A.I. School Is in Session: Two Takes on the Future of Education

This episode is supported by KPMG.

AI agents, buzzworthy, right?

But what do they really mean for you?

KPMG's agent framework demystifies agents, creating clarity on how they can accelerate business-critical outcomes.

From strategy to execution, KPMG helps you harness AI agents with secure architecture and a smart plan for your workforce's future.

Dive into their insights on how to scale agent value in your enterprise.

Curious?

Head to www.kpmg.us/slash agents to learn more.

Well, Casey, today we've got a very special back to school episode where we're going to talk about AI and education.

Oh, I thought we were going to start school over from scratch like Billy Madison.

No, that would be fun, though.

But this is roughly the year 3 BCE in the world of education.

By that, I mean before ChatGPT existed.

How do you like that?

Wait, we're now three years before ChatGPT existed?

God damn it.

Somebody needs to do his math homework.

God.

That one sounded great in my head.

That was a real banger.

I'm Kevin Roos, a tech columnist at the New York Times.

I'm Casey Newton from Platformer.

And this is Hard4.

This week, it's our back to school episode.

Big thoughts on how AI is changing classrooms from K-12 to college, featuring interviews with Alpha School co-founder Mackenzie Price and Princeton professor Dee Graham Burnett.

Plus, students weigh in on how they see education changing in real time.

Have you ever thought of getting an education, Kevin?

No, never.

So, Casey, lots of kids, including mine, are going going back to school around this time.

And we thought it'd be a good time to check in on the state of AI in education.

This is roughly the third school year since chatbots like ChatGPT became widely available.

And there's been a lot of drama and consternation, and school districts trying to figure out how to or whether to incorporate this technology into their classroom or whether to ban it.

And a lot of fascinating discussions have been had about that.

Yeah.

And so today we want to give you a range of of perspectives on that subject.

We want to talk a little bit about what's going on in K-12 schools.

We want to talk about what's going on in college.

And then we're going to hear from students, the people who are truly on the front lines of learning.

They're going to tell us what they're up to and how they're trying to reckon with what ChatGPT has wrought.

So first up, we're going to hear about a new approach to K through 12 schools.

Our first guest is Mackenzie Price.

She's the co-founder of Alpha Schools, a private school network that has gotten a lot of attention recently for the way that they are reimagining the school day.

In an alpha school, you don't have a traditional teacher or a traditional lesson.

You don't sit in a classroom with a bunch of other kids learning the same thing at the same time as everyone else.

Instead, alpha school students spend just two hours a day doing personalized lessons and coursework online using an AI-powered education app.

And then the rest of the day, they devote to group activities that the school calls life skills.

There are still adults in these classrooms and in these schools.

They call them guides rather than teachers, but their job is not to transmit.

information to students.

Instead, they are sort of helping students stay motivated, supporting them, and helping them with their other activities throughout the day.

I first heard about alpha school from a parent I know who was very interested.

And I just thought, well, that sounds really weird and different and interesting.

And I actually want to know if that works.

Yeah, it is a very different approach and quite controversial.

But I think that Mackenzie has some frankly fascinating ideas about what education could be in a world where AI exists.

Yep.

So let's bring in Alpha School co-founder, Mackenzie Price.

Mackenzie Price, welcome to Hard Fork.

Thanks for having me.

I'm so excited to be here.

So Alpha School has been getting a lot of attention recently.

It's not strictly new, but I think it's fair to say it's having a bit of a burst of energy and excitement around it and a lot of questions about the premise that you all are offering.

This notion of doing two hours of lessons a day using AI software for curriculum and guides instead of teachers.

So walk us through an average day for a student at Alpha School.

They get dropped off in the morning.

What time are they getting there?

And what's the first thing they're doing?

Yeah, so we are a full-time school.

This is not like weird, dystopian, robot-led, kids are locked up at home type of school.

So kids show up at about 8:30 in the morning, and we always start each day with what we call a limitless launch.

Think like Tony Robbins for kids.

It's a chance for kids to come together.

They usually do some sort of like impossible task.

For example, this morning, we had a bunch of our students who were paired up in groups of four to six kids and they had these long strings and we just tied them up in their hands, and they had to figure out how to unravel the strings and work as a team and communicate to make the string unraveled.

So it's kind of like a human noodle project.

After they've done that, then our students sit down.

It's time to do the core academic block.

So you're exactly right.

Our kids are crushing their academics in only two hours a day.

We practice what's called a Pomodoro technique.

So kids are basically doing like 25 minutes of focused attention in the core academics of math, reading, language, and science.

They get breaks in between.

And then by lunchtime, academics are done for the day and it's time to do other things.

So in the afternoon is where it gets really exciting because when kids don't have to sit at a classroom desk all day long just grinding through academics, we instead use that time for project-based collaborative life skills projects.

And these are workshops that are led by our teachers.

We call them guides.

And they're learning skills like entrepreneurship and financial literacy and leadership and teamwork and communication and socialization skills.

And at about 3.15 in the afternoon, the kids all come back together and we do what's called shout outs.

And then they head home after school.

And the cool thing about what AI is allowing in personalized learning is kids can crush their academics in the day.

So they're not having to do homework at night when they get home.

They have the rest of the evening to enjoy time with their family and friends and go play sports and do all the things they love.

So that's kind of a typical day at Alpha and it's pretty fun.

It does sound fun to me.

I'm not sure it would sound fun to everyone.

And I want to dig into a lot of the different elements that you just described.

But I wonder if first you could give us a sense of where all of this came from.

What was the set of problems that led you to reimagine school in the way that you have?

Yeah.

So fundamentally, I was a mom who was frustrated by the experience that I was seeing my girls have in traditional school.

And when I say traditional school, what I mean by that is the one size fits all, everybody kind of learning at the same pace, teacher lecturing in front of the classroom model.

I saw when I sent my girls off to kindergarten, first, second grade.

They were so curious and interested, which by the way, every kid is like that.

And over time, like that light and that fire was just sort of going out.

And finally halfway through my daughter's second grade year, she looked at me and she said, I don't want to go to school tomorrow.

I said, what do you mean?

You love school.

And she just said, school's so boring.

And so I looked around.

I did realize it wasn't about moving to a different public school or going to private school versus public.

It was truly that.

model of education.

And so we started the first school in 2014.

And then, you know, even back then, we knew we could use app-based learning to provide the ability for kids to kind of go at their own pace and level.

But in the last three years, with the advent of generative AI, we can incorporate those into the precise and accurate measurement and personalization of a student's learning plan.

And as a result, you know, our kids are crushing their academics and they're doing it in a fraction of the time.

Now, tell us a little bit more about the AI technology that you're talking about here.

My understanding from reading a little bit about the curriculum that you all are using is that it's not really based on generative AI.

It's sort of more like taking an online quiz.

You have students who are at their computers who are sort of getting guided through these lessons and having some sort of spaced repetition, maybe to help them understand or memorize some difficult concepts.

And then they're sort of being evaluated on that.

Is there more to it than that?

And how do you see generative AI playing in?

Yeah.

You know, when people think about AI in education, they immediately think of, you know, either some robot teacher at the front of the classroom, which is totally not accurate, or they think of chat bots.

And we actually aren't using either of those.

Instead, what we do is we use generative AI to create personalized lesson plans based on K through 8 Common Core and at the high school level advanced placement curriculum to basically provide the lessons that each and every kid needs based on where they're at.

And so one of the things that's kind of cool about what artificial intelligence is allowing us to do is we can assess where a student's at, what do they know, what don't they know, and then generate a lesson plan that goes and fills those holes.

The other part that we use is basically the vision model that allows us to be able to measure engagement and effectiveness of how a student is learning.

So what's their rate of accuracy?

How quickly are they moving through the problems?

Are they reading the explanations?

Are they watching them?

Are they guessing?

How are they doing that?

And then a student will get coaching on how they can be learning more effectively.

But basically, we're using that aspect of artificial intelligence to create the personalized learning plans and then measure engagement.

And then the other thing that's really exciting about this, and we're just at the very beginning of what's going to get exciting here, is we can take not only the knowledge graph that a student has, but we can overlay it with their interest graph.

So for example, did you guys like to read growing up?

Were you readers?

I was a big reader.

Yeah, me too.

You were.

good.

Well, you know, you can take that kid who maybe doesn't love to read, and suddenly you can have them be the main hero along with their best buddies from their soccer team.

And they're in an Avengers-style save the world, you know, type of story that's at exactly the right Lexile level for that kid.

And when you can do something like that, it makes reading come alive.

So that's one example.

Think about, you know, learning math and being able to combine that with interest in fashion design or sports statistics.

And that's just something that we are, we are right at the very beginning of being able to incorporate in our academic learning.

And it makes kids super excited.

I'm sure that people ask you this all the time, but I imagine our listeners might be sitting there thinking, how do you compress a whole school day into two hours?

So can you speak to that?

Well, first of all, go back in time and remember when you would walk in your math class.

First of all, you're spending, you know, 10 minutes just sitting down, getting people to be quiet, pull your papers out, turn your book to this page, whatever it is.

Then you would have a teacher that would go through a lesson, right?

And remember, at the beginning of that lesson, there were kids in that classroom who already know what's going on and they're just sitting there like, okay, this is boring.

And at the end of the teacher's explanation of whatever concept she or he was teaching, there are kids who are like, I have.

absolutely no idea what's going on, right?

And either way at the end of class, you know, the teacher has no choice but to move forward with that material.

And so you think about how inefficient a traditional classroom is.

And so if you're spending 50 minutes on math class, the amount of real information that's really being given out is much, much less than that.

And so basically, when we can have our kids spend 25 minutes on math and they're learning at exactly the level and pace that they need to know, that's where we see so much more efficiency.

And I think that's the thing I would love to help parents understand.

Your kid does not have to sit in class all day long in order to get not just their academics done, but be excellent in academics.

Now, I can see how this kind of self-guided

AI-based curriculum would work really well for like a student who's very self-guided, who doesn't need a lot of structure.

But I'm curious, how does this work for kids who may not be super self-directed?

Yeah.

So personalized learning allows us to help every kid wherever they're at.

So you take that kid who maybe struggles a little bit more or maybe has fallen behind.

First of all,

when a child is able to learn the material that they're at, let's say you got a fifth grader who needs to go back and review second or third or fourth grade knowledge, it's easier for them.

But the key critical feature to our success is that we have transformed the role of the teacher in our classroom to be able to focus on the most important thing when it comes to great learning, which is motivation.

How do you get that kid who's maybe not motivated to be able to lock in?

And that is the role of the adults in our classroom.

So we have a little boy who loves bird watching.

And every week when he would hit his academic goals, he and his guide would have time to go back in our green belt and bird watch.

And there's all kinds of things that our teachers basically are able to spend their time really getting to know kids and understand what are the things that motivate them when they're challenged what are the things that they're saying to themselves and how can we help them incorporate more growth mindset strategies and connect there and so fundamentally at the end of the day what we're trying to create is self-driven learners who have the life skill of learning how to learn and so you know you could have a kid who is working during our core academics and they might ask one of our guides like hey i'm struggling with this lesson our guide does not go and say hey let me teach you how to carry the one and divide over here instead they're going to say hey did you use your resources you know were you able to read the explanation did you go into the resource library and search for a video uh that is shown to be helpful in learning this concept what is it that you're doing and that's why i think you know i'm really proud of the fact that our classes are 99th percentile across the board you know all grades all subjects one exception fifth grade math we're 93rd percentile but there.

And

unlike

probably most private schools, we don't care where a kid comes to us at.

If a newly admitted student has got holes and is behind grade levels, no problem.

We can help catch them up.

Let's talk about another way that you're motivating your students.

Alpha Bucks.

Your schools have an incentive structure where students can get paid and toys and prizes for completing work or sort of going above and beyond their expectations.

Tell us a bit about how that works and how it ties into something that you just said, which is that you're trying to create kids who are sort of intrinsically motivated to be lifelong learners and maybe less reliant on outside incentives.

Absolutely.

Well, you know, let's first think about the current motivation model that's in traditional school.

If a kid's not doing well, it's generally considered the kid's fault, right?

He's not the sharpest knife in the drawer or he's disengaged and he's not willing to do the things he needs to do to study.

In our model, we believe if a kid is not thriving, it's our fault.

We're doing something wrong.

And what we're missing is that motivation part.

And so when we think about, okay, we've got a kid who's going to come in or any of our students come in, how do we get them excited and motivated?

And one of the things we have found is that cold, hard cash is an incredible motivator.

Just like, you know what?

All three of us on this conversation sort of respond to cash as well in our jobs, right?

And

he loves loves cash.

Yeah.

Well, we start, you know, we start with introducing financial currency.

So we have this thing called alpha bucks and kids earn the alpha currency by getting their academic goals done each day.

And not only are they earning, but then they also learn how to spend because they have the ability to spend in our school emporium.

They learn how to save.

They learn how to invest and they learn how to donate.

And so we start introducing these financial literacy skills that get kids connected.

And, you know, from there, we have all these other motivational models too, to get kids excited about what they're doing.

And, you know, I get a lot of criticism where people will say, oh, you know, extrinsic motivation measures are so bad.

And here's my story.

First of all, 5% of the population is just naturally, intrinsically motivated and wants to learn for the sake of learning, right?

But what do you do with all those other kids who maybe aren't motivated?

And the answer shouldn't be, well, tough on them.

They're not there.

It's like whatever we need to do to turn on that excitement and get kids pumped.

Mackenzie, I'm curious.

There's been some pushback to the alpha school approach.

And I want to just run through a few of these criticisms

and have you respond.

One is.

selection bias.

Alpha school, people say, you know, just gets these strong outcomes for its students because it has the kind of students whose parents can pay roughly $40,000 a year to send them to an innovative private school.

And you've not actually demonstrated that this can work on all kids.

Yeah.

So first of all, I will absolutely admit there is definitely selection bias.

The fact that it's a private school means that the parent has somehow raised their hand to say, you know what, I've decided that I'm not just going to put my kid on whatever school bus passes our street and send them to the local school.

So I totally understand that.

However, there's a couple of things that I like to argue back on when it comes to this.

First of all, we have students who come from various socioeconomic backgrounds.

We've got a school that's as low as $10,000 a year for students to go to.

And actually this year, we're even rolling out schools that are free for students to go to.

So our data selection is going to get larger and more rigorous.

But the other thing that I think is really important to note is unlike most all private schools, we're not looking at kids' transcripts.

We're not looking at assessments that they've done to say, sorry, you're too far behind for us to take you or you're not an A student already.

We really are looking for kids who are basically showing some coachability, like they're excited to engage in this model.

And that's one of the things that I'm most proud of is that we can take a kid who shows up to us in the 10th percentile and we can get him to the 90th percentile within two years by providing that personalized learning experience.

And so one of the things I'm focused on is just continuing to expand the number of students that we're serving and doing it in various price points that can help with that kind of selection bias criticism.

Got it.

Okay, next criticism is basically one that I hear from educators more than students or parents.

It's this kind of fear of fads, right?

I think a lot of educators and people who have been involved in education for a long time, they're used to people showing up, usually from Silicon Valley or somewhere and saying, we've got the technology that's going to fix education forever.

You know, I remember the one laptop per child program was supposed to revolutionize the way that people learned.

And a decade or so ago, there was this whole wave of personalized learning startups that were trying to do a version of what you're doing in alpha school.

And these things just never really stick and they never really take.

And so I think people are wary of people, I know you're not from Silicon Valley, but from a sort of an innovative ed tech experimental setting coming in and saying the way you're doing it is wrong.

We've got the solution here.

So, what do you say to people who have this more sort of basic objection to this thing that sounds very futuristic?

Absolutely.

Ed tech

is never the answer in and of itself.

And that has been proven over and over again

because there are two factors that create a great learning experience.

Number one is

getting the right pace and level of learning.

And that's about 10% of what kind of creates this.

Now, some EdTech products from the past and actually current ones that are out there,

you can hand a kid an iPad and have them do some EdTech app, but they may play around on it.

They may just randomly push buttons.

They could be spinning in their chairs, not even engaging, right?

And so one of the things that is better in this model is that the AI is actually monitoring and understanding if a kid is engaging effectively.

But 90% of what creates a great learner is a motivated student.

And what's so beautiful about this new role of the teacher is that we finally are going to give them the opportunity to focus on what only humans can do well, which is that personal connection and getting to know students to understand what motivates them.

So I think my biggest answer to that is that we have to figure out how to focus on the motivation aspect.

And I will, I will absolutely be the first to admit, this does require a true shift in our mental models, right?

Everybody has a personal experience with being educated and especially educators, right?

They're like, this is how it's done.

In fact, every parent would say, you know what?

The key to learning is a great teacher.

And, you know, 15 years ago, if you'd asked me what the key to learning is, I would have said, well, you have to have a great teacher.

And now what we realize is that that's not totally true.

Now, it doesn't mean a teacher is not in there.

There's never going to be a classroom that I have where there's no teacher, right?

But we're just shifting the focus to work on that motivation.

And I think that's been what is missing in EdTech.

I appreciate what you're saying, that this is not like just a case of we came up with a new technological idea and built a school around it.

At the same time, I think one of the main reasons Kevin and I wanted to talk to you today is we are so fascinated by how AI is transforming schools, the way it can enable things like personalized learning.

So I'm curious as you think about like everything that Alpha School is, how big a piece is the technology for you?

Like how do you think about the tech piece in relation to everything else?

Yeah, I mean, the tech is certainly an important component, especially making sure that kids are using it well and efficiently, which is really what happened about three years ago.

I started my first school in 2014 and we were using apps during that time.

But in 2022, when generative AI started to come out, that's when we realized like, okay, now we have an opportunity to really make sure that kids are efficiently and effectively learning and we can do a better job of building the personalized learning plans to meet kids where they are.

It also completely changed the role of assessments and how great of a feedback tool they are.

And again, contrast standardized tests with a traditional school.

There's a reason that standardized testing has such a bad reputation.

It's because nothing changes for a student based on how they do on a standardized assessment.

They get a grade and no matter what, they move on, right?

And I understand why in a traditional model, a teacher doesn't even have the ability to figure out, okay, how can I go fill holes, you know, here?

What's incredible now is that we can take these assessments, we can plug them into our AI tutor and then generate the lessons that fill those holes.

You know, it's really hard to be good at algebra if your fractions are shaky.

It's hard to be good at fractions if you don't remember your multiplication table.

And so that's kind of the key ed tech feature that we've built in our system.

But then the other part that's kind of exciting about EdTech and what AI is doing is.

the afternoons.

And when we're thinking about fundamentally how we need to prepare young people to be successful out in the world, one of the things they got to do is be literate in all of these incredible AI tools.

So in the afternoons, we're helping kids learn how they can use AI tools to improve their public speaking abilities, to design movies and interactive books,

to do research, to help them become experts.

Like there's so many cool tools that are out there.

And I believe in order to be really successful out there, you've got to have knowledge in your brain that you can draw upon, and you have to be well versed in all of those types of things to be successful.

You know, I think it's something like 60% of the jobs that young people are going to do when they graduate don't even exist yet.

And so we need to make sure kids have those like future ready skills along with life skills.

Yeah.

I mean, it seems like there are sort of two different ways of looking at the opportunity and the challenge that AI presents in education.

One of them is sort of about how we educate, how we teach kids, how kids learn.

And that sounds like what it is that you and Alpha School have leaned into.

The other view is that AI really challenges our notions of what people should be learning.

There's this view that we hear a lot out here in the Bay Area that AI is going to be smarter than most or all humans at most or all things relatively soon, and that that might render a lot of our traditional curriculum obsolete, that students might not need to learn how to do chemistry experiments in a world where AI is just there at their fingertips.

Do you have a view on that?

And I'm curious how you think about the sort of next phase of AI when we do get to these more powerful systems.

Yeah, I think what we're going to see in education is that young people are going to have the ability to become PhD level so much earlier than they would in a traditional system, right?

And I don't believe that learning is any less important in the age of AI, but having a framework for how you incorporate the knowledge that you have and use the tools that are at your disposal to be successful is a skill that we need to teach people.

And one of the things that, you know, AI is really good at, it knows the black and white of knowledge, right?

It knows what's out there, but what it doesn't know is how to push forward into new knowledge.

And I think that's where humans are going to need to live and where the world is going to see incredible progress when we can allow our young people to be successful in that that gray zone,

well, thanks so much for joining us, Mackenzie.

Really appreciate it.

Thank you.

Thank you.

Well, Casey, we are graduating from our segment on K through 12 education to one about colleges and universities.

When we come back, we'll talk to Princeton professor DeGraham Burnett about how institutions of higher education should adapt to the age of AI.

This episode is supported by KPMG.

AI agents.

Buzzworthy, right?

But what do they really mean for you?

KPMG's agent framework demystifies agents, creating clarity on how they can accelerate business-critical outcomes.

From strategy to execution, KPMG helps you harness AI agents with secure architecture and a smart plan for your workforce's future.

Dive into their insights on how to scale agent value in your enterprise.

Curious?

Head to www.kpmg.us/slash agents to learn more.

Can your software engineering team plan, track, and ship faster?

Monday Dev says yes.

With Monday Dev, you get fully customizable workflows, AI-powered context, and integrations that work within your IDE.

No more admin bottlenecks, no add-ons, no BS.

Just a frictionless platform built for developers.

Try it for free at monday.com slash dev.

You'll love it.

Huge savings on Dell AI PCs with Intel Core Ultra processors are here, and they are newly designed to help you do more faster.

They can generate code, edit images, multitask without lag, draft emails, summarize documents, create live translations, and even extend your battery life.

That's the power of Dell AI with Intel inside.

Upgrade today by visiting dell.com slash deals.

Well, Kevin, recently we both read an essay in the New Yorker called Will the Humanities Survive Artificial Intelligence?

And it gave both of us a lot to think about when it comes to education and AI.

Well, I actually didn't read it.

I had AI summarize it for me.

And that really gets to the heart of today's segment, which is what is AI doing to higher education?

We just talked about what it is doing to K-12 education and some of the new models that are being proposed and experimented with for younger kids.

But this is also a conversation that is playing out at colleges and universities, which is why I was so excited to talk with De Graham Burnett.

He is a historian of science and technology at Princeton, and he wrote this really great essay in the New Yorker.

I did actually read it.

I was joking.

Called, Will the Humanities Survive Artificial Intelligence?

And, you know, there's been no shortage of professors and college administrators and presidents sort of opining on how AI and education will intersect.

But what I really appreciated about Graham Burnett's essay was that he argued that this could actually be a good thing for the humanities in particular.

That's his area of focus.

And he has been sort of tracing the decline of the humanities over the years.

And he now thinks that AI could help revitalize them by encouraging colleges and universities to focus on the life of the mind and some of these deeper questions about what it means to be a human.

He also just writes about some really interesting ways that he and his students have been using AI.

He has been encouraging them to have these extended conversations with chatbots, edit that down into these conversations that reveal something both about the student and the language model that they are learning.

And so while no one professor is going to have a perfect prescription for what you should be doing with AI in your classroom, in fact, as he tells us, the humanities are very different from computer science, for example, when it comes to how you might use AI.

I do think he just has some really interesting conversation starters and thoughts for us as we enter a new school year.

Yeah.

So let's talk with Graham Burnett.

Graham Burnett, welcome to Hard Fork.

Oh, it's so good to be here.

Where are we catching you today?

Are you at your school?

Are you at Princeton?

I am.

I am in my office here in History of Science History in Dickinson Hall in Princeton, New Jersey, Princeton University.

You have many books behind you for our viewing audience.

Wait, it's called the History of Science History Hall.

No, history of science is like a program within the history department.

Oh, I see.

And then the history of science, technology, that kind of stuff.

Sorry, Casey didn't go to college, so he's a little challenged here.

Graham, I wonder if you could start by just describing the picture of AI and higher education as you see it.

You wrote this great piece in the New Yorker about your sort of realization that AI was going to be a big deal, maybe a bigger deal than a lot of your colleagues in academia gave it credit for.

So what convinced you that AI was going to be unlike other technologies that universities and schools have faced?

Well, these technologies are changing very fast.

So it's always a little dangerous to predict.

But there can be no doubt that these technologies emulate human thinking, analyzing, and expressive powers and do so with sufficient sophistication as to make their output indistinguishable from human work.

And that means that a bunch of the ways that we've operated to educate students are off the table.

You write about this really fascinating exercise that you did with your students, where you gave them the assignment to have a conversation with a chatbot about the history of attention, edit the text down to four pages and turn it in.

And you write, reading the results on your living room couch turned out to be the most profound experience of your teaching career.

Tell us about what you thought you were going to get out of students doing that exercise and then why it wound up being so profound for you.

Yeah.

So it's true.

I wanted to experiment with developing some new kinds of assignments in my classes that feel equal to this moment.

And the class in question is a class on the history of human attentional capabilities.

And so the students had read their way through a bunch of really dense stuff about the history of monasticism all the way through to the attention economy.

And

because

the attention economy is driven by intelligent algorithms that figure out how to maximize time on device.

I thought it would be interesting for the students to go into a conversation with AI, a chatbot, about a topic that in some sense, AI is pretty good at, analyzing human attention, optimizing feeds for human engagement, attentional engagement.

And I didn't really know what

would come out of the thing.

I want to emphasize that I didn't especially have an expectation.

So

the power of what emerged was in line with my high hopes rather than exactly defying my expectations.

But I will say this.

I've spent 30 years

reading student work.

And

part of the power of sitting and reading this stack of papers was

the uncanny feeling that I was watching a generation

feel out the emergence of a kind of alien, familiar, ghost, god,

monster, child.

And that was intense because the kids, my students are primed and the machine.

is special right now.

It's, it's kind of feeling its wild oats, if you know what I mean.

And so to watch them kind of like test each other out in the ring, what have you got?

Oh, you know, and they know they're going to be working with each other for some time to come.

I would say that was the

overarching feeling of the uncanny that settled on me with those papers.

You also recount the experience of a student who finds herself feeling like when she is in conversation with the chat bot, she is able to get a better understanding of her own

beliefs, of her own knowledge, because it is so patient and it is so private.

And she doesn't have to interrupt everyone.

She doesn't have to come to your office hours and wait in the line of students.

And I was so struck by that idea because we've heard it now in a number of contexts.

We recently had someone telling us about the experience of chefs who have had the exact same experience and now they feel like they can explore so many other cuisines because they don't have to be embarrassed and call, you know, some chef at another restaurant and ask them, you know, how to make the pad thai or whatever it might be.

So I'm wondering what you can tell us about that, that sort of use case.

Are more students finding that there actually is this intellectual avenue that they can pursue with ChatGPT?

And it is not about writing the essay for them.

It is sort of about like being in communion with the subject matter.

Yeah, I think that this moment with this particular student was very affecting because it wasn't just that the machine was so patient and was willing to pay attention to her like practically no human ever had.

It went a step beyond that.

What she actually said was,

I got to be inside my intelligence in a way that felt kind of new to me because I wasn't worried

about the person I was talking to.

Very smart young woman, very sensitive young woman, conscious of the people around her.

And I don't want to speak for her, but I think it's fair to say she had, like a number of my, you know, very brilliant, often female students, spend a certain amount of time trying to make sure that she didn't like, blow away her interlocutor with her own intellect.

So kind of just being a little like keeping it under a wrap, going slow, making sure that whoever she was talking to wasn't feeling intimidated.

You can sort of imagine that dynamic.

I can't imagine.

This is not a problem I have.

Kevin.

I'm blowing away people with my intellect.

Actually, what it is, Kevin, is that certain people had mothers who loved them so much that they never actually think that anyone could want anything but to hear more from them.

And I'm one of those guys.

But this girl was not like that.

And so she was like, wait, with the machine, I don't have to take care of this machine.

I can bring full force.

And it was like, whoa, she was rolling out her powers.

And I thought that was an extraordinary kind of insight into a thing these machines can provide in that context.

So if this is where students are at the beginning of this school year, sort of, you know, engaging with these chatbots,

having these sort of profound or engaging experiences with these AI systems.

I want to ask about where you think teachers are, faculty members at Princeton and other elite universities, college administrators.

What is the posture of universities toward AI heading into the school year?

And how has it changed from maybe a year or two ago?

It's all moving so fast, Kevin, and so difficult to

get a vector, but still pretty defensive, I would say.

Still concerned, you know, with the police function.

of the professoriate.

You know, we're like the sheriffs.

And so

the concern is all my assignments are now useless.

I can't assign papers.

Am I going to have to like do a blue book exam?

Are the students going to do the reading or are they just going to feed it to Notebook LM and turn it into a podcast?

You know, we're a relatively conservative cohort, maybe not politically, but in terms of the like.

lag time on change within the humanities in particular.

So there's a concern that I'm going to have to like change my class, which is a pain because I'm going to have to do things differently.

So that would be my vibe check, speaking very generally.

There are, of course, exceptions, but

I would say that's still basically where things are.

And what's your vibe check on like where you've arrived for your own approach to education?

You and your piece you write about what I think sound like some really cool assignments.

But how are you thinking about this coming year?

Are you still kind of developing new approaches?

Do you feel like you've got a toolkit?

Where are you?

I'm on sabbatical this year, so I'm working on congratulations

i'm hunkered down and uh trying to get a book finished before i have a newborn baby on my hands makes sense but i will speak to that because here's where you go to the core what do you want from a university what do you want from education and

i don't really kind of like care especially about their job training functionality There are plenty of other people who can worry about that.

What I care about in education, what got me into this business and keeps me in it, my like mission critical love of what I do in this kind of space and in my classrooms,

working

to give shape to persons equal to the conditions of freedom.

That's what matters to me.

I believe that there really is something about taking shape as a human being who is responsible to the inheritance of your freedom in relation to others and in relation to yourself.

And I believe that colleges and universities are a space where that work must continue to happen.

So if you care about that, I actually think that this moment is very exciting

because

all of our kind of police functions are kind of going down the tubes.

And a lot of what we've previously asked students to do, which is to perform a kind of karaoke dance of being a professor, producing new knowledge about Shakespeare, just like I do.

So write a paper that looks like a mini and slightly sucky version of an academic article that I would write after studying Shakespeare for 26 years.

That's pointless.

In fact, even the article that you, my fellow professor colleague, were going to write about Shakespeare, that's kind of increasingly useless because the systems can produce a shit ton of interesting articles about Hamlet iteratively.

And so if it becomes like prompt engineering, the question is not, can we write 500 new papers about Hamlet?

It's does anyone want to read them?

And so asking the students to produce several hundred new papers on Shakespeare, meh, it's, you can't actually make them do it anymore.

So the good news from my perspective is while I think those activities are interesting and important, I never really thought of them as at the heart of what you would want a college or a university for.

So I'm kind of happy to kind of let them like lightly slip away.

Now, that's a pretty positive vision.

I want to be clear about something.

There's another like much less happy possibility, which is that everybody's like, you know what, this humanities stuff, we never really understood it anyway.

The whole point of colleges and universities is to prepare people for the job market and to maximize the ROI of their tuition.

And then the humanities come to end.

in colleges and universities.

I think that is frankly the more likely scenario.

So let me be bleak for a moment.

I think that the most important forms of that work are going to need to happen outside of colleges and universities.

It's another way of saying that colleges and universities, it's not really their fault, but the shifting economics of education has actually meant that just about everybody, the administrators, the parents, the kids, are all thinking so much in terms of return on their tuition investment and interpolating themselves into an increasingly exigent job market.

Right.

I mean, I think that's the key tension here is there are a lot of ways that you could spend four years discovering yourself and your highest purpose that don't involve paying hundreds of thousands of dollars for a college education and potentially going into debt that you're going to be working off for decades to come.

What is the value proposition

to sort of borrow the language of companies that colleges and universities can offer to students and parents who are staring down the barrel of this enormous tuition bill to say, this is going to be worth it for you?

Yeah, it's a beautiful question.

It's not really my question in a way.

You know, I basically believe at this point that the good work,

the work I care about, is going to need to happen in thousands of new schools that are going to arise outside of the traditional framework of extremely expensive university education, and that that is actually a very exciting moment in the history of education.

And I think that AI is going to

help break, in a sense, the university model that has anyway reached a certain kind of end game.

So we're now talking very big, but this kind of thing has happened before.

Remember, I'm a historian, like, you know, the emergence of schools that taught in secular settings in the vernacular was itself a very specific historical development.

So, you know, it was the end of a certain kind of monastic training in the Renaissance.

This is happening again.

Like the university model and that value proposition that you're calling for, other people have to answer for that.

I'm interested in creating new kinds of schools in which people can continue to explore how to live and what to do, what historical consciousness means, and think about tradition and where we're headed.

And that doesn't, you know, the humanities exist as a set of fields precisely because they don't monetize.

One thing that I'm just realizing through this conversation and our last segment is how important the shape of education is to what actually ends up being transmitted and learned.

You're going to have a very different experience learning something one-on-one from

a computer program than you are from being taught by a teacher standing at the front of the class.

And that experience might be very different from learning in a more like collaborative and active way with a group of your peers.

So I guess I'm curious, Graham, what you think the ideal shape of a humanities education in the kind of post-AI age is?

Is it students who are sitting around a classroom discussing things that they've previously discussed with their AI tools and done sort of self-guided learning on?

Is it professors still giving lectures and students supplementing that with what they're learning from AI?

Like, what is the shape that is most conducive to the type of humanities education that you want students to get?

From my perspective, the biggest change that we need to implement urgently and that we are implementing, but it's not happening as fast as it needs to is a shift in literacy.

It will pain me to say this and it may pain some of your readers to hear this, but long-form immersive literacy is coming to an end as a widespread cultural phenomenon.

People do not read now like they did 10 years ago.

They don't read like they did two years ago.

They definitely don't read like they did 30 years ago.

Now this gets complicated.

Who people, where, how much literacy was there ever, bracket all that.

Even in the most elite circles, the ability to give yourself to an immersive textual experience is ending.

This has enormous implications

for a tradition that has meaningfully inscribed itself textually.

And part of what you came to university in the humanities to do is to read books, but you can't anymore because nobody can read them.

So now what we can do is give people short sections from books and ask them to do stuff with those short sections of books, dramatize them, discuss them, perform them,

enact them, sing them, memorize them.

cut them up into little pieces and stick them on the walls of their friends' dorm rooms.

Because the stuff that's in those books, we aren't going to lose.

We're going to carry elements of the tradition forward.

And people are not going to get stupider.

People are getting smarter.

But their ability to immerse themselves in a tradition that presents itself like this is finished.

I mean, that's why y'all do a podcast.

Basically, this is orality.

Like once upon a time, you'd have been journalists writing long pieces.

You each kind of have been that for a minute.

But in fact, now this is what journalism increasingly is.

So we are moving into a culture of orality, but we have a humanistic humanistic structure within modern colleges and universities that is still rooted in textuality, but that's finished.

Do you think that we've sort of reached peak college

10, 15 years from now,

because

the humanities may be switching to this sort of deeper form of inquiry rather than getting people a credential to go out and get a job?

Do you think fewer people need to go to college, should go to college, will go to college?

My mother was a college president,

Lovely place called Kun College.

My dad was a dean.

I grew up like in on campus.

Like my earliest memories was getting fruit loops out of the dispenser in a college cafeteria at Indiana University in Bloomington.

So like you're looking at a dude who's like spent his entire life on a campus.

I love these spaces.

I think it's a very important time for people who care about decency and goodness to rally to the defense of colleges and universities under trying conditions.

Comma.

However,

yes,

the thing that has made colleges and universities in America so special over the last 75 or 80 years

is winding out.

Some of the old motherships, Princeton, Harvard, Yale, Stanford, they're not going to change that much.

But a lot of the other institutions were seeing it already.

They got to get.

dynamic or die.

But we are going to create lots of new schools where that work of soul craft can happen.

It's going to be part of people's working lives.

It's going to be financed totally differently.

It may not be like accredited at all or provide any kind of certificate or diploma, but people want still to do that work and they need that.

I have to say, it makes me a little sad.

I feel like I was able to have my cake and eat it too in college because I did a journalism and an English major.

And the journalism major was the ROI because I was going to need a job when I graduated.

And the English major was an excuse to read novels because I wanted to read a lot of novels.

And so I was able to have the kind of combined experience.

And I'm wondering if there's a way to keep those two things together, or if that truly is just an artifact of the time that I grew up.

I do think basically, like everything else, that was a historical moment.

Yeah.

If you could incept one sentence, one idea about AI into the minds of humanities professors

and college and university administrators around the world.

What would it be?

I'll try one.

I mean, it's a very difficult question, obviously.

Let's try this.

This is not a test.

This is real.

Gather up what you've got.

what you think the college and the university experience are for, and get ready to bring it because this is going to be the crucible for the kind of work you say you care about.

Yeah.

Well, Graham, I think that's a really good place to end it.

Thank you so much for chatting with us, and uh, people should go read your piece.

It's in the New Yorker, it's called Will the Humanities Survive Artificial Intelligence.

Thanks for the opportunity to talk with you all.

It's really, really interesting.

Thanks, Graham.

It was super interesting.

Thanks, guys.

Take care and be well.

When we come back, what students have to say about how AI is changing education?

This episode is supported by KPMG, AI agents, buzzworthy, right?

But what do they really mean for you?

KPMG's agent framework demystifies agents, creating clarity on how they can accelerate business-critical outcomes.

From strategy to execution, KPMG helps you harness AI agents with secure architecture and a smart plan for your workforce's future.

Dive into their insights on how to scale agent value in your enterprise.

Curious?

Head to www.kpmg.us slash agents to learn more.

Can your software engineering team plan, track, and ship faster?

Monday Dev says yes.

With Monday Dev, you get fully customizable workflows, AI-powered context, and integrations that work within your IDE.

No more admin bottlenecks, no add-ons, no BS.

Just a frictionless platform built for developers.

Try it for free at monday.com slash dev.

You'll love it.

Dell AI PCs are newly designed to help you do more faster.

That's the power of Dell AI powered by Intel Core Ultra Processors.

Upgrade today by visiting dell.com/slash deals.

Well, Casey, we wanted to end today by hearing from a number of students about what they are feeling and experiencing at the beginning of this school year.

And we put out a couple calls for responses in previous episodes and got just a ton of great responses back in our inbox.

Yeah, and thanks to all of the students who sent us in their thoughts despite being massively hungover from the party at the frat last night.

Yes, thank you for taking a break from vaping and collecting laboo boos

to write in.

So we should say this is not meant to be a scientific sampling.

It's just what a number of young people who listen to hard fork around the country and around the world are seeing and experiencing.

So we're going to play some clips and talk about them.

All right.

So this first clip comes from Keith, a student at Princeton who is using AI to connect dots and fill in the the gaps from professors' explanations.

And let's hear what he has to say.

Hey, Kevin, hey, Casey, huge fan of the podcast.

I'm a computer science student.

So sometimes we have like math classes or math proof classes.

And sometimes professors will often skip over the derivation of a proof or they'll kind of like hand wave away like the intuition behind certain equations.

And they'll be like, hey, like, here's the equation, like, go it.

And you're kind of left trying to reverse-engineer whatever the professor wrote on the board or the notes that you took.

You're looking back at it like three weeks later, trying to prepare for an exam.

And you're like, How did they come up with this?

I found AI to be really helpful for kind of connecting the dots between different moments of clarity that you might have gotten when you were sitting in lecture or were like focused and then unfocused one time.

So, you kind of use AI to fill in the gaps between when you kind of lost focus or lost understanding.

Yeah.

Thanks.

I mean, here's where I get so jealous of the students today.

How many times was I sitting in a lecture class and the professor said something that I didn't totally understand?

I wasn't going to raise my hand and stop the whole lecture.

I just wanted a little bit of extra information about something.

Now you can just ask a chat bot right there on your laptop in the middle of class and kind of catch up.

So that feels like a superpower to me.

Yeah.

And I also think it points to the very real limitations that many human professors and and teachers have.

And like, I love and adore teachers and professors.

And some of the people who have changed my life the most have been teachers.

But there's also like a large distribution of skill among teachers, right?

There are people who are amazing researchers, but not very good at actual teaching.

Or I'll give another example, Kevin.

There are people who are hired to be football coaches who all of a sudden were teaching my high school chemistry class.

And let's just say that was not their forte.

Right.

So without denigrating the, I think, the amazing teachers out there, I just think like this is a good use of chatbots.

I think often when we talk about AI in the classroom, we're sort of comparing the tools that exist to some sort of platonic ideal of the world's best, most patient, most helpful classroom instructor.

And in reality, like all professors and teachers have gaps in their knowledge and abilities, and it's useful for AI to come in and help fill in the gaps.

All right.

This next one comes to us from Greta.

This is a little bit of a longer clip.

Greta is studying at MIT, where she's studying AI and business analytics.

And she gave us an overview of her study workflow, which includes some vibe-coded automation.

So I'm interested to hear about this.

Hey, Kevin and Casey.

My name is Greta, and I'm an incoming junior at MIT.

I have three kind of interesting AI use cases.

So first of all, most of our homework assignments are given in what's called problem sets.

And so sometimes when I look at a problem set and I have no idea where to start, I'll upload it into Gemini and then ask it to explain the background concepts of a given question.

So instead of like telling me how to approach the problem, it'll be like, okay, you need to know this part of, for example, graph theory, so that I have more of a background understanding and then I can approach it myself.

And then secondly, I was studying for an optimization test in the fall and I realized that they had given me a full list of all the topics that were going to be covered on the exam.

And I had all the lecture notes.

So I put all of that into perplexity.

And then I told it to go through the topics one by one, explain everything I should know for my exam, and then quiz me on that topic until it was sure that I understood it so that it wouldn't move on before I was ready.

And then finally, one cool recent use case I found is that I take my notes on Notability, which is just an an iPad app, but then I have them backing up to Google Drive.

And so I had Gemini write a Google Apps script that every time my notes back up to Google Drive, it will take all of the notes, upload it into Gemini, and for each note, it'll generate a full summary of everything that was covered in that lecture or recitation.

And then it creates a 10-question quiz so that I can make sure I'm still understanding the material, say like a week later.

And then it puts that into a Google Doc and puts it back in my Google Drive folder.

So that's a little bit more of a complicated use case, but I found it extremely useful.

And honestly, Gemini kind of did it all itself.

It walked me through how to create that integration.

So that was really cool.

Anyways, love the show.

Thank you so much.

What do you think, Kevin?

I think this is amazing.

I mean,

I think this is just an example of how we often just sort of underestimate how good students are going to be at using this stuff on their own.

I think in a lot of schools, they have it exactly backwards.

I think students should be teaching the teachers how to use these tools effectively for things like studying and preparation for exams.

This is obviously a very, you know, sophisticated MIT student who's like familiar with this stuff, but this is the kind of stuff I'm hearing from students at all levels that this is just very helpful for them in creating little mnemonic devices to remember things or sort of a spaced repetition style app that can be used to teach themselves.

So I just think like very self-motivated and capable people like Greta are just going to like zoom ahead in their classes using this tools.

Yeah, I think you nailed on the important thing, which is that for a student who wants to learn, AI is maybe one of the best things that has ever happened to them, right?

She clearly wants to understand the material.

She's not just saying, give me the answer.

She's building systems that ensure she is learning the material, that will not let her move forward until she learns the material.

And that's something that her professor was not going to do for her, right?

So to me, like this is the template.

If you were a college student right now, I would listen to what Greta just told us and try to be thinking, how could I put similar systems into my own life?

Because I think you're going to do great in your classes.

And on the flip side, though, if you don't want to learn, then these tools really cannot help you.

And so I think that as we continue to talk about the role of AI and education, we have to be thinking about that motivation piece because I really think that that's the difference between success and failure.

And that probably is a problem that long predated the arrival of AI.

100%.

All right, this next one comes to us from Claire, a student at Fordham University.

Let's hear this.

Hello.

My name is Claire Krieger.

I'm going to be a senior at Fordham University this year.

I don't personally really like AI.

It just creates too many corner cuts.

Like you you gave the example of book summaries.

And yes, I

we will talk about books in class and I will not have read them and I will look at the ChatGPT summary and I'm not proud of it, but it is like way too easy, I think.

However, a specific example of when

AI was used against like me as a student that I think is pretty telling is I wrote an essay for a class once and my teacher accused me of plagiarizing

because I had a quote that was similar to another student's.

And I'm pretty sure she thought I had used AI to write my paper because of that.

I think AI in education is terrifying because it leads to examples like these where people are wrongly accused of using it.

It happened to one of my friends with one of her professors as well.

She was accused of using AI when she didn't on a paper.

So yeah, it's my two cents.

Thank you.

I'm glad Clara said this because I think this is the flip side of the ones that we've just heard from the students who are using AI for helping themselves learn new things and draw new connections between their knowledge.

There is a lot of corner cutting going on out there using AI.

There is a lot of people cheating their way through school and depriving themselves of education.

And the second thing she brings up is really important too.

There are a lot of falsely accused students out there whose teachers and professors are saying you wrote this using AI.

And I will say this every fall.

I think this is an absolute travesty that these companies are still selling these tools to schools and universities.

They do not work.

Every study of these tools shows that they generate huge numbers of false positives.

You cannot ask an AI system, did you generate this, and have it give you a verified, reliable answer.

And there is no way at scale to determine whether something was written by AI or not.

There just just isn't.

And I worry that that message is not getting across to the educators of America because we still hear stories like this all the time.

Well, is there any good solution to the cutting corners problem that Claire brings up?

I mean, I think it's just about reinventing the class.

My intuition is that you just have to assign work differently and evaluate work differently.

Because if people are able to do things like just use a summary of a book from ChatGPT rather than actually doing the reading themselves, my assumption is they will.

And by the way, that use of AI is not new at all, right?

When we were in college, there were SparkNotes.

Before that, there was Cliff Notes.

Like this has always been possible for students who want to skip the reading.

Have I used SparkNotes in my life?

Yes.

Am I proud of that?

No.

But I think any realistic take on AI in education has to account for the number of students who are just overworked.

They've got too much going on.

Maybe they don't feel like reading the book.

Maybe it's not a very good book.

And so they're just going to cut corners.

Yeah.

All right.

This next one comes to us from Pia, a student studying in Germany.

German is not her native language, but she's finding that AI has been very helpful in translation.

Hi, Casey.

Hi, Kevin.

My name is Pia, and I'm an American living in Germany.

I've been applying to master's programs here, and ChatGPT has been super helpful from the start.

It's been really important to me to do my own work in the sense of my ideas.

But given that I'm not a native speaker, ChatGPT is so helpful to just translate or to sort of smooth those rough edges of my German.

The biggest thing where it was really a godsend was I had to write a 10-page essay, which I did in English with all of my own research and thinking, and you know, no ChatGPT there.

But then I use ChatGPT and the help of my girlfriend to translate it.

On the other hand, I'm very cognizant of the idea that like cheating is cheating yourself.

And so I don't want to shortchange my learning.

And I feel like, especially once I really start, I will have to find those boundaries for myself.

Casey, I love this one because this is an argument that I've had many times with my sort of more

worldly polyglot friends.

I have been saying for years now that it was a waste of time for me to try to learn another language in part because I wanted to justify the fact that I only speak one language,

but also because I just assumed that we would have these like earbuds that just magically translated between every language using AI.

And that actually exists now.

Like you can pretty much travel anywhere in the world and talk into your phone and have the phone translate for you into whatever language of the country you're in.

But my polyglot friends push back on this and they say, you idiot, you absolute rube.

The reason to learn another language is not purely to communicate in that language with people who are native speakers.

It's because you are learning about a culture.

You are flexing your brain in different ways.

You are

depriving yourself by using the easy out of these translation apps.

So where do you come down on this question of AI and language?

Well, I would agree with your polyglot friends.

Of course, diet refers to a person who's in a relationship with more than one glot.

Learning another language, as you say, is not just about communicating.

It is effectively a different system of thinking.

You know, like I took five years of Spanish, so I can get by in Spanish, and it's a beautiful language.

And there are just certain words and phrases that evoke concepts that don't exist in English.

And so you're not going to get that with your AI earbud, bro.

Okay.

Okay.

So as far as, you know, what does Pia do?

Look, if she's living in Germany and going to school there, I imagine some of those rough edges are going to get sanded off of her German and she's going to start writing more fluently in German as time goes on.

There's probably a German word for the feeling that AI is slowly taking over your cognitive capability.

Yeah, we should ask the AI what that is.

All right.

And finally, Vikram Ananta has sent in a video.

And let's take a look at that.

Hey, Kevin and Casey.

I'm Vikram.

I'm a CS student at the University of Michigan.

And it's kind of gotten to a point where AI is just so normalized.

It's like the internet or like a calculator.

You kind of just have to use it.

And obviously, there's some classes where they don't want you to use it and they explain why.

But for like a CS project, like I try not to use it because I want to be able to learn.

But some of my friends use tab to autocomplete in cursor and it's just so good or like generating a presentation from scratch.

And it's really just gotten to a point where if you're not using AI, you're kind of just behind.

Like in non-graded things like clubs or like organizations, they want you to use AI because it's like productive.

So yeah, it's just everywhere.

So AI is ubiquitous now, Kevin.

Yeah, I mean, this feels true to me.

I have spent a bunch of time at schools over the past year or two.

And a lot of people tell me what Vikram did, which is that this is just everywhere.

It's in the water.

It's part of every student's workflow now.

And good luck telling them they can't use it.

I think it also gets to a point about equity in schools.

Another thing that I've come to believe is that schools should be paying for premium subscriptions for students because the premium subscriptions to these products generally are better than the free versions.

And if some of your students have the means to buy the paid versions and some don't, then the students with access to the paid versions are just going to have better answers and better homework and more queries to give to the models before they run out.

And I just think that puts your sort of lower income students at a tactical disadvantage.

So I haven't gotten a lot of

uptake on that suggestion, but I do think that is increasingly an issue.

I mean, what Vikram's comment here makes me think about, though, is just the risk that some of our other students have talked to us about today, which is, yeah, AI is everywhere.

Yeah, you can use it for many or even most of the tasks that you're being assigned.

But at what point are you leaning so hard on the AI that you are not learning the stuff that you went to college to learn?

I've been thinking a lot as we've been listening to all these comments about what is the point of college.

Honestly, I went to school to study English because I loved reading novels and to study journalism because I wanted a job.

And I think I probably would have relied heavily on AI in school as long as I still felt like I was going to get a job at the end of that and be pretty good at that job.

So if Vikram thinks that he's using AI in a way that's going to make him really good at computer science at the end of his time there, then I think he's probably fine.

But if all he's really learned at the end of four years is how to autocomplete, he may have some regrets at the end.

Vikram, I'm sure you're not going to do that, but other people might.

Yeah, put some respect on

Vikram is great.

Vikram is great.

But other people are going to be using that autocomplete a lot.

And I do worry about what the net result of that is.

The thing that this makes me think about is like, if this is really

a transition that is happening at a lot of schools where students are just kind of auto-completing their way through their degrees, then that may not be very durable as a model because pretty soon the AI won't need the human to start them off.

It'll just sort of do all the coding.

And so I think that's what we're worried about is not that today's students will be cheating themselves out of an education, but that we are headed towards a future in which all of this sort of becomes moot because the jobs that people are going to college to train themselves for no longer exist or no longer exist in anywhere near the current number.

So that's what I think about when I think about Vikram's description of college just being sort of full of AI is like, are we just sort of at a halfway point on the continuum where these tools still can't really like use themselves?

They still need the students to sort of guide them along.

And will all of that just sort of be rendered obsolete by just better and more capable systems?

Well, the only way we'll know is if we continue to listen to the hard fork podcast.

So, everyone, campuses all across the country, make sure to stay subscribed.

And we'll tell you when we've crossed the terrifying point of no return.

And on a more serious note, thank you to all the students who wrote in.

Really fun to hear from you all.

And good luck with your fall semester.

This episode is supported by KPMG.

AI agents, buzzworthy, right?

But what do they really mean for you?

KPMG's agent framework demystifies agents, creating clarity on how they can accelerate business-critical outcomes.

From strategy to execution, KPMG helps helps you harness AI agents with secure architecture and a smart plan for your workforce's future.

Dive into their insights on how to scale agent value in your enterprise.

Curious?

Head to www.kpmg.us slash agents to learn more.

Can your software engineering team plan, track, and ship faster?

Monday Dev says yes.

With Monday Dev, you get fully customizable workflows, AI-powered context, and integrations that work within your IDE.

No more admin bottlenecks, no add-ons, no BS.

Just a frictionless platform built for developers.

Try it for free at monday.com/slash dev.

You'll love it.

Dell AI PCs are newly designed to help you do more faster.

That's the power of Dell AI powered by Intel Core Ultra Processors.

Upgrade today by visiting dell.com/slash deals.

Heartfork is produced by Whitney Jones and Rachel Cohn.

We're edited by Jen Poynat.

We're fact-checked by Caitlin Love.

Today's show was engineered by Alyssa Moxley.

Original music by Alicia B'Itube, Marion Lozano, Sophia Landman, Rowan Nemisto, and Dan Powell.

Video production by Sawyer Roquet, Pat Gunther, Jake Nicol, and Chris Schott.

You can watch this whole episode on YouTube at youtube.com/slash heartfork.

Special thanks to Paula Schuman, Puy Wink Tam, Dahlia Haddad, and Jeffrey Miranda.

You can email us at hardfork at nytimes.com with your AI and education take.

Gun injuries are the leading cause of death for children and teens in the United States.

Some people avoid talking about gun violence because they don't think they can make a difference, but every conversation matters.

When it comes to gun violence, we agree on more than we think.

And having productive conversations about gun violence can help protect children and teens.

Learn how to have the conversation at agreetoagree.org, brought to you by the Ad Council.