thinking, or
Cara
the kids who are motivated by something else.
Rebecca Winthrop
Yeah, yeah, yeah, yeah.
Vanessa
Right?
Rebecca Winthrop
Like, absolutely.
Vanessa
What about when this research says it's making us dumber?
Rebecca Winthrop
So that is the downside.
Vanessa
So to elegantly transition to the downsides,
Rebecca Winthrop
benefits are AI harnessed in the right context, linked with, you know, and bounded, can really help learning. The downside is that most AI use at the moment is not that AI, as we've established, is everywhere. And so what's happening is most kids, when it comes, most students are replacing their learning experiences with AI because actually, as we know in the US 50% of middle school and high school students describe experiences at school where they're just like, I'm coasting. In our words, Jenny and me and I, it's the passenger mode. I'm just coasting along. And I really think AI is making that group bigger. So what we found is because kids are offloading their work through AI use outside of school especially, it's really harming kids cognitive development. This is through cognitive offloading. So every parent out there should know what that word means. That's one sort of term I would grab, which is basically letting AI do your work, your thinking for you. And as we established, kids develop brains develop the way they're used. So if they aren't actually being exercised like muscles in the body, they aren't being exercised to think critically. They're going to develop those neural connections less and less. And I always say the classic sort of struggle is this is a poor English teacher who assigns a seventh grade essay not because she expects great works of literature to come out of it, but because she wants to train kids to read text, identify arguments, look at data, assess which side the data lands on, and then develop a thesis and then persuasively describe their position. That's all, all what's going on when you write an essay or write a good essay. So we are going to have to have other ways of doing that or do it in class. So cognitive offloading, another piece is undermining emotional development. This is gets to the sort of social, emotional learning piece with AI friends and companions. One out of three young people in the US say that they prefer that they, you know, talking to an AI friend. They're equally happy to talk to them or prefer them over real human beings. And that is, you know, just, they're only going to get better. That's all I'm saying.
Vanessa
The AI friends, not the children themselves.
Rebecca Winthrop
AI friends, not the children. So that's a huge worry. It's a huge worry. We're also really worried about children's ability, basically resilience. Learning is hard. And the way kids develop learning is through productive struggle. There's a lot of evidence around this, which is you have to try things that are a little out of your reach and you gotta struggle. Not massively out of your reach. You don't give a kindergartner who speaks English like a second, you know, a 12th grade German textbook and be like, go. But a bit out of your reach. And you've got to struggle through it. And that builds a learning Muscle that every kid is gonna need in their life. And AI is frictionless and removes that struggle.
Cara
Can I ask Rebecca that? You know, you talked earlier about the importance of social interaction at schools and how one of the primary responsibilities of a school is to build not just cognitive learning, but also social learning. Right. And to that end, now when we talk about the risks of AI, it's both this cognitive offloading and also there's this risk of AI companionship. Right. And now I'm trying to balance that in my mind with schools that are exploring the use of AI, student facing AI in their classrooms, in their regular daily schedule and who are, who are saying, our kids love school. They love school more because when they come to school, meets them where they are, they're not bored in the classroom, they're not over to your point, you know, of the friction, frictionless, like they're getting what they need from AI. And I'm wondering how, as this ball is traveling so fast down the field, how do we measure that? And I guess more specifically, just because a kid likes it doesn't mean it's the better way. But if a kid likes it and they show up in the environment, we have so many problems with school recidivism, kids not wanting to go, like, what's the balance there?
Rebecca Winthrop
I'm less worried. So in the wonky education terms, it's the zone of proximal development, which is literally the sweet spot for kids learning, where it's not too easy and it's not too hard. And guess what? Guess who's great at that? Video game designers. So these tech folks can totally do this. I'm less worried about the examples we talked about where it is helping the English language learner, the kids with dyslexia, the kid who, who just doesn't come from a family that understands the cultural references in a text, right. To really access it. The example I talked about of like, you know, interactive text based, digital based learning materials or it could even be virtual reality, which is amazing. I'm less worried about that because those are really helping kids get into their sweet spot of learning. What I'm worried about is kids getting used to in an instant having an essay written for them or math problem set be done for them. And a lot of it again is because we have this task completion approach to passing our grades. I think teachers are going to have to do a lot more in class presentations, a lot more group projects that ask for more, that ask more for kids, not just can you master this knowledge and do some homework and you know, take some assessments, but that's an input. Master this knowledge and show me what you can do with it.
Vanessa
Rebecca, let's build on that idea. What are your other recommendations along those lines?
Rebecca Winthrop
Yes, we have a bunch of recommendations that are, that we have 12 specific recommendations that basically fall into three buckets. We call them Prosperity, Prepare and protect. Prosper is shift how education is done so that it can help kids prosper. And AI can be a tool in that. That's the project to redesign a park. And you use AI to help understand what the native species are. The prepare piece is really important. It's about getting education systems ready for an AI world. And a lot of it is around holistic AI literacy. And a lot of holistic AI literacy doesn't necessarily mean kids are sitting in front of a chatbot. You can start with ethical discussions about what are humans good at, what is AI good at, what should we do? And by the way, I've seen teachers do this with sixth graders, fifth graders. Kids are thinking about this already. They're thinking a lot about it. And we should engage them in that discussion. And kids should be at the table co designing AI use policies and procurement policies. Honestly, like they're really at the helm. And when we talk about holistic AI literacy, again in the prepare bucket, it's not just how do you use these tools, it's how do you create new things with them that that come from within you, which is really important. And how do you design AI? Because it is not the laws of physics. This is not some, you know, God given property from on high. We made it, we can change it. Kids should know that they can change it. So that's prepare and then prot is really all the safety guardrails that we absolutely need to be put in place, especially around AI friends and companions, so that AI can be safe to use for kids.
Vanessa
And do you have resources, guidelines like for a school administrator, a teacher, an involved parent, like where can they go to to use this guidance?
Rebecca Winthrop
Yes, well, we have a very long policy report, we have a bunch of smaller briefs on it, little condensed excerpts with our recommendations by actor and our executive summary and stuff like that. We also have the start of a parent facing set of briefs which are digesting the main messages of this task force. Because one of the things we found to bring, to bring this back to you guys is that families have been left out entirely of this conversation. It's part of why I'm so thrilled you're having me on. I so appreciate it. And families have to work hand in hand with teachers. Yeah, more than ever because this is actually a case where it's the out of school use that is really undermining the in school time learning. And we need to bring our families alongside. So we will have a whole. We have several briefs on what's in the task force and we're going to be adding a couple each week and would encourage people to check it out. I'll let the sort of Brookings Edu website and I will be posting a lot of it on LinkedIn. I will, I have a newsletter, Winthrop's World of education on LinkedIn. So I'll be, you know, posting that, you know, weekly there.
Vanessa
So we will link to whatever's available when the episode comes out. And I think the area we didn't touch on, which is that an ongoing frontier and wild west is policy around AI use, disciplinary measures, all of those things because we hear about that at every level of education.
Rebecca Winthrop
So it's a huge issue, especially with the false accusations which are proven by Stanford for neurodivergent and English language learners to be worse.
Vanessa
So if you are an administrator, an educator and you're interested in joining us to talk about the challenges involved in taking everything Rebecca's talked about and even having a pedagogically, ethically sound approach to AI use in your educational community, and yet you are still facing week after week all sorts of complicated realities around false accusations. What is the blurry line? How do you draw a line? What is the disciplinary actions needed or not needed with AI use? All of these big questions that affect the everyday lives of kids, families and educators. Write to us podcast.com we would love, love, love to continue this really rich and complex conversation that we started with you, Rebecca. We're so grateful for the work you do, for the thoughtfulness, for the balance with all of the different elements that go into this really, really uncertain and open ended reality.
Rebecca Winthrop
Well, it's a pleasure to be here. Thank you for doing it.
Vanessa
Thank you so much for listening. You can email us with questions, feedback or episode requests@podcastawkward.com if you want to
Cara
learn more about what we do to make this whole stage of life less awkward for everyone involved. Our parent membership, our school health ed curriculum, our keynote talks and more are
Vanessa
all@lessawkward.com and if you want products that make puberty so much more comfortable, visit myumla.com.
Podcast: This Is So Awkward
Hosts: Dr. Cara Natterson, Vanessa Kroll Bennett
Guest: Rebecca Winthrop, Director of the Center for Universal Education at Brookings
Date: January 20, 2026
This episode welcomes back Rebecca Winthrop to discuss the rapidly evolving intersection of generative AI and education. Tackling both excitement and fear, the conversation centers on the profound changes AI is driving in schools, how it impacts teaching, learning, and social development, and what lessons the education system can draw from the rollout of social media. Using findings from a global, multi-method Brookings study, the hosts and guest dissect risks, benefits, and critical recommendations for navigating this "wild west" era.
"Move the autopsy forward, take evidence and use your imagination." (Rebecca, 04:13)
“I think all the activism around technology, children’s mental health, has really done a great job at bringing the public to be wide awake to these issues.” (Rebecca, 08:59)
Rebecca outlines three research-backed foundational beliefs:
“If you just sprinkle technology...without good pedagogical practice...it might distract and make worse outcomes.” (Rebecca, 13:15)
“Pretend you’re a student in your class...if you can easily outsource it [the assignment], change your assignment.” (Rebecca, 25:54)
“Anything that marries knowledge acquisition with knowledge application is the direction that we need to go.” (Rebecca, 27:26)
"Currently with the current deployment of AI in education, the risks of AI are overshadowing the benefits." (Rebecca, 29:40)
(34:00–38:50)
“Kids with communication problems...can all of a sudden communicate in the classroom with their educators.” (Rebecca, 34:00)
“I take all my class notes, I put them in AI, make a podcast...I listen to it three times while I’m walking the dog and cleaning my room...” (Student, via Rebecca, 38:19)
(39:19–44:00)
“If [brains] aren’t being exercised...they aren’t being exercised to think critically, they’re going to develop those neural connections less and less.” (Rebecca, 41:19)
(46:31)
Prosper:
Prepare:
Protect:
Resource Guide:
On the missed chance with social media:
“It was a huge missed opportunity...I really feel we’re in a different place this time around...Parents have been a huge driver trying to demand accountability and change.” (Rebecca, 08:59)
On how AI will force classroom change:
“Educators are going to have to totally change their teaching and learning practice. My main recommendation...do an AI audit...If you can easily outsource it, change your assignment.” (Rebecca, 24:10 & 25:54)
On the spectrum of tech comfort among teachers:
“There’s a very big spectrum...some extraordinarily comfortable...and on the other end...the Luddite educators...they know kids, they know what works...” (Cara, 21:48)
On the key downside—cognitive offloading:
"...brains develop the way they're used. So if they aren't actually being exercised like muscles in the body...they're going to develop those neural connections less and less." (Rebecca, 41:19)
| Time | Segment Description | |------------|------------------------------------------------------------------| | 00:00–04:13 | Introduction, urgency of AI's rapid evolution in schools | | 04:13–07:45 | Study methodology and scope | | 08:59–11:17 | Lessons from social media rollout & parental activism | | 11:18–15:20 | Foundational beliefs about learning and tech in education | | 18:38–21:27 | Hot debates: teacher vs. student tech use, phone bans | | 24:10–26:40 | Need for radical shift in classroom practices | | 29:40–32:15 | Headline research findings: risks outweigh benefits (for now) | | 33:06–38:50 | Enriching, bounded AI use—special education, motivated learners | | 39:19–44:00 | Core risks: cognitive offloading, emotional downsides | | 46:31–49:44 | Recommendations: Prosper, Prepare, Protect, where to find resources |
This lively, pragmatic, and deeply-informed episode pushes adults in education and parenting to move beyond fear or Pollyanna optimism. Instead, it urges a collective, evidence-based reckoning with how AI transforms not just what kids learn, but how and why they do. The stakes: mental, emotional, and cognitive development for the next generation—making it crucial not to repeat the mistakes of the past as education crosses into a new technological frontier.
For resources and ongoing updates on these policies and recommendations, visit the Brookings Center for Universal Education website or follow Rebecca Winthrop's LinkedIn newsletter.