Rethinking School in the Age of AI

Hey everyone, this is Tristan and welcome to your undivided attention. AI

is set to disrupt every part of our lives in the near future. Healthcare,

finances, the job market, you name it.

And some of this disruption is a few years away, but there's one place where it's immediate, and that's the classroom. Students can plug their

classroom. Students can plug their homework into Chat GBT and it spits out an answer within seconds. It can write their essays for them, give them personalized cliffnotes, and even answer complex math and science questions. And

there's no way for teachers to tell.

It's like the rug has been pulled out from the entire system. Yeah. And even

if these students don't want to use these systems to cheat, they often feel like they have to, or else they're going to fall behind their peers who are. When

your grade feels like it's the only thing that matters, then all of the incentives push kids towards using and abusing these tools. And of course, teachers are struggling to figure out how to grade assignments. And even if the whole concept of assignments and

grading has to change, the old way of running education seems suddenly and pretty fundamentally broken. So in a way, AI is forcing us to rethink what education is for and what the education

system does. And that's critical because

system does. And that's critical because education is obviously the foundation of our society. If we do it right, it will

our society. If we do it right, it will set up our society to thrive. But if we do it wrong, the consequences can be disastrous. So we're at an inflection

disastrous. So we're at an inflection point where we can actually reexamine some fundamental questions about what is the purpose of education? what is it actually for? So to begin to answer that

actually for? So to begin to answer that question, we've invited two guests on the show who've thought deeply about the structure and purpose of teaching for a very long time. Maryann Wolf is a cognitive neuroscientist and expert on

the development of the learning brain.

She's the author of P and the Squid and Reader Come Home, which explore how reading, writing, and thinking affect brain development. And Rebecca Winthre

brain development. And Rebecca Winthre is the director of the Center for Universal Education at the Brookings Institution, where she's on the global task force for AI and education. Her

book, The Disengaged Team, came out in January of this year. So, Mary and Rebecca, welcome to Your Undivided Attention. And you have our very

Attention. And you have our very undivided attention.

We're going to talk about undivided attention. Great. So, we wanted to have

attention. Great. So, we wanted to have you both on the show today because you're both lifelong educators who've thought a lot about the role of technology in education and you've written books on the damage that social media has done and that smartphones have

done to kids attention spans and and critical thinking skills. Um, but you're also optimistic about the kinds of positive role that technology could play in the classroom. So, I was reading your work in preparation for this interview

and both of you in different ways talked about a kind of um fouian bargain that technology is in the classroom where you give up something of deep moral importance in exchange for something of material importance and I was really

struck by that and so I wondering like what did you mean by that and how have we seen this appear in education in the past? Maybe Rebecca, let's start with

past? Maybe Rebecca, let's start with you. A Foustian bargain. Okay, Daniel,

you. A Foustian bargain. Okay, Daniel,

you don't you don't start with the lightweight questions. Um, never. So, I

lightweight questions. Um, never. So, I

I think the thing I am most haunted by as I watch AI, generative AI develop um,

and think about how good it is getting.

It's starting to reason. It can possibly deceive. And they learn how to do

deceive. And they learn how to do algorithmic thinking which is basically logical thinking. And that is a very

logical thinking. And that is a very foundational skill that helps them be better and give better answers in other parts of life. And that is exactly what

we want kids to do. That's the reason we want kids to write essays. It is really the process of learning to think in a

logical clear way. a thesis statement with evidence underneath and and organizing it and making an argument

that transfers to life and being able to think across domains and subjects. And

so I just keep thinking of little rooms around the world with incredibly smart developers training AI to think that way and then chat GPT which students use a

ton. it might be the majority user of

ton. it might be the majority user of chat GBT or students actually using it to deskkill themselves by having the essay write for them and so like what

are we doing here um so that's the thing that keeps me up at night Maryanne so let me begin with what Rebecca said in

the beginning and that's about thinking how do we develop thinking so I look at it not as a Foustian bargain But in a Socratic search for the first principal

and the Socratic search for the first principal always for me what child learns best under what conditions. So in

my book reader come home the reading brain in the digital world. I suggested,

not unlike you, Rebecca, that there are several sets of skills that we want our children to learn, but that for the reading brain to develop, I think the

research points to it being developed most, if you will, not efficiently, but best in terms of deep reading as the

goal on print. And so I suggested that between zero and 10 12 all along we're working on what not is just decoding but

the ability to develop these what I call deep reading skills that are analogical inferential empathic and most

importantly the sum of which is critical thinking that that happens best in print when we have this beautiful reading brain and I'll begin with

reading but we It could same thing with math but I'm just going to use it because that's my easy point. When we

build this circuit the circuit has a very basic function. The whole point of learning is to use the effort the and

this is what I really want to say that is so worrisome to me about AI. It's the

efforts to build the circuits elaboration to make analogical skills to make inferential skills. One of the most beautiful parts of learning is to pass

over from the perspective of that egocentric learner into the perspective of others. This is a a effective

of others. This is a a effective cognitive process that's going on. But

that all adds to the ability of that learner to become critically analytic about what they're reading. Now the

problem with AI for me is what we call cognitive offloading that the in the interest of efficiency we can do all

this faster and better if we're using these technological devices that augment and blah blah blah. The reality is that

that what we need as learners are the efforts. Even Emerson said it's when we

efforts. Even Emerson said it's when we are braced by labors. That's where

thinking begins. And all of that is in the interest of the imaginative insights of the individual. All of that should never be

shortcircuited in the interest back to fast of efficiency or the best grade. If

we could somehow model the importance of effort and labor that's what builds the circuit that I want, you know, of the deep reader. But we we have to dig in

deep reader. But we we have to dig in here because in some way this feels a lot like um teachers talking about the advent of calculators in mathematics.

like kids won't remember their their multiplication tables. And to a certain

multiplication tables. And to a certain extent, I have to say like I'm all for that. I'm all I'm all for not having

that. I'm all I'm all for not having spent three years of my life, you know, two hours every day of the week trying to memorize multiplication tables. That

just seems like a win to me. So, how do we tell the difference between things that are we're fine being dependent on our technology versus the things you really don't want to be? I would say we

have over history as a species evolved through cognitive offloading. Like not

none of us I I think could be dropped in the middle of the woods and know which berries are poisonous and which berries are not which is something we would have

known many many years ago. Um so there is to me there is things that are things we might live with but the you know the

things that I worry about are sort of the core and this is where the you know you think about AI being introduced education building on what Marian said the core of what it means to be a

learner. What Marian is talking about in

learner. What Marian is talking about in terms of deep reading is a skill that is not just about reading. It's about

thinking. It's about understanding yourself versus the rest of the world.

It's about coming up with new ideas.

Those are things that we should not cognitively offload. And I'm curious,

cognitively offload. And I'm curious, like we're going to spend a lot of time in this episode on artificial intelligence, but I'm sort of curious starting in the past. So over the last 20 years, you've seen this happen in a

lot of different classrooms um in different nations, in different municipalities trying to integrate technology. And there's been this

technology. And there's been this similar kind of fouian bargain, right?

This this this gain of efficiency, but this real loss of social fabric or ability to teach kids what you're trying to. Can you talk a little bit about what

to. Can you talk a little bit about what just happened and then we'll move towards what's about to happen? Sure. I

mean, over the last couple decades, we've seen several waves of technology come into schools. So, one of the big

ones was hardware. Let's get devices in schools, not in kids hand. like desktop

computers, computer labs, remember the days of having computer labs. Um, and at the end of the day, and the OECD has done research on

this across many countries, about 70 countries, the the education systems that really push getting computer labs in, desktop computers, technology in the

classroom, their kids didn't learn more than the systems that didn't push it.

And it's not that technology can't help with learning. It's because you can't

with learning. It's because you can't just take learning, which is a profoundly social process, and take teachers who are people who are doing

many things to help kids learn. It's

that you have to integrate it into the teaching and learning process. So, we've

seen this a little bit time and time again. Um, certainly can talk about cell

again. Um, certainly can talk about cell phones. I don't know if you want to go

phones. I don't know if you want to go there. that that's a that's a whole

there. that that's a that's a whole different um topic and issue, but it permeates schools. But I think the thing

permeates schools. But I think the thing that we've learned is you have to be very intentional about how you introduce technology into education. It will not

magically make things better. So, so

what I what I think you I heard from that is that underneath this big skill that we call reading are a bunch of cognitive development. You have to learn

cognitive development. You have to learn what an analogy means. You have to learn a bunch of cognition. And I hear you saying that that is just way better done on print than than on devices. Have we

seen that as schools have rotated towards devices? Have we seen those

towards devices? Have we seen those skills drop? I often feel that I'm in

skills drop? I often feel that I'm in the wild west frontier of knowledge in this area. Um I have people in Norway

this area. Um I have people in Norway asking me to testify about what is happening with the digitizing of their

textbooks and their libraries. and

they're worried that like Sweden, they will do this massive digitization only to find the grades slipping and that

academic performance is declining. So

Norway now and Sweden have decided not to. Meanwhile, Korea is beginning to

to. Meanwhile, Korea is beginning to digitize their third grade textbooks.

And I'm I'm literally on their NPRs at PBS saying, "Wait, you know, let us get more information, but for heaven's sakes, do

not do mass digitization." So, I think there's enough, if you will, data about this. Um I think all of you know the

this. Um I think all of you know the Singapore study that was released last year in which you have data from 0ero to 8 from Singapore, McGill and Harvard

showing that the more digital exposure between zero and 8, the less the attentional mechanisms are working in the same way and the difference in

academic performance. So we have these

academic performance. So we have these different databases at different developmental epochs. some at infancy,

developmental epochs. some at infancy, some 0ero to eight, some young adults.

And I can talk about that later if you wish. But yes, I think the data to this

wish. But yes, I think the data to this moment in time suggests that reading is best with print, not to be not

complimented, but to be learned during these very pivotal developmental times.

Daniel, I want to come in because one of the things that Maryanne didn't um say explicitly, but I certainly have taken from her work because I think it's actually very relevant beyond reading is

that when you're reading on digital devices, you're more often skimming. Like

Maryanne, you've found like I you're you found that there's like an F pattern.

You read the top and you go down a little bit in the second line or a Z pattern. Like we're really not we're

pattern. Like we're really not we're what we're doing what is called surface learning. Yeah. Which you're basically

learning. Yeah. Which you're basically going quickly and sort of getting the top headline and you're missing a lot of nuance and actually stuff that's interesting. You really are trying to

interesting. You really are trying to operate as like a little bit like a machine. Input in, input out. Let me get

machine. Input in, input out. Let me get it on the test. Let me get it on the on the worksheet. Let me try to get the

the worksheet. Let me try to get the right answer. And that's actually not

right answer. And that's actually not exciting for kids. And so one of the things that we found in in the book research we did with my colleague Johnny Anderson um was that kids are super

disengaged from school. They are not motivated. They are not engaged. They

motivated. They are not engaged. They

are not enjoying learning. And we found these different modes that kids show up in. One of them being passenger mode

in. One of them being passenger mode which is basically doing surface learning and coasting. And that is out not just reading it's across the board.

So I think there's something deeper. I

can't be more thankful to you for making one of my major points which is that reading is so much more than decoding the surface and that first circuit

that's what it does and what we are actually teaching our children is to elaborate that circuit and that is engaging because it engages not only uh

what a word looks like but what it's connected to what are the thoughts that that evoke that elicit. So we're really teaching how to think. Han Bongul, the

Korean philosopher says that we are so accelerated that we're moving from thing to thing, stimulus to stimulus. And

that's what reading is like a canary of the mind for that because it shows you when you skim, you just don't have time to allocate attention. We have to have

time for beauty. We have to have time for your own connections to this to this content. But you know, well, I love I

content. But you know, well, I love I love how live this conversation is and this is why I'm so thrilled to have you both in the same room because Maryanne, you bring this like deep neuroscience of learning and Rebecca, you have this big

macro how are different countries thinking about this and what are the systems doing? So, it's just a thrill to

systems doing? So, it's just a thrill to have you both in the same conversation together. So, one of the things I I I

together. So, one of the things I I I hear you talking about is we've replaced not just the way that kids learn, but we've replaced some of the fundamental motivations for learning. And I'm

curious if we can dig into that a little bit, like how has the the motivation for students of what they're getting out of every moment of learning, how has that changed and how do we want them to be

motivated? I can definitely talk to

motivated? I can definitely talk to that. Um, I'll kick it off, Marian, and

that. Um, I'll kick it off, Marian, and pass it to you. Great. So Jenny and I, my co-author and I just did three years of research. Um, and we were looking at

of research. Um, and we were looking at the question of why kids don't really like school. And it is basically it's

like school. And it is basically it's not actually that they don't like school. We they like their friends. They

school. We they like their friends. They

like going to school to see their friends. We saw that in COVID. They

friends. We saw that in COVID. They

don't like what they do in school.

That's what they don't like. Um, and we found that kids show up to their learning, they're sort of motivated and engaged in kind of four ways. They're in

passenger mode, which I just talked about. They're coasting, doing the bare

about. They're coasting, doing the bare minimum. This is roughly the experience

minimum. This is roughly the experience of half of the middle school and high school kids in the US. Or they're in achiever mode where they're really excited to do perfect on every

assignment and get a gold star in everything that's put in front of them.

We have thought for a very long time in education and in society that achiever mode was the top of the engagement mountain. Getting the right answer was

mountain. Getting the right answer was the top of the engagement mountain. We

have learned from our research. It is

not. And actually kids who get stuck in achiever mode are very fragile. They're

risk averse. They're less able to adapt.

And they do not have the resilient skills that if they have a bad day and get a bad grade, they can just pick themselves up. They're really focused on

themselves up. They're really focused on the outcome, not the process. And then

you've got kids, kids in resistor mode.

These are quote unquote the problem children. They're avoiding and

children. They're avoiding and disrupting their learning. But they

actually have a fair bit of gumption and agency because they are saying often inappropriately class clown, not turning in their homework, skipping school.

These are all kids way of telling adults, hey, this is not working for me.

And they actually those kids if you shift the learning environment can actually flip to explore mode which is the top of the engagement mountain. And

we know from two decades of research explore mode is where they get to explore their curiosity and they are driven and they really do become unstoppable. And when they get that

unstoppable. And when they get that opportunity, they actually do better academically and they are being prepared to swim in the AI world that they are

entering because they will be able to navigate all the shifts and changes that come their way. But less than 4% of kids we found in middle school and high

school get a chance to regularly be in explore mode in school. And so, wow, that to me is what we need to hold in our mind and figure out how technology

and AI can help kids get into explorer mode, not reduce them to passenger mode.

I was going to say, does technology put us into explorer mode uh or achievement mode or when the other modes? Totally

depends on how technology is used. There

are times when it can be really good.

So, for example, Arizona State University, sort of the very um creative and inventive um higher ed institution, a couple years ago, piloted a new

approach with virtual reality in their biology 101 class. So, this is an introductory to biology class. Everybody

had to take it. Not a lot of people loved it. Kids didn't do really well

loved it. Kids didn't do really well except when they started introducing for 10 minutes. Each class, it's the normal

10 minutes. Each class, it's the normal class. 10 minutes sort of do a lecture

class. 10 minutes sort of do a lecture on a concept photosynthesis or you know endangered species I don't know make it up and they go into this virtual reality

beautifully created world that they have to explore and go find the example of what they're learning in the textbook in

this world that ability to actually explore in sort of a semi-mbbodied state even though it's not it's in virtual reality you know kids did so much better

on biology 101 that they're that's a methodology that they're using. The

great promise of education technology was always spending the ass and the teens in technology was always that you could let kids follow their interest.

You could let them explore the topics they were interested in that instead of reading one textbook it would react to you and if the more you show interest in something the more you could pull from it. But that's that doesn't feel like

it. But that's that doesn't feel like what I've seen or certainly I haven't seen results from that in education. Why

didn't we live up to that great promise of giving kids tools that would let them explore more? I think that

explore more? I think that we gave kids tools that were focused on adaptive and personalized learning which

is not exploration. That is I want you to learn fractions in third grade and I will give you an adapted sequence where you know if you get the wrong answer we'll give you a couple more questions

till you master it and then you can move on to the next. very effective in mastering third grade fractions. Um, not

hugely exploratory. And then the other thing I think is we made the mistake on the other end of just letting into our classrooms a wash of

overwhelming technology from cell phones to the internet to, you know, Chromebooks. I can't tell you how many

Chromebooks. I can't tell you how many times my um seventh grader came home last year and I was like, "How was school? What did you do? What did you do

school? What did you do? What did you do in math?" and he was like, "I played

in math?" and he was like, "I played Minecraft unblocked. I played Zelda the

Minecraft unblocked. I played Zelda the W unblocked. I You can get any video

W unblocked. I You can get any video game unblocked on any Chromebook." Like

kids will find a way around it. So, I

think we kind of never really nailed that piece. Well, so one question is why

that piece. Well, so one question is why are we just throwing technology at children as if it's going to help? I

mean, I'll just say when I was at Google and I knew product managers who were building, you know, the Google classroom sort of suite, they were not experts in child developmental psychology. These

were just people who were trying to build products and get marketing, you know, get the thing adopted by as many schools as possible. Not that they didn't care about kids. They did. It's

just that they weren't actually fundamentally developed. That was not

fundamentally developed. That was not their core education as they were making design choices that would influence, you know, the developmental uh brain that Maryanne uh you speak about so so eloquently. Um so we just should we talk

eloquently. Um so we just should we talk for a moment just about the incentives of you know schools don't want to look like they're behind and not adopting the latest technology. As the other schools

latest technology. As the other schools get the Chromebooks, we should adopt them too. Oh, the other schools are

them too. Oh, the other schools are getting the iPads. We should get them to, oh well, kids are going to grow up in this phone world, so we got to make sure that kids are using phones in the classroom. But all of this is just

classroom. But all of this is just misguided and and you know, very naive thinking. Um, can we talk about some of

thinking. Um, can we talk about some of these sort of social pressures and what's driving this mass naive and sort of almost counterproductive adoption of technology. I think there's multiple

technology. I think there's multiple incentives and you center for humane technology talks about incentives a lot.

So you guys know what they are broadly in the tech space and I think you do a very good job of of uncovering them. So

a lot of tech companies are trying to sell into schools and the incentive is to create products that will um be

easily adopted at scale and make money.

We analyze um edtech based on does it substitute for an analog function? Does

it augment what you know we're doing in real life? Does it modify or redefine?

real life? Does it modify or redefine?

And so most of what tech does is substitute or augment because that is the incentive is to sell into schools in

a way they can plug and play very easily and you can scale and get money. I did a large study several years ago for a book called Leaprogging Inequality and we looked at 3,000 education innovations

across 160 countries and 1500 of them were edtech and 80% of those innovations were just substitution and augmentation.

That would mean if you're doing a paper multiplication worksheet and you um digitize it on a tablet, it could help.

It could augment because it could automatically grade and save the teacher time. But you are not

time. But you are not profoundly changing what education is like. That

like. That said, your point Tristan I about the fact that a lot of edtech is developed not by educationalists is is the

perennial discussion in every single education conference I go to which is why can't we get educators at the table because we do know that when educators

are at the table better products are made. I think about clever which is in a

made. I think about clever which is in a lot of classrooms which if you guys don't know or your listeners don't know

is a very simple single sign on portal for teachers and parents and students. Other areas that I think do

students. Other areas that I think do really well in edtech are uh incredible work on supporting neurode divergent kids with technology. There's incredible

dyslexia um software my son uses from text to speech, speech to text and it's you know developed by educators. So when

it is developed by educators and solves a problem in education it can be quite effective. Well so I want to build on

effective. Well so I want to build on that though. Go ahead. which is it's

that though. Go ahead. which is it's easy to talk about the supply side and say you know teachers aren't aren't enough at the table or the people to Tristan's point the people who know how to build tech aren't the people who know

about human development but there's another side of it too which is you know since the '9s we've gone from a period of real information scarcity to this flood of information and the other side

of it that I've seen is educators teachers parents even say you know we need to educate kids for the world that we have now and that means means making sure that we educate them for the flood

of information. So maybe the question I

of information. So maybe the question I have is what does it mean? Forget about

what exists now and how broken it is.

What does it mean to actually educate the next generation for this absolute title wave of information, this confusing, often contradictory, often overwhelming information we get through

the internet?

Well, I think we have for good reasons, good intentions believed that if we had test scores, we could see how well education was doing. And when it isn't

doing well, there's this, if you will, almost reflex, okay, let's do something more. And the something more inevitably

more. And the something more inevitably in the last decade has been tech technological fixes.

The reality, and I'm actually going back to what Rebecca said about our children being disengaged. Um, the reality is that we

disengaged. Um, the reality is that we put so much on the backs of teachers who have to use all kinds of flexibilities

to move from one thing to the next and they're expected to whatever it is that year, they're expected do it to do it.

So that third grade, fourth grade, and eighth grade scores show how well they're doing when in fact we're all doing poorly. The NAPE scores of the

doing poorly. The NAPE scores of the country which were released just you know last month show abysmal results that if my goal is deep reading for the

world only onethird of our eighth graders in the United States are even close to that and of that onethird only

one half of the kids of color are in that third even worse 40% of our eighth graders are not at a basic level of

reading. Now, technology is not going to

reading. Now, technology is not going to fix that. I mean, I think the top note

fix that. I mean, I think the top note you're talking about is kind of an exasperation. All this technology has

exasperation. All this technology has moved so quickly and educators and parents and teachers are all struggling to integrate it. And Marian, to your point, it integrating it has meant this very linear approach to education. And

so, that's what just happened to us and we're still recovering from that. But

now AI comes into the picture. Well, how

will AI change education? Because how do you not take away the essay when you have an essay writing machine? What does

that mean? Because you know, you don't want to take away the essay, but the essay is now broken. I think Daniel, what you're bringing up is the purpose of education. And the purpose of

of education. And the purpose of education in schools is profoundly Yeah. shaken to its core. Yeah. Because

Yeah. shaken to its core. Yeah. Because

we have in what Jenny, my co-author and I talk about as we're moving in from an age of achievement where the purpose of

school has been primarily to rank and sort kids to what I would call an age of agency and lean into a lot of the other

purposes of school that have evolved over the years. There are many purposes that are crucially important. one

custodial care. What would we do? We

knew we actually we found this experiment in CO schools are the number one ways in which governments provide child care. Number two, um

child care. Number two, um socialization, which it doesn't have to just happen in school, but that is a big way. And if you live in a democracy,

way. And if you live in a democracy, it's about citizenship development, everybody having having a similar school experience. because if we don't have a

experience. because if we don't have a shared understanding and experience, we will devolve and lose our democratic um way of life. Let me add a fourth and

that is and this I I don't want to sound um too philosophical and yet I think there's a philosophical aspiration that our schools and our learners can have. I

think what we have right now is this almost TS Elliott bifurcation or triurcation of information, knowledge

and wisdom. And when we are only after

and wisdom. And when we are only after information which AI is so good at and its translation into knowledge which we

hope it will complement us we nevertheless must never forget what does that all mean for humanity the future of the species and that's the wisdom part

and so what I hope that the school can give is this sense of translation that we are taking information we are transmitting it to you so that you will

have knowledge from which you will help propel us wisely. So I completely agree and I just want to acknowledge like that the stakes of this are are really high, right? We're talking about democracy.

right? We're talking about democracy.

We're talking about wisdom. We're

talking about not losing deep human skills and and yet I feel like less and less sure that I know what wisdom means in an AI empowered world, right? Right.

And so we are really sitting at this precipice of a really deep change to what it means to grow up inside of a world imbued with with artificial intelligence.

And to my ear, some of these solutions sound great, but I'm wondering they're very much focused on sort of the continuation of an old tradition. So for

example, you talked about, you know, school used to be a place where you memorize things because if it wasn't with you in your brain, it was really hard to look it up. And then schools

became I I forgot how you said this but it's where you find it not what it is.

It's we increasingly had to live with the internet in knowing where I can look for things or knowing where the knowledge sits. And with AI, I think

knowledge sits. And with AI, I think there's a new skill that's coming online, which is almost how to manage.

Like it used to be management was something that you learned late in your career because really for the first 5 to 10 years of your job, you were just an individual contributor. And now I'm

individual contributor. And now I'm actually seeing management skills in kids the age of 10 because I they need to manage their AI and doing certain tasks. And I'm curious as as people look

tasks. And I'm curious as as people look at education as a as a set of metacognitive tasks, as a set of learning how to do certain things, what tasks that we haven't been teaching kids

that we suddenly need to teach them in a world of AI. I I think there's an opportunity, Daniel, I've spent my life looking at education innovations and how

to transform education systems. And so in some way I'm very excited about AI because it will it has to move us from

the age of achievement into the sort of age of agency as I call it. Uh where you could have schools break open that

sorting and ranking and really bring much closer together knowledge acquisition with knowledge application.

So that is how many schools are trying this. There's great models around which

this. There's great models around which is okay we what are we going to do for this quarter? We're going to try to

this quarter? We're going to try to solve the problem of trash in the streets. This is an example from uh a

streets. This is an example from uh a conference I was just at with um the former under secretary of Rio de Janeiro who did this in the favllas radical agency in their schools and they

actually learned a lot better on the content because when you're trying to solve a problem that's meaningful and relevant you have to it included everything included math included

geography and social sciences and they had to survey and they had to do interviews and they had to look at the history of trash and They remembered that so much more, right, than a skills-based curriculum, right, rather

than input input out because we humans make meaning. We learn things when we

make meaning. We learn things when we have a make meaning of them and they're relevant to our lives. So, can AI help us do that more? Maybe if it if it's

used in that way and in that sort of explore mode, that would be a great example of explore mode. Yes. But I I also think just because AI is

here doesn't mean we have to force it on our kids. Like we should feel free to

our kids. Like we should feel free to say no, we don't want to use this for our young kids at this moment in time.

So, one of the big promises of AI is obviously tutors that we are going to have individualized tutors. And that's

what we're being sold as a story that AI is going to enable that for the masses.

We're about to enter this age of abundance, the best education we've ever had. And obviously, there's conflicting

had. And obviously, there's conflicting views on this. But recently, the World Bank ran a program in Nigeria where they used an AI tutor to help students learn English. And the early results were

English. And the early results were extraordinary that just after 6 weeks, the students achieved the equivalent of something like 2 years of instruction.

and students who had those tutors performed performed much better in their end of year exams and uh the longer that they worked with those tutors, the better they did. And I'm just curious to sort of dissect this example because it

sits on the optimistic side. Yeah, I do think that um we've seen before this AI tutor with more predictive AI, personalized learning.

So part of it has to do with the context. I have worked across the globe

context. I have worked across the globe and in many uh countries there are 150 kids in one class in a first grade,

second grade, third grade class. Um and

teachers teach their heart out but there is no way they are able to reach every last kid. So the Nigeria example six

last kid. So the Nigeria example six weeks after school it was twice a week with an AI tutor. It was about a.3

standard deviation improvement which is is quite good in education. We also saw a.3 standard deviation improvement

a.3 standard deviation improvement during COVID in Batswana 12 weeks not through AI. It was through teachers

through AI. It was through teachers sending text messages on flip phones to parents. Parents opening it up, kids

parents. Parents opening it up, kids doing the it's math problems and then they teachers would call and say, "Can you put the kid on speaker phone? Let's

talk through this math problem." If the kid got it, they would send a harder problem the next week. And kids also improved.3 standard deviations. So what

improved.3 standard deviations. So what

you're seeing is that there's very little instruction going on. So it is not a replacement for

on. So it is not a replacement for teachers in education all these contexts and there's many more examples that that get that these are wonderful apps.

They're in 45 languages now and we're all for teaching those early precursors of literacy which is what they were doing in Nigeria too. But it's what

happens next. And what happens next can

happens next. And what happens next can have all kinds of differences. I work in Johannesburg sometimes where the one school Bella Vista teaches the schools

in the settlement with 100 kids in a classroom. There is no question that

classroom. There is no question that those technical aids are essential. So I

I I there is never a binary here. It's

what works best in what context for which children. We need to watch out,

which children. We need to watch out, right, that we don't we're not pulling up the ladder of education behind us.

Like these tools can supercharge adult learners and people who have those cognitive skills. And the worry is that

cognitive skills. And the worry is that actually for early learners, you're not just not helping, you're actively hurting them having the abilities that you want later. Basically, if you give Chad GPT to uh someone who does not yet

have the critical thinking skills, they're doing more cognitive offloading versus if you have someone who, let's say, goes all the way through high school, has a full developmental paradigm, then they use CHP, then they're getting the uplift and they're getting more enrichment. And it's it's

not totally lasting. There's still some offloading, but it's like there's an enriching there's a non- diminishment process. And I feel like when we talk

process. And I feel like when we talk about what an ideal world looks like, I feel like landing that distinction is very important. We really have to look

very important. We really have to look at adult users versus children users very differently. Is AI helping adults

very differently. Is AI helping adults in the education system do their work better and more efficiently? Is it

making bus schedules more efficient?

Yes, it's amazing. Is it doing um calendaring which is always a pain in the butt for schools more efficiently?

Yes, it is. It um there's an incredible um examples from around the world of walled garden GPTs being given to teachers to who are just experimenting

coming up with really interesting things that make their lives better like being able to assess kids who are learning English for the first time much more quickly and saving a lot of time. Um,

and that is because those adults in the system have critical thinking skills and can you and and have their hand on the steering wheel. They have agency over

steering wheel. They have agency over the AI. Rebecca, you're running a premortem

AI. Rebecca, you're running a premortem on AI and education. Tell us about that because I imagine it has to do with these skipped skills or these these places where we're just going to do it wrong. What does it look like to run a

wrong. What does it look like to run a premortem on AI? So, a pre-mortem, there's a science behind a premortem. It

is the opposite of a post-mortem where you move um the debrief, the autopsy forward. We should have done this when

forward. We should have done this when social media rolled out a decade ago.

And we learned our lesson. And so, our task force is collectively with many many people across the globe um asking

two questions. one, what are the um

two questions. one, what are the um possible risks to to AI and children's learning and education? And get those all out on paper and really imagine, use

your big imagination. And then question two, what

imagination. And then question two, what can we do today to mitigate those risks and harness the really exciting possibilities of AI to help kids learn

and grow? And so that's what we're

and grow? And so that's what we're doing. And do you have any intermediate

doing. And do you have any intermediate findings or we just stay tuned? Oh, some

of the things we are seeing is that people are feeling like AI is inevitable and that they can't say no. We don't

want to use it in our classroom at this point or that point. Um, which I think is worrisome because it isn't. We are a

agentic people. We can decide what we

agentic people. We can decide what we want to do with technology. So that is one thing that has come through loud and clear that concerns me. I think our

parents need to know that like last year JAMAMA comes out and shows the more digital exposure the less language development is happening. So and what is

happening between zero and five is this massive distraction. I who was it Linda

massive distraction. I who was it Linda Hunter or someone called this 20 years ago she called it continuous partial attention of our children. Well, that's

Linda Stone. It's a continuous partial attendance. Yeah, it was 1998. I mean,

attendance. Yeah, it was 1998. I mean,

it was a long time ago, but it was right. And the reality is that our kids

right. And the reality is that our kids are constantly being bombarded by the iPads, which I love on a plane and nowhere else. But this is this is these

nowhere else. But this is this is these are these are kids who are learning to be overstimulated. Um, Han Bungch talks

be overstimulated. Um, Han Bungch talks about it. This is the the deadly

about it. This is the the deadly hyperactivity of our young. Well,

whether it's deadly or not, it's certainly not it's not enhancing their ability to have focused attention, to have a better memory for things that

will be consolidated and used later in all these other processes. So 0 to five is also part of this. We've got to really think about what we can do with

our parents even before school. Just to

pick up on Marian's point about interactivity and socialization of young people and what technology does. Um, you

know, one of the things that we people are quite worried about is young kids are being socialized to interact in society with other people. And when

they're interacting with an AI, you can interrupt it, you can be rude to it, you can call it names. And that is a form of socialization. And kids have a hard time

socialization. And kids have a hard time understanding do I do that with a chatbot but not my brother or not my friend. And we we saw the damage that

friend. And we we saw the damage that social media did to kids interaction.

And I think we risk if we don't do it right really scaling that more broadly.

But okay, I will ask you a question.

Where are both of you headed with this topic? What do you what will you do next

topic? What do you what will you do next with it with just this conversation?

Well, I'll tell you one thing, which is we here at CHT are very worried about the last thing that Rebecca brought up, which is it it changing the very nature of what it means to relate to each

other. Yes. That not just for children,

other. Yes. That not just for children, for adults as well as AI begins to insert itself into our relationships, into our institutions.

Uh how do we as human beings deal with that? And how do you design the

that? And how do you design the technology such that it's not inadvertently creating huge harms to to commons that we haven't even named yet?

If in the 201s it was about the attentional commons, we destroyed all of our attention. And as side effects, we

our attention. And as side effects, we began to not only polarize but destabilize, you know, democracies and and so much more. Well, what are the side effects of this AI wave? And can we

learn what they are and can we educate people on what they are before it's 10 years later and we're just learning what we did to ourselves. So that's what keeps Tristan and I up at night. And so

having people on like you who can speak to this and trying to trying to make this conversation progress at the speed of change and everybody has kids no matter which political party they are

part of and they see the impacts of technology on their kids. Yeah. 100%.

With children we we really recognize that we have a duty of care. We need to protect our children. We need to design for our children. And that's why we often focus on it at CHT. I love this

sign. the duty of care. You know, the

sign. the duty of care. You know, the the because the pope has been so ill of late, I wanted to quote him at some point about children and he said that

children are our world's best diagnostic for the health not only of our society but of our whole world and I think that's you know that duty of care is

part of that. Well, that's a good place of any to end it. I'm so thrilled to have both of you on as deep experts here. is an amazing conversation and

here. is an amazing conversation and thank you for coming on your undivided attention. Thank you for having us. This

attention. Thank you for having us. This

was fun. It was

was fun. It was