March 10, 2025 • 46 mins
How can we rethink schools to meet the future? What does this have to do with the invention of the printing press, the prevalence of desk calculators, or the spread of Google? And how is this connected to the writer Goethe, a digital replica of the philosopher Aristotle, or the two lasting bequests that we should give our children? Join Eagleman this week for surprises about what AI means for the next generation.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:05):
What does AI mean for the future of education?
Speaker 2 (00:09):
How should we be rethinking schools to meet the future?
And what does this have to do with the invention
of the printing press in fourteen forty or the calculator
or the spread of Google? And how is this connected
to the eighteenth century German writer Gerta or a digital
replica of the philosopher Aristotle or the two lasting bequeaths
(00:33):
that we should give our children. Welcome to Inner Cosmos
with me David Eagleman. I'm a neuroscientist and an author
at Stanford and in these episodes we dive deeply into
our three pound universe to uncover some of the most
surprising aspects of our lives. Today's episode is all about
(01:04):
the future of education. After all, we're still running our
school systems essentially as we have since the Industrial Revolution,
but now we're in a world of AI. And the
question is are we optimally educating our children for the
future or are we mostly reiterating what we did mispreparing
(01:27):
for a world that will quickly be unrecognizable. After all,
the price of knowledge has dropped precipitously and is rapidly
approaching zero.
Speaker 1 (01:38):
We are so used to having.
Speaker 2 (01:40):
Lots of jobs where you're paid to have specialized knowledge,
like in medicine or law, or engineers or actuaries or
patent examiners or whatever. But with AI that absorbs everything
that we've written and can do an increasingly precise job
of knowing things better than we can.
Speaker 1 (01:59):
In a world where AI.
Speaker 2 (02:02):
Aces the SATs and the MCATs and the LSATs without
breaking a sweat, this is an open question.
Speaker 1 (02:08):
What the future is going to look like.
Speaker 2 (02:11):
I think it would be madness to assume that it's
going to look much like our present in a decade
or two from now. This is what many of us
around Silicon Valley and the world are talking about all
the time. But how does this affect how we are
training the next generation? So for the next two episodes,
we're going to dive deeply into this and we'll see
(02:32):
some very cool directions and surprises. In today's episode, I'm
going to lay the groundwork for how we can think
about brains and education and the two important things that
we need to be doing for the future. And in
next week's episode, I'm going to interview Salkn, the creator
of the con Academy, which is an extraordinary endeavor, which
(02:55):
educates about one hundred million people in the world every year,
and specifically Salini are going to talk about what the
future of education looks like from his point of view
and what tools he's building to meet the moment. So
for today, let's start at the center of the issue,
which is the human brain. The most remarkable thing about
the brain is what we call neuroplasticity, which is the
(03:18):
brain's ability to rewrite itself, to be flexible, to absorb
the experiences of the world and etch them directly into
the circuits of the brain. So brains come into the
world half baked, and they absorb the world around them.
This is how you come to master all the details
of your language and your culture and the knowledge and
(03:42):
the mores of your generation. And human brains are more
plastic than any others in the animal kingdom. They pop
out uniquely unfinished, and this is why a baby giraffe
can walk within forty five minutes. But if you've ever
seen a baby Homo sapiens get born, you know that
it takes.
Speaker 1 (04:01):
Years to figure it out.
Speaker 2 (04:03):
So we have these massively extended infancies because we are
wiring ourselves to the world as opposed to dropping in
more pre programmed. Now, at first blush, it would seem
that having a very long infancy would be a disadvantage,
and in some ways it is, But the end result
(04:24):
is that humans gain capacities that our neighbors in the
animal kingdom just don't have. This is why a giraffe
is doing the same thing giraffes were doing a million
years ago. But humans launch satellites and compose symphonies, and
erect skyscrapers and build quantum computers. No other species builds
a fraction of what we build. And that's because although
(04:48):
most others tend to be tougher than we are, they
don't have the capacity to absorb the world around them
to the degree that we do. In each generation, they
just run the program of being a giraffe. But we
drop into the world and in a handful of years
we have learned almost every important insight and discovery of
(05:09):
the billions of humans before us. We get that, and
then we springboard into the next generation.
Speaker 1 (05:16):
So what does that mean for the way we.
Speaker 2 (05:19):
Think about educating our children. It means our responsibility is
to shape this malleable brain.
Speaker 1 (05:26):
Each child has a forest of tens.
Speaker 2 (05:29):
Of billions of little neurons that are spending every moment
changing the strength of their connections, and plugging and unplugging
and seeking new locations. Whatever they learn in their childhoods,
it's stored in the brain, and it's coded in a
reconfiguration of the network. That's how brains absorb the world.
(05:51):
So our job with education comes down to one main thing,
which is to feed the brains of our children as
rich a diet of high calorie information as we can.
The problem is that schools are largely being run the
same way we've done it for hundreds of years, when
we needed to train children for particular jobs, careers that
(06:15):
we could reasonably assume would be around and lucrative and
solid in twenty years from now. But we're entering a
new world now, a world where previously unimaginable technology, artificial
intelligence that can answer any kind of question this is
available the way that running water is available. Just imagine
(06:36):
if you were Mark Twain or President Garfield, or Ada
Lovelace or Florence Nightingale, the thought of running water would
have seemed insane to you. But in the nineteen thirties
there was suddenly a pivotal decade where most of America
got running water, and suddenly, for the next generation of kids,
it was just like background furniture, such that by the
(06:59):
time you were born, you don't even devote a single
neuron to the fact that you can produce a clean
little river at whatever temperature you want, the second you
want it. That's what artificial intelligence is going to be
for this next generation. Of course, you can open an
app that has read everything ever written by humans, and
(07:19):
of course you can ask it anything at all in
plain English and it will generate an absolutely beautiful answer.
Of course, it can synthesize pieces of data that would
take you lifetimes to read, and it does it instantly
and brilliantly, and soon, of course, you can tell an
AI to act as an agent and go out into
the world and do tasks for you. So here's the
(07:41):
thing I want to emphasize. This world of AI is
going to require different skills than what we were raised
with and it will offer extraordinarily different opportunities. So this
is what we're diving into today. How does current technology
change the future landscape and how can we leverage it
to rewrite our school systems. So first we know that
(08:06):
children are growing up in a different world than the
one that many of us did from their first moments.
Speaker 1 (08:12):
They're in a world of technology.
Speaker 2 (08:14):
News of the pregnancy is posted on TikTok or Instagram,
the news of the birth is spread on WhatsApp and X.
Relatives anywhere on the planet immediately see their pictures. And
they grow up in this world where control all delete
is as basic as ABC. They have no idea why
that key on the keyboard is labeled return. Some of
(08:36):
you may remember typewriters where you had to return the
carriage to the other side of the page.
Speaker 1 (08:41):
All our kids grow up in.
Speaker 2 (08:42):
A world where photographs are instant, whereas when many of
us were children, we had to get the film processed
and it took three days. In the space of a generation,
things change fast, and the world of very young people
now consists of millions of podcasts, billions of web pages,
trily of videos.
Speaker 1 (09:01):
They spend tens.
Speaker 2 (09:03):
Of thousands of hours on mobile phones, video games, texts, emails,
social media. And this is to say nothing about the
continental plate shifts that AI is starting to make, which
will return to in a moment. So as a result,
by the time they're in high school, they have a
different world than many of us grew up with. So
(09:24):
does this change how their brains wire up and how
they view themselves compared to, say, their parents.
Speaker 1 (09:30):
Of course it does.
Speaker 2 (09:32):
Their exposure has been to fast paced, shortened, interactive information,
and what that means is the ways that students learn
and communicate are different. You've probably heard the term digital
immigrants for those teachers who carry old world traditions and assumptions.
The idea is that you couldn't live and teach in
(09:53):
another country unless you learned the language. For many teachers,
the digital world is a second language that was learned later,
and any teacher who was a little behind the technology
ball is now really behind now that AI is here.
So what does all this mean for the brain of
a digital native? Well I mentioned that we come into
(10:15):
the world with brains half baked, and we absorb the
world around us. And the way this works is that
in the brain of a newborn baby, the neurons are
only starting to communicate with one another, but then over
the first two years of life, those neurons begin connecting
extremely rapidly, so by the time you hit two years old,
you have hundreds of trillions of connections and this is
(10:38):
where you max out. You have as many connections as
you're ever going to have, and after that it's all
about pruning.
Speaker 1 (10:46):
Like an overgrown garden.
Speaker 2 (10:49):
Your brain is encoding what it experiences out there. That's
what those connections represent. Whatever you encounter in the world
strengthens some pathways, while other pathways eventually go away. And
that means the process of becoming who you are isn't
about what you gain in the brain, but about what
(11:09):
you lose. Think about it like Michelangelo carving away at
the marble to find the statue inside. That's what your
brain is doing by keeping those connections that resonate with
the world. So how your brain turns out depends upon
what you are exposed to. So given all this is
(11:31):
of course no surprise that technology affects the brain, and
that the current generation carries out tasks in different ways.
With different technologies. They have increased use of some pathways
and decreased use of others, and that gears today's students
towards a different kind of learning. Now, I want to
(11:52):
take a one second tangent, which is that it's very
difficult to do a rigorous scientific study about growing up
digitally and how that affects the brain. Why is it difficult,
It's because you never really have a good control group.
You can't just look at a bunch of students now
and run a comparison against a previous generation, because there
(12:13):
are one hundred other differences between them, including food and
politics and air quality and on and on and on.
And you can't easily find another group of students who
grew up in this moment without being digital, unless they're
amish or deeply impoverished.
Speaker 1 (12:30):
But then there are one.
Speaker 2 (12:31):
Hundred other differences in how they're being raised, and you
don't know what you can attribute to the digital technology
versus something else. So whatever people say are the differences
with this generation, it's difficult to know which of those
has to do with growing up digitally and which has
to do with other issues like politics or pollution, or
(12:53):
changes in sugar intake or less secondhand smoke or legalized
marijuana or whatever, none the less. We can point to
a few things that people have noticed when comparing pre
Internet and post Internet patterns of reading a page of text.
For example, so digital natives tend to make slightly different
eye movements that are essentially f shaped meaning they're more
(13:16):
likely to quickly scan headlines and then the first few
lines of paragraphs before moving farther down the page to
do that. Again, readers from older generations tend to be
more thorough in they're reading and don't skip.
Speaker 1 (13:28):
Around as much.
Speaker 2 (13:30):
Digital natives are also more accustomed to taking in content
that has text and images and videos and interactive sliders
and buttons or whatever, and presumably as a result, it's
tougher for a digital native to concentrate on a long
book that is only walls of text, and so digital
natives seem to be more prone to distractions, as measured
(13:52):
by them moving their eyes off the text entirely. The
cool part is that these details of eye movements can
be tracked, and they're all totally unconscious. In other words,
it's not a signal that can be easily faked because
people don't even know they're doing it. I mentioned this
to illustrate the bigger point that the digital generation is
(14:12):
exposed to different stimuli and that strengthens different pathways in
their brains, and that changes how they learn.
Speaker 1 (14:21):
Fastest and most efficiently.
Speaker 2 (14:23):
So for many of us, when we were growing up,
we got a lot of information taught to us just
in case we ever needed it. So the Battle of
Hastings was in ten sixty six, because you might need
to know that someday, or you might need to do
geometry later in life, or you might need to know
the capital of Indonesia. But education has been transforming because
(14:47):
of the Internet. It's gone from just in case information
to just in time information. Now young people increasingly pull
information on demand. New skills are acquired as needed. Do
you need to understand how to change a bike tire,
you look it up. You need to understand how this
(15:08):
math problem works, or what chemicals to use in this situation,
you look it up.
Speaker 1 (15:27):
Now.
Speaker 2 (15:28):
The reason this new just in time learning is important
is because the learning is not out of context with
the question. When you want to know how something works,
you have the right cocktail of neurotransmitters present, and that
makes all the difference as far as the neuroplasticity goes.
(15:48):
When you are curious, the information sticks and in fact
without any of the neurotransmitter knowledge. The spirit of this
observation reaches back to the ancient Greeks who loose dated
seven levels of learning, with the highest level being when
you're curious about something, then it really sticks. So when
(16:08):
a child asks Alexa or Siri a question and gets
the answer, their curiosity allows the information to have a
better chance of getting edged in than if you simply
stopped the child and said, hey, I need you to
memorize this fact that has nothing to do with your
life right now, and I'm going to ask you about
this later.
Speaker 1 (16:28):
So when you follow some.
Speaker 2 (16:31):
Rabbit hole on Wikipedia, you click on this link because
you're curious, and then the next because you want to
understand that word that you've heard once, and then the
next because there's some fascinating distinction you've never even thought of,
and so on, and eventually you end up twenty clicks
away from where you started. That's an amazing way to
learn because you are maximizing your curiosity at every step.
(16:55):
So each successive generation gets the opportunity for more self direction,
and not surprisingly, our comfort zones come to be different.
Back in the day, most of what we learned was
by text or lecture. It was for the most part sequential,
and one topic was tackled at a time, and most
(17:16):
of our learning was done independently. Now learning has a
very different character. There's way more multitasking, and we learn
from pictures and sound and video, which makes everything a
lot more engaging. A lot of our learning is interactive
with websites or games, and we're often networked with other learners.
Much of our learning now is random access like Wikipedia
(17:40):
or Google or chat GPT. We connect ideas ourselves. Learning
is far more hands on than it ever used to be,
so the way brains seek information has changed, and this
leads me to a one second tangent about the question
of ADHD. Current students are used to constantly changing input
(18:02):
and information sources. They're switching from WhatsApp to instead a
chat GPT. Does that make a student struggle to concentrate
in class? Yes, almost certainly it's a factor. But is
that ADHD in many or most cases? Probably not. This
is a normal response of their brains to the world.
The way they scan and multitask means that sitting down
(18:26):
and being taught from a boring textbook or a droning
lecturer is not sufficiently interactive or engaging for them. It's
more difficult to concentrate on, so the onus becomes on
the educator to meet them halfway. Now, what does that
look like? Well, now, we have the table set for
us to move into the world that we find ourselves
(18:49):
in in twenty twenty five, and specifically the world of
AI as we all know, modern AI is blowing everyone's mind.
Every day we see large language models lms pull things
off that no one programmed them to do or even
really expected them to be able to do, including all
(19:09):
the stuff we're now taking for granted, like summarizing books
or making a video from a text prompt or crushing
the SATs or LSATs or MCATs. And this is why
we find ourselves in an era of discovery more than invention.
And I want to point out that a lot of
the arguments people have been making about AI not being
(19:31):
so good at this or that these have been changing
really rapidly. For example, just a year ago people were
arguing that AI would make silly mistakes about certain things,
and couldn't add thirty two digit numbers with each other,
and sometimes would get the wrong answers in a word problem.
Speaker 1 (19:48):
But in a shockingly.
Speaker 2 (19:49):
Brief amount of time, these shortcomings have all been mastered.
So it's yet to be seen what challenges will remain
and for how long we are in for fit a
ride and it's going to move faster and faster. So
what does this mean for our school systems? Our students
are now in a world where the entirety of humankind's
(20:11):
knowledge has been digested and any piece of information or
any sophisticated analysis is available to you instantly. What does
that mean for the next generation? How will the future
change and what should we be doing about it now? So,
first of all, let's put this in historical context, because
in times like these, it's always useful to remember that
(20:34):
there have always been times like these. So in fourteen forty,
when the printing press was invented in Europe, there were
many thinkers who asserted that this was going to ruin
the next generation. Why because now the answers were just
sitting there.
Speaker 1 (20:49):
They were right there. You don't have to remember much
of anything.
Speaker 2 (20:53):
You just needed to remember where to get the information,
and you could pull it right off the shelf.
Speaker 1 (20:59):
A lot of it.
Speaker 2 (21:00):
Educators in the fifteenth and sixteenth centuries lamented what was
going on here. They were certain this was going to
make everyone intellectually lazy. What they overlooked was the way
that knowledge would continue to explode and the way that
society could springboard off the top of this new technology
of books. Memorizing things was not, in fact the important
(21:23):
part of education. Instead, it was about storing pointers to
where the information could be obtained. Twenty five years ago,
I was having a similar conversation with educators as we
transitioned into a world of the Internet, and a teacher
would ask a question and the kids would just go
home and look up the answer. I don't know if
(21:43):
you remember those days, but everyone's concern was that it
was cheating.
Speaker 1 (21:48):
So what happened.
Speaker 2 (21:49):
It's not that students got dumber, It's that educators got smarter.
We changed our approach to posing the questions to obviate
copy and paste homework. We just stopped asking simple memorizable facts,
and we started doing other things like getting students to
present out loud to the class, or grading each other's
(22:10):
work what new things did they learn, or getting students
to grade the quality of information from different websites, and
generally anything that made the questions an active exploration instead
of a simple fill in the blank. Well, now we
are faced with a new challenge from AI, and the
question is is this the same kind of challenge we've
(22:32):
always had, or is something different this time, and if so,
what and how do we as a society best prepare
ourselves and our students for this future. So let's start
with a prediction about AI that I heard from Steve
Jobs in the nineteen nineties. Jobs was giving a talk
where he imagined that computers could reach this point where
(22:54):
they would be able to understand everything that the great
philosopher Aristotle knew and would be able to simulate his mind,
and you could have a conversation with him. You could
ask Aristotle anything, just like Alexander the Great did. And
I have to say, when Steve Jobs talked about this
in the nineties, it seemed like a full fledged computer
(23:17):
world fantasy.
Speaker 1 (23:18):
On the AI mountain.
Speaker 2 (23:20):
We hadn't even gotten our boots strapped on at the
bottom yet, and yet.
Speaker 1 (23:24):
Here we are.
Speaker 2 (23:25):
Now, We're so far up that mountain that now having
a simulacrum of Aristotle just seems like part of the
background furniture.
Speaker 1 (23:34):
You can do this with any LM.
Speaker 2 (23:36):
You just feed in all the books by someone and
then you have a full, rich conversation. Now, why did
Steve jobs suggestion of having a personalized tutor matter? Well,
it matters because every classroom goes too fast for half
the kids and too slow for the other half. So
a tailored individualized education is a dream that many think
(24:00):
in the education space have had for a long time,
having an expert who could teach you with passion and
patience and in every possible language. That's something that hasn't
existed before, and now it's here. But here's the question
I want to pose. If students were able to sit
and have a conversation with Aristotle, would they or would
(24:23):
they rather be hanging out with other students and making
jokes and getting dates and so on. So I asked
my thirteen year old boy this question of whether he
would use a digital twin tutor, and he found it
not that interesting because he was aware that he would
run out of things to ask pretty quickly, because in fact,
he wants to be with friends and play video games
(24:45):
and run around. Also, there's what I call the problem
of the mismatched internal model.
Speaker 1 (24:51):
A typical student.
Speaker 2 (24:53):
Simply doesn't have in his or her head the same
questions or concepts that Aristotle had in his head, so
there's not much of an interface there. They wouldn't think
to ask him about his notion of four different causes
that underlie everything, or his distinction between substantial properties and
accidental properties of objects, or his theories about virtue or
(25:16):
teleology or the unmoved mover. So if you were a
student lucky enough to sit down with Aristotle, you just
might not know what to ask. You're not that likely
to probe him with really terrific questions. But I think
the dream of a tutor isn't lost. It simply has
to be cast correctly. There's a different way to think
(25:38):
about the concept of tutor, and I think one of
the best examples of this is coming out of the
con Academy. It's called con Migo. Con Migo doesn't tell
you what it knows. Instead, it challenges you to think
critically and to solve problems without ever giving you direct answers.
So you can learn algebra or programming or essay writing.
(25:59):
But the key is that, unlike chatchpt, con migo does
not tell you the answer. It guides you to find
the answer yourself, and it does so with limitless patience.
I won't say more about this now because next week
we'll have Saul Khn of the con Academy on Intercosmos
(26:19):
to talk with us about this, how it came about.
Speaker 1 (26:22):
And where it's going.
Speaker 2 (26:23):
But the thing I want us to note for now
is that this is not a tutor that replaces a teacher,
but one that works with the teacher, and that leads
me to the role that AI is going to be
great at, which is AI as Tha.
Speaker 1 (26:38):
In other words, as a teaching.
Speaker 2 (26:39):
Assistant, AI can do more than just help with the teaching.
It can also help with the job of teaching. There
are all kinds of ways that teachers can use AI
to help with grading, with creating schedules, with tracking attendance
that gives the teachers more time and energy to spend
with the students. AI is super helpful a teacher needs
(27:00):
to write comments to parents on an individual student performance.
If it saves the teachers fifty percent of their time,
that's an incredible win. Or when a teacher has to
make multiple versions of the same test because kids are
going off to debate tournaments, or writing teacher letters of recommendation.
All these are things that AI helps with to reduce
(27:22):
the burden on educators. And the effect of all these
efficiency gains is that it gives the teachers more time
to have real human relationships with the students, and AI
is also going to improve the structure of teaching. A
lot of researchers feel that the big end of semester
(27:42):
exam is going to go away in place of AI
monitoring every problem along the way, so the teacher knows
exactly how a student is doing on the fly. In
other words, with the enormous amount of data available to
collect about learners, I can process and translate that to
provide useful insights to everybody, to teachers, to students, to parents,
(28:06):
and it can give in the moment formative feedback to
the students, so even before they turn things into the teacher,
the AI can differentiate the learning experience based on that
student's need, in other words, tailored education. The main thing
is it can digest all this data on the fly
and make it understandable to everyone. Okay, so what should
(28:28):
we do next in our classrooms. Let's return to what
I said about teachers having to change the way they
gave assignments when Google appeared.
Speaker 1 (28:36):
What do we do with chat GPT?
Speaker 2 (28:39):
Well, first of all, last week I was giving a
talk at a different university and I talked with several
of my colleagues on the faculty there and I asked them,
how are you changing the way that you do assignments
to prevent students from using chat GPT, and many of
them said, well, I changed my assignment up. Instead of
(28:59):
just asking them to tell me the answer, I told
the students to do something new, like compare two ideas
or come up with three different ways of doing something.
And I said, wait a minute, they could still just
do that with their favorite LLM right, And the professors
seemed to have one of two responses. One was, well,
that would violate the honor code if they were doing that,
(29:21):
and the other was I'm pretty sure my students are
not doing that. But the fact is every student is
doing it. There's no reason for a student not to
do it. If you are a student and not using
the available technology while your competition is using it, your
chances of success are slim. So I think we should
(29:57):
assume that all students are leveraging this technology. We need
to work with that reality. But happily, as far as
things to do, it's not that hard for an educator
to modify the assignments. I'll give you an example. This
is what I'm doing in my current class at Stanford.
Speaker 1 (30:14):
We used to have.
Speaker 2 (30:15):
The final assignment as the writing of a twelve page paper.
Now the students run a science experiment on their fellow
classmates and they write up the results. Even if they
get help refining the question that they want to pursue
from an LLM, that's fine. They're still optimizing the question
that they want to pursue and designing the experiment, and
(30:36):
then they're contacting their classmates and running the experiment and
understanding what goes wrong in real life and how they
need to communicate, and finally how to analyze data and
make charts, even.
Speaker 1 (30:47):
If they use GPT for help with that.
Speaker 2 (30:50):
Anyway, it was an easy change for me to go
from a final paper to a final experiment. And we
also have quizzes every class where at the end of
the lecture there are four sessions on a slide and
the students write down the answers on paper. A lot
of teachers are successfully moving in these directions. For example,
many teachers I talk with say they used to give
(31:11):
a big take home exam at the end of the semester.
Now they give a bunch of smaller in class quizzes throughout.
So I just want to emphasize the importance of being
one hundred percent aware of the tools out there so
that educators don't have to pretend that our students are
not using them. And one way that educators can think
about this is actually giving AI assisted assignments where the
(31:34):
students are meant to use AI and couldn't complete it
without it, and then you have oral defenses about what
the student just learned. So just imagine the difference between
having a student making a little diorama or whatever I
did when I was a kid, versus having them build
an AI enhanced science fiction world where the student works
(31:56):
with AI to generate details about the new world, including
the science, the cultures, the creatures, the technology that comes
out of that, and then you have the student give
a performance in front of the class. So it's not
a matter of simply generating it, but refining it and
knowing it and believing it and being able to be
quizzed deeply on the fly about it. You can imagine
(32:19):
these same concepts with AI assisted spoken word performance, where
in combination with AI, a student might generate richer content
to produce on the stage or AI designed next decade
planning from social structures to technology and education.
Speaker 1 (32:37):
And daily life. This is how you give students a great.
Speaker 2 (32:41):
Education because they are learning how to use the tools
of the day to build something creative and understand it deeply. Okay,
so this is a direction where things can go. But
things I think are even more exciting when we look
out a few years. A colleague of mine, Rich Branyak,
runs open Stacks, which is a service that provides textbooks
(33:03):
in eighty five subjects for free. Now I just spoke
with him, and he's put years into this project, but
they're no longer certain that textbooks are the optimal way
to carry information. So open Stacks is trying out many
different things. And by the way, we are in a
long period of discovery here, so some things will work
(33:23):
well and some things won't. One of the things they're
trying is this. Imagine that you need the students to
read chapter seven of this book. You click a button
to pass chapter seven to Google's Notebook LM, or even
better now Google Illuminate, which turns the book into a
podcast with two artificial podcasters having a conversation about the content.
Speaker 1 (33:49):
And you can set.
Speaker 2 (33:49):
All the parameters like which age group this is for,
and the level of difficulty, and so on. I'll play
a brief clip for you from one of Google Illuminates podcasts.
Here it's absorbed Charles Darwin's Origin of Species, and you
can learn about the book in a conversation.
Speaker 1 (34:07):
Let's talk about it. Absolutely.
Speaker 3 (34:10):
This is a classic foundational work in the field of
evolutionary biology. It's amazing to think that it was published
way back in eighteen fifty nine.
Speaker 1 (34:20):
Eighteen fifty nine, that's quite a while ago. Can you
tell us a bit about what makes this work so important?
Speaker 3 (34:24):
You bet, Darwin's theory of evolution by natural selection was
a total game changer.
Speaker 2 (34:30):
Now I'm only playing a bit of this, but it
goes on to mention the key ideas and to break
those down even further, such that in a short podcast
you get a really good summary of the key ideas. Now,
the thing I want to emphasize is that by this
clever trick of turning it into two people talking, it
becomes more interesting than a bulleted list of facts about
(34:55):
the book. Why because it gets closer to a story,
and we have a story shaped hole in our brain,
and that's what allows things to go in rather than
simply bullet points of facts. Now, it's not necessarily easy
to turn everything into a story. But this technology that
turns a book into a podcast gets us part way
(35:17):
there because instead of somebody writing, here's a fact, here's
a fact, here's a fact, now you have someone saying, wait,
I'm confused, and the other person says, ah, hang on,
this is amazing, and the first person says, wow, that
really blows my mind. And now suddenly you're recruiting parts
of the brain that we study in a subfield called
social neuroscience, which is just to say, when other people
(35:39):
are involved, we care more. And that's a big part
of why we care about story. So this quick trick
of converting a chapter into a podcast dialogue with a
click allows story to be approximated and gets you part
way there. And it's instant and free. And that's not all.
If you, as a teacher, want to emphasize some aspect
(36:03):
of the chapter in particular, you just steer the prompt
that way. So open Stax is experimenting with saying, here's
chapter seven of this textbook, but I really want my
kids to concentrate on this aspect of it and not
spend much time on.
Speaker 1 (36:15):
This other aspect.
Speaker 2 (36:16):
And even better than that, Open Stax wants to make
it so the teacher can add his or her own voice,
so that one of the voices on the podcast is
the teacher saying the words.
Speaker 1 (36:28):
All of this is.
Speaker 2 (36:29):
Easily done with the current technology, and it allows this
story shaped hole to be hit. It's not a Kloak
and Dagger mystery about Darwin, but it's a heck of
a lot better than.
Speaker 1 (36:40):
Reading a long list of facts.
Speaker 2 (36:43):
All this leads to the big question, which is what
is it we want our students to learn in school. Well,
one of the most important aspects of school is this
social education, learning all about other people and how to
deal with them, and that's obviously much more than simple
fact gathering. Beyond this, there are two fundamental things we
(37:03):
need them to learn and to get there, I'll step
back to the great German thinker Gerta, who in the
eighteenth century suggested that children need two things from their parents,
roots and wings. That is the ability to know what
has come before them that's the roots, and the courage
(37:24):
to move into the future and create something new that's
the wings. So what does this mean in the context
of our future world? We need to be thinking about
what this translates into for our students, because lots of
jobs will get outdated, just as always happens. Think about
buggy drivers and phonograph makers and telephone operators. There's always
(37:47):
been no point in teaching towards particular vocations, but that's
especially true now. In ten years, it is really difficult
to know what jobs are going to exist. Twenty years
from now, it's probably impossible. Our students of today are
going to have job titles we can't even imagine right now,
(38:08):
and the question is how do we prepare them for it. Well,
in this context, my interpretation of the Gerta quote about
roots and wings translates to this, for schools, everything is
going to change. So the roots we want to teach
our children that's critical thinking, and the wings we want
to give them that's creativity. And I'm going to argue that,
(38:31):
even as everything else changes rapidly around us, these are
the two finest gifts that we can bequeath to the
next generation. So let's concentrate on critical thinking first. The
idea of writing an end of term paper, or memorizing
a bunch of formulas or dates, all of that is
becoming less important because AI can take care of that instead,
(38:54):
the central job for us is to focus on how
to think, how to critically reason, how to make good decisions.
And let me tell you what I think is a
beautiful example of how AI can be leveraged.
Speaker 1 (39:08):
Here.
Speaker 2 (39:09):
One of the things I'm going to discuss with Saul
Kahan next week is the issue of teaching critical thinking
by debate. So, without giving that discussion away, I'll just
say we're all interested in teaching students not to pipe
off with opinions that they believe are incontrovertible, but instead
to teach them how to think by being challenged. Now,
(39:32):
somehow our schools have gotten less good.
Speaker 1 (39:34):
At that training.
Speaker 2 (39:36):
How can we make sure that our students are given
that challenge, that training such that they're able to understand
the other side of an argument well enough to steal
man it. Steel manning is the opposite of straw manning.
It's where you hear someone else's opposing argument and then
you try to reproduce it and possibly even strengthen it
(39:57):
to demonstrate your understanding of it. So, in this way
of using AI, a student takes a hot button topic
and debates it with the friendly but firm AI, and
the student is graded on how well she defended her
position with logical arguments as well as how well she
(40:17):
was able to steal man the computer's argument, and then
the sides are switched. Now, to my mind, this is
one of the most powerful approaches to how AI can
be leveraged in leveling up education. A student can get
the time and patience and practice to meaningfully learn how
to think this way. As I said, we're going to
(40:40):
come back to this in more detail next week, and
now let's turn to creativity in theory. The aim of
the National curriculum is to prepare children for life, right,
but many worry that it has drifted into the realm
of teaching for the next test, and we sometimes miss
the fact that the content of those tests isn't in
the right aim, because we're still training kids for what
(41:04):
was important in the last generation, not the upcoming one.
And I mentioned a moment ago the issue of jobs
that won't even exist anymore and the unimagined jobs that
are coming down the pike. So presumably the important thing
is not to teach rote knowledge, but instead to teach
the ability to learn to be cognitively flexible, to be creative.
(41:27):
And AI is extraordinary at helping with this, students and
all of us can throw the ball farther every time
by having a thought partner. Fundamentally, what's happening in the
brain is that creativity happens because we absorb information from
the world and then remix it by bending it, by
breaking it, by blending it. If you're interested in this topic,
(41:50):
please check out my book The Runaway Species or listen
to episode fifty eight for more on that. And what
AI gives is plenty of practice with this bending, breaking
and blending. It's so good at working with us to
do these remixes. And if we zoom out to AI
and our technology more generally, what it is extraordinary at
(42:12):
is giving us the ability to consume a larger diet,
which gives us the warehouse of materials that we can remix.
This is now something that students have the opportunity to
do at a whole new level, to have an incredibly
expansive storehouse of knowledge, bigger than any generation before them.
(42:34):
For many of us, when we were young, we'd get
our information by going down to the library and pulling
out the Encyclopedia Britannica.
Speaker 1 (42:42):
But now you can get the entirety.
Speaker 2 (42:44):
Of humankind's knowledge from a rectangle in your pocket. That
makes an enormous difference. And you had whatever homeroom teacher
you had in your little town. But now students get
to absorb a larger world due to technology. Our devices
have flattened the world and zero the distances. I run
(43:06):
into kids all the time who say something really smart,
and I.
Speaker 1 (43:10):
Say, how did you know that?
Speaker 2 (43:12):
And they say, oh, I learned that from a Ted
talk where you've got somebody giving the best talk of
their lives in fifteen minutes. To my mind, this is
one of the most beautiful things about AI and technology
in our lives. It's about consuming a larger diet. Children
now have an opportunity which we never had. They can
(43:32):
build whole new worlds with AI. They can figure out
how to implement anything not on a year's timescale, but
in a few dedicated hours. As a result of our
increasingly large diet, we have become a species with a
runaway imagination. Our innate cognitive software has produced a society
(43:54):
with increasingly faster innovation when that feeds upon its latest ideas,
and our students now are on the steepest part of
the curve that we've ever had in the history of
the species. There are now more materials than ever for
them to absorb and remix.
Speaker 1 (44:14):
And the question for us as educators is.
Speaker 2 (44:17):
How do we meet them there on their terms, with
current technology and with their method of learning.
Speaker 1 (44:23):
We're entering a.
Speaker 2 (44:24):
New world with AI in scientific discovery, AI and art
and music, AI and social sciences and engineering and legislation.
And there are patterns emerging that we can't conceive of
with our twenty twenty five minds, but will look obvious
enough in twenty thirty five. So to wrap up this thought,
(44:47):
few endeavors hold as much power for our students of
today as critical thinking and creativity. Everything is going to
look very different from the world we know today. There's
going to be new medical treatments, fl forms of communication,
new works of art, and the road to that future
begins in the classrooms of today. So let's do a
good job making sure that our students have the tools
(45:10):
to go out and build the next.
Speaker 1 (45:13):
Generation of our world.
Speaker 2 (45:15):
So please join me for the next episode, Part two,
where I'll be interviewing Saul Khan, founder of the kN Academy.
As a preview, he's not opposed to teaching, wrote knowledge,
and he'll argue why that is to some degree a
really important thing, and he'll talk about the incredible work
he's done with the con Academy over the years, and
also where things are going with our new world of technology.
(45:38):
So until next week, I'll just end by reminding us
that very little is going to remain status quo in
our society. There are many fronts on which we'll watch
and see what happens, but I suggest the most important
place for us to take action now is with the
education of our children. As the slope of change increases,
(45:59):
we need to keep in how we can make the
right moves now so that we can successfully bequeath to
our children roots and wings. Go to Eagleman dot com
slash podcast for more information and to find further reading.
Send me an email at podcast at eagleman dot com
(46:21):
with questions or discussion, and check out and subscribe to
Inner Cosmos on YouTube for videos of each episode and
to leave comments until next time. I'm David Eagleman, and
this is Inner Cosmos.
What does AI mean for the future of education?
Speaker 2 (00:09):
How should we be rethinking schools to meet the future?
And what does this have to do with the invention
of the printing press in fourteen forty or the calculator
or the spread of Google? And how is this connected
to the eighteenth century German writer Gerta or a digital
replica of the philosopher Aristotle or the two lasting bequeaths
(00:33):
that we should give our children. Welcome to Inner Cosmos
with me David Eagleman. I'm a neuroscientist and an author
at Stanford and in these episodes we dive deeply into
our three pound universe to uncover some of the most
surprising aspects of our lives. Today's episode is all about
(01:04):
the future of education. After all, we're still running our
school systems essentially as we have since the Industrial Revolution,
but now we're in a world of AI. And the
question is are we optimally educating our children for the
future or are we mostly reiterating what we did mispreparing
(01:27):
for a world that will quickly be unrecognizable. After all,
the price of knowledge has dropped precipitously and is rapidly
approaching zero.
Speaker 1 (01:38):
We are so used to having.
Speaker 2 (01:40):
Lots of jobs where you're paid to have specialized knowledge,
like in medicine or law, or engineers or actuaries or
patent examiners or whatever. But with AI that absorbs everything
that we've written and can do an increasingly precise job
of knowing things better than we can.
Speaker 1 (01:59):
In a world where AI.
Speaker 2 (02:02):
Aces the SATs and the MCATs and the LSATs without
breaking a sweat, this is an open question.
Speaker 1 (02:08):
What the future is going to look like.
Speaker 2 (02:11):
I think it would be madness to assume that it's
going to look much like our present in a decade
or two from now. This is what many of us
around Silicon Valley and the world are talking about all
the time. But how does this affect how we are
training the next generation? So for the next two episodes,
we're going to dive deeply into this and we'll see
(02:32):
some very cool directions and surprises. In today's episode, I'm
going to lay the groundwork for how we can think
about brains and education and the two important things that
we need to be doing for the future. And in
next week's episode, I'm going to interview Salkn, the creator
of the con Academy, which is an extraordinary endeavor, which
(02:55):
educates about one hundred million people in the world every year,
and specifically Salini are going to talk about what the
future of education looks like from his point of view
and what tools he's building to meet the moment. So
for today, let's start at the center of the issue,
which is the human brain. The most remarkable thing about
the brain is what we call neuroplasticity, which is the
(03:18):
brain's ability to rewrite itself, to be flexible, to absorb
the experiences of the world and etch them directly into
the circuits of the brain. So brains come into the
world half baked, and they absorb the world around them.
This is how you come to master all the details
of your language and your culture and the knowledge and
(03:42):
the mores of your generation. And human brains are more
plastic than any others in the animal kingdom. They pop
out uniquely unfinished, and this is why a baby giraffe
can walk within forty five minutes. But if you've ever
seen a baby Homo sapiens get born, you know that
it takes.
Speaker 1 (04:01):
Years to figure it out.
Speaker 2 (04:03):
So we have these massively extended infancies because we are
wiring ourselves to the world as opposed to dropping in
more pre programmed. Now, at first blush, it would seem
that having a very long infancy would be a disadvantage,
and in some ways it is, But the end result
(04:24):
is that humans gain capacities that our neighbors in the
animal kingdom just don't have. This is why a giraffe
is doing the same thing giraffes were doing a million
years ago. But humans launch satellites and compose symphonies, and
erect skyscrapers and build quantum computers. No other species builds
a fraction of what we build. And that's because although
(04:48):
most others tend to be tougher than we are, they
don't have the capacity to absorb the world around them
to the degree that we do. In each generation, they
just run the program of being a giraffe. But we
drop into the world and in a handful of years
we have learned almost every important insight and discovery of
(05:09):
the billions of humans before us. We get that, and
then we springboard into the next generation.
Speaker 1 (05:16):
So what does that mean for the way we.
Speaker 2 (05:19):
Think about educating our children. It means our responsibility is
to shape this malleable brain.
Speaker 1 (05:26):
Each child has a forest of tens.
Speaker 2 (05:29):
Of billions of little neurons that are spending every moment
changing the strength of their connections, and plugging and unplugging
and seeking new locations. Whatever they learn in their childhoods,
it's stored in the brain, and it's coded in a
reconfiguration of the network. That's how brains absorb the world.
(05:51):
So our job with education comes down to one main thing,
which is to feed the brains of our children as
rich a diet of high calorie information as we can.
The problem is that schools are largely being run the
same way we've done it for hundreds of years, when
we needed to train children for particular jobs, careers that
(06:15):
we could reasonably assume would be around and lucrative and
solid in twenty years from now. But we're entering a
new world now, a world where previously unimaginable technology, artificial
intelligence that can answer any kind of question this is
available the way that running water is available. Just imagine
(06:36):
if you were Mark Twain or President Garfield, or Ada
Lovelace or Florence Nightingale, the thought of running water would
have seemed insane to you. But in the nineteen thirties
there was suddenly a pivotal decade where most of America
got running water, and suddenly, for the next generation of kids,
it was just like background furniture, such that by the
(06:59):
time you were born, you don't even devote a single
neuron to the fact that you can produce a clean
little river at whatever temperature you want, the second you
want it. That's what artificial intelligence is going to be
for this next generation. Of course, you can open an
app that has read everything ever written by humans, and
(07:19):
of course you can ask it anything at all in
plain English and it will generate an absolutely beautiful answer.
Of course, it can synthesize pieces of data that would
take you lifetimes to read, and it does it instantly
and brilliantly, and soon, of course, you can tell an
AI to act as an agent and go out into
the world and do tasks for you. So here's the
(07:41):
thing I want to emphasize. This world of AI is
going to require different skills than what we were raised
with and it will offer extraordinarily different opportunities. So this
is what we're diving into today. How does current technology
change the future landscape and how can we leverage it
to rewrite our school systems. So first we know that
(08:06):
children are growing up in a different world than the
one that many of us did from their first moments.
Speaker 1 (08:12):
They're in a world of technology.
Speaker 2 (08:14):
News of the pregnancy is posted on TikTok or Instagram,
the news of the birth is spread on WhatsApp and X.
Relatives anywhere on the planet immediately see their pictures. And
they grow up in this world where control all delete
is as basic as ABC. They have no idea why
that key on the keyboard is labeled return. Some of
(08:36):
you may remember typewriters where you had to return the
carriage to the other side of the page.
Speaker 1 (08:41):
All our kids grow up in.
Speaker 2 (08:42):
A world where photographs are instant, whereas when many of
us were children, we had to get the film processed
and it took three days. In the space of a generation,
things change fast, and the world of very young people
now consists of millions of podcasts, billions of web pages,
trily of videos.
Speaker 1 (09:01):
They spend tens.
Speaker 2 (09:03):
Of thousands of hours on mobile phones, video games, texts, emails,
social media. And this is to say nothing about the
continental plate shifts that AI is starting to make, which
will return to in a moment. So as a result,
by the time they're in high school, they have a
different world than many of us grew up with. So
(09:24):
does this change how their brains wire up and how
they view themselves compared to, say, their parents.
Speaker 1 (09:30):
Of course it does.
Speaker 2 (09:32):
Their exposure has been to fast paced, shortened, interactive information,
and what that means is the ways that students learn
and communicate are different. You've probably heard the term digital
immigrants for those teachers who carry old world traditions and assumptions.
The idea is that you couldn't live and teach in
(09:53):
another country unless you learned the language. For many teachers,
the digital world is a second language that was learned later,
and any teacher who was a little behind the technology
ball is now really behind now that AI is here.
So what does all this mean for the brain of
a digital native? Well I mentioned that we come into
(10:15):
the world with brains half baked, and we absorb the
world around us. And the way this works is that
in the brain of a newborn baby, the neurons are
only starting to communicate with one another, but then over
the first two years of life, those neurons begin connecting
extremely rapidly, so by the time you hit two years old,
you have hundreds of trillions of connections and this is
(10:38):
where you max out. You have as many connections as
you're ever going to have, and after that it's all
about pruning.
Speaker 1 (10:46):
Like an overgrown garden.
Speaker 2 (10:49):
Your brain is encoding what it experiences out there. That's
what those connections represent. Whatever you encounter in the world
strengthens some pathways, while other pathways eventually go away. And
that means the process of becoming who you are isn't
about what you gain in the brain, but about what
(11:09):
you lose. Think about it like Michelangelo carving away at
the marble to find the statue inside. That's what your
brain is doing by keeping those connections that resonate with
the world. So how your brain turns out depends upon
what you are exposed to. So given all this is
(11:31):
of course no surprise that technology affects the brain, and
that the current generation carries out tasks in different ways.
With different technologies. They have increased use of some pathways
and decreased use of others, and that gears today's students
towards a different kind of learning. Now, I want to
(11:52):
take a one second tangent, which is that it's very
difficult to do a rigorous scientific study about growing up
digitally and how that affects the brain. Why is it difficult,
It's because you never really have a good control group.
You can't just look at a bunch of students now
and run a comparison against a previous generation, because there
(12:13):
are one hundred other differences between them, including food and
politics and air quality and on and on and on.
And you can't easily find another group of students who
grew up in this moment without being digital, unless they're
amish or deeply impoverished.
Speaker 1 (12:30):
But then there are one.
Speaker 2 (12:31):
Hundred other differences in how they're being raised, and you
don't know what you can attribute to the digital technology
versus something else. So whatever people say are the differences
with this generation, it's difficult to know which of those
has to do with growing up digitally and which has
to do with other issues like politics or pollution, or
(12:53):
changes in sugar intake or less secondhand smoke or legalized
marijuana or whatever, none the less. We can point to
a few things that people have noticed when comparing pre
Internet and post Internet patterns of reading a page of text.
For example, so digital natives tend to make slightly different
eye movements that are essentially f shaped meaning they're more
(13:16):
likely to quickly scan headlines and then the first few
lines of paragraphs before moving farther down the page to
do that. Again, readers from older generations tend to be
more thorough in they're reading and don't skip.
Speaker 1 (13:28):
Around as much.
Speaker 2 (13:30):
Digital natives are also more accustomed to taking in content
that has text and images and videos and interactive sliders
and buttons or whatever, and presumably as a result, it's
tougher for a digital native to concentrate on a long
book that is only walls of text, and so digital
natives seem to be more prone to distractions, as measured
(13:52):
by them moving their eyes off the text entirely. The
cool part is that these details of eye movements can
be tracked, and they're all totally unconscious. In other words,
it's not a signal that can be easily faked because
people don't even know they're doing it. I mentioned this
to illustrate the bigger point that the digital generation is
(14:12):
exposed to different stimuli and that strengthens different pathways in
their brains, and that changes how they learn.
Speaker 1 (14:21):
Fastest and most efficiently.
Speaker 2 (14:23):
So for many of us, when we were growing up,
we got a lot of information taught to us just
in case we ever needed it. So the Battle of
Hastings was in ten sixty six, because you might need
to know that someday, or you might need to do
geometry later in life, or you might need to know
the capital of Indonesia. But education has been transforming because
(14:47):
of the Internet. It's gone from just in case information
to just in time information. Now young people increasingly pull
information on demand. New skills are acquired as needed. Do
you need to understand how to change a bike tire,
you look it up. You need to understand how this
(15:08):
math problem works, or what chemicals to use in this situation,
you look it up.
Speaker 1 (15:27):
Now.
Speaker 2 (15:28):
The reason this new just in time learning is important
is because the learning is not out of context with
the question. When you want to know how something works,
you have the right cocktail of neurotransmitters present, and that
makes all the difference as far as the neuroplasticity goes.
(15:48):
When you are curious, the information sticks and in fact
without any of the neurotransmitter knowledge. The spirit of this
observation reaches back to the ancient Greeks who loose dated
seven levels of learning, with the highest level being when
you're curious about something, then it really sticks. So when
(16:08):
a child asks Alexa or Siri a question and gets
the answer, their curiosity allows the information to have a
better chance of getting edged in than if you simply
stopped the child and said, hey, I need you to
memorize this fact that has nothing to do with your
life right now, and I'm going to ask you about
this later.
Speaker 1 (16:28):
So when you follow some.
Speaker 2 (16:31):
Rabbit hole on Wikipedia, you click on this link because
you're curious, and then the next because you want to
understand that word that you've heard once, and then the
next because there's some fascinating distinction you've never even thought of,
and so on, and eventually you end up twenty clicks
away from where you started. That's an amazing way to
learn because you are maximizing your curiosity at every step.
(16:55):
So each successive generation gets the opportunity for more self direction,
and not surprisingly, our comfort zones come to be different.
Back in the day, most of what we learned was
by text or lecture. It was for the most part sequential,
and one topic was tackled at a time, and most
(17:16):
of our learning was done independently. Now learning has a
very different character. There's way more multitasking, and we learn
from pictures and sound and video, which makes everything a
lot more engaging. A lot of our learning is interactive
with websites or games, and we're often networked with other learners.
Much of our learning now is random access like Wikipedia
(17:40):
or Google or chat GPT. We connect ideas ourselves. Learning
is far more hands on than it ever used to be,
so the way brains seek information has changed, and this
leads me to a one second tangent about the question
of ADHD. Current students are used to constantly changing input
(18:02):
and information sources. They're switching from WhatsApp to instead a
chat GPT. Does that make a student struggle to concentrate
in class? Yes, almost certainly it's a factor. But is
that ADHD in many or most cases? Probably not. This
is a normal response of their brains to the world.
The way they scan and multitask means that sitting down
(18:26):
and being taught from a boring textbook or a droning
lecturer is not sufficiently interactive or engaging for them. It's
more difficult to concentrate on, so the onus becomes on
the educator to meet them halfway. Now, what does that
look like? Well, now, we have the table set for
us to move into the world that we find ourselves
(18:49):
in in twenty twenty five, and specifically the world of
AI as we all know, modern AI is blowing everyone's mind.
Every day we see large language models lms pull things
off that no one programmed them to do or even
really expected them to be able to do, including all
(19:09):
the stuff we're now taking for granted, like summarizing books
or making a video from a text prompt or crushing
the SATs or LSATs or MCATs. And this is why
we find ourselves in an era of discovery more than invention.
And I want to point out that a lot of
the arguments people have been making about AI not being
(19:31):
so good at this or that these have been changing
really rapidly. For example, just a year ago people were
arguing that AI would make silly mistakes about certain things,
and couldn't add thirty two digit numbers with each other,
and sometimes would get the wrong answers in a word problem.
Speaker 1 (19:48):
But in a shockingly.
Speaker 2 (19:49):
Brief amount of time, these shortcomings have all been mastered.
So it's yet to be seen what challenges will remain
and for how long we are in for fit a
ride and it's going to move faster and faster. So
what does this mean for our school systems? Our students
are now in a world where the entirety of humankind's
(20:11):
knowledge has been digested and any piece of information or
any sophisticated analysis is available to you instantly. What does
that mean for the next generation? How will the future
change and what should we be doing about it now? So,
first of all, let's put this in historical context, because
in times like these, it's always useful to remember that
(20:34):
there have always been times like these. So in fourteen forty,
when the printing press was invented in Europe, there were
many thinkers who asserted that this was going to ruin
the next generation. Why because now the answers were just
sitting there.
Speaker 1 (20:49):
They were right there. You don't have to remember much
of anything.
Speaker 2 (20:53):
You just needed to remember where to get the information,
and you could pull it right off the shelf.
Speaker 1 (20:59):
A lot of it.
Speaker 2 (21:00):
Educators in the fifteenth and sixteenth centuries lamented what was
going on here. They were certain this was going to
make everyone intellectually lazy. What they overlooked was the way
that knowledge would continue to explode and the way that
society could springboard off the top of this new technology
of books. Memorizing things was not, in fact the important
(21:23):
part of education. Instead, it was about storing pointers to
where the information could be obtained. Twenty five years ago,
I was having a similar conversation with educators as we
transitioned into a world of the Internet, and a teacher
would ask a question and the kids would just go
home and look up the answer. I don't know if
(21:43):
you remember those days, but everyone's concern was that it
was cheating.
Speaker 1 (21:48):
So what happened.
Speaker 2 (21:49):
It's not that students got dumber, It's that educators got smarter.
We changed our approach to posing the questions to obviate
copy and paste homework. We just stopped asking simple memorizable facts,
and we started doing other things like getting students to
present out loud to the class, or grading each other's
(22:10):
work what new things did they learn, or getting students
to grade the quality of information from different websites, and
generally anything that made the questions an active exploration instead
of a simple fill in the blank. Well, now we
are faced with a new challenge from AI, and the
question is is this the same kind of challenge we've
(22:32):
always had, or is something different this time, and if so,
what and how do we as a society best prepare
ourselves and our students for this future. So let's start
with a prediction about AI that I heard from Steve
Jobs in the nineteen nineties. Jobs was giving a talk
where he imagined that computers could reach this point where
(22:54):
they would be able to understand everything that the great
philosopher Aristotle knew and would be able to simulate his mind,
and you could have a conversation with him. You could
ask Aristotle anything, just like Alexander the Great did. And
I have to say, when Steve Jobs talked about this
in the nineties, it seemed like a full fledged computer
(23:17):
world fantasy.
Speaker 1 (23:18):
On the AI mountain.
Speaker 2 (23:20):
We hadn't even gotten our boots strapped on at the
bottom yet, and yet.
Speaker 1 (23:24):
Here we are.
Speaker 2 (23:25):
Now, We're so far up that mountain that now having
a simulacrum of Aristotle just seems like part of the
background furniture.
Speaker 1 (23:34):
You can do this with any LM.
Speaker 2 (23:36):
You just feed in all the books by someone and
then you have a full, rich conversation. Now, why did
Steve jobs suggestion of having a personalized tutor matter? Well,
it matters because every classroom goes too fast for half
the kids and too slow for the other half. So
a tailored individualized education is a dream that many think
(24:00):
in the education space have had for a long time,
having an expert who could teach you with passion and
patience and in every possible language. That's something that hasn't
existed before, and now it's here. But here's the question
I want to pose. If students were able to sit
and have a conversation with Aristotle, would they or would
(24:23):
they rather be hanging out with other students and making
jokes and getting dates and so on. So I asked
my thirteen year old boy this question of whether he
would use a digital twin tutor, and he found it
not that interesting because he was aware that he would
run out of things to ask pretty quickly, because in fact,
he wants to be with friends and play video games
(24:45):
and run around. Also, there's what I call the problem
of the mismatched internal model.
Speaker 1 (24:51):
A typical student.
Speaker 2 (24:53):
Simply doesn't have in his or her head the same
questions or concepts that Aristotle had in his head, so
there's not much of an interface there. They wouldn't think
to ask him about his notion of four different causes
that underlie everything, or his distinction between substantial properties and
accidental properties of objects, or his theories about virtue or
(25:16):
teleology or the unmoved mover. So if you were a
student lucky enough to sit down with Aristotle, you just
might not know what to ask. You're not that likely
to probe him with really terrific questions. But I think
the dream of a tutor isn't lost. It simply has
to be cast correctly. There's a different way to think
(25:38):
about the concept of tutor, and I think one of
the best examples of this is coming out of the
con Academy. It's called con Migo. Con Migo doesn't tell
you what it knows. Instead, it challenges you to think
critically and to solve problems without ever giving you direct answers.
So you can learn algebra or programming or essay writing.
(25:59):
But the key is that, unlike chatchpt, con migo does
not tell you the answer. It guides you to find
the answer yourself, and it does so with limitless patience.
I won't say more about this now because next week
we'll have Saul Khn of the con Academy on Intercosmos
(26:19):
to talk with us about this, how it came about.
Speaker 1 (26:22):
And where it's going.
Speaker 2 (26:23):
But the thing I want us to note for now
is that this is not a tutor that replaces a teacher,
but one that works with the teacher, and that leads
me to the role that AI is going to be
great at, which is AI as Tha.
Speaker 1 (26:38):
In other words, as a teaching.
Speaker 2 (26:39):
Assistant, AI can do more than just help with the teaching.
It can also help with the job of teaching. There
are all kinds of ways that teachers can use AI
to help with grading, with creating schedules, with tracking attendance
that gives the teachers more time and energy to spend
with the students. AI is super helpful a teacher needs
(27:00):
to write comments to parents on an individual student performance.
If it saves the teachers fifty percent of their time,
that's an incredible win. Or when a teacher has to
make multiple versions of the same test because kids are
going off to debate tournaments, or writing teacher letters of recommendation.
All these are things that AI helps with to reduce
(27:22):
the burden on educators. And the effect of all these
efficiency gains is that it gives the teachers more time
to have real human relationships with the students, and AI
is also going to improve the structure of teaching. A
lot of researchers feel that the big end of semester
(27:42):
exam is going to go away in place of AI
monitoring every problem along the way, so the teacher knows
exactly how a student is doing on the fly. In
other words, with the enormous amount of data available to
collect about learners, I can process and translate that to
provide useful insights to everybody, to teachers, to students, to parents,
(28:06):
and it can give in the moment formative feedback to
the students, so even before they turn things into the teacher,
the AI can differentiate the learning experience based on that
student's need, in other words, tailored education. The main thing
is it can digest all this data on the fly
and make it understandable to everyone. Okay, so what should
(28:28):
we do next in our classrooms. Let's return to what
I said about teachers having to change the way they
gave assignments when Google appeared.
Speaker 1 (28:36):
What do we do with chat GPT?
Speaker 2 (28:39):
Well, first of all, last week I was giving a
talk at a different university and I talked with several
of my colleagues on the faculty there and I asked them,
how are you changing the way that you do assignments
to prevent students from using chat GPT, and many of
them said, well, I changed my assignment up. Instead of
(28:59):
just asking them to tell me the answer, I told
the students to do something new, like compare two ideas
or come up with three different ways of doing something.
And I said, wait a minute, they could still just
do that with their favorite LLM right, And the professors
seemed to have one of two responses. One was, well,
that would violate the honor code if they were doing that,
(29:21):
and the other was I'm pretty sure my students are
not doing that. But the fact is every student is
doing it. There's no reason for a student not to
do it. If you are a student and not using
the available technology while your competition is using it, your
chances of success are slim. So I think we should
(29:57):
assume that all students are leveraging this technology. We need
to work with that reality. But happily, as far as
things to do, it's not that hard for an educator
to modify the assignments. I'll give you an example. This
is what I'm doing in my current class at Stanford.
Speaker 1 (30:14):
We used to have.
Speaker 2 (30:15):
The final assignment as the writing of a twelve page paper.
Now the students run a science experiment on their fellow
classmates and they write up the results. Even if they
get help refining the question that they want to pursue
from an LLM, that's fine. They're still optimizing the question
that they want to pursue and designing the experiment, and
(30:36):
then they're contacting their classmates and running the experiment and
understanding what goes wrong in real life and how they
need to communicate, and finally how to analyze data and
make charts, even.
Speaker 1 (30:47):
If they use GPT for help with that.
Speaker 2 (30:50):
Anyway, it was an easy change for me to go
from a final paper to a final experiment. And we
also have quizzes every class where at the end of
the lecture there are four sessions on a slide and
the students write down the answers on paper. A lot
of teachers are successfully moving in these directions. For example,
many teachers I talk with say they used to give
(31:11):
a big take home exam at the end of the semester.
Now they give a bunch of smaller in class quizzes throughout.
So I just want to emphasize the importance of being
one hundred percent aware of the tools out there so
that educators don't have to pretend that our students are
not using them. And one way that educators can think
about this is actually giving AI assisted assignments where the
(31:34):
students are meant to use AI and couldn't complete it
without it, and then you have oral defenses about what
the student just learned. So just imagine the difference between
having a student making a little diorama or whatever I
did when I was a kid, versus having them build
an AI enhanced science fiction world where the student works
(31:56):
with AI to generate details about the new world, including
the science, the cultures, the creatures, the technology that comes
out of that, and then you have the student give
a performance in front of the class. So it's not
a matter of simply generating it, but refining it and
knowing it and believing it and being able to be
quizzed deeply on the fly about it. You can imagine
(32:19):
these same concepts with AI assisted spoken word performance, where
in combination with AI, a student might generate richer content
to produce on the stage or AI designed next decade
planning from social structures to technology and education.
Speaker 1 (32:37):
And daily life. This is how you give students a great.
Speaker 2 (32:41):
Education because they are learning how to use the tools
of the day to build something creative and understand it deeply. Okay,
so this is a direction where things can go. But
things I think are even more exciting when we look
out a few years. A colleague of mine, Rich Branyak,
runs open Stacks, which is a service that provides textbooks
(33:03):
in eighty five subjects for free. Now I just spoke
with him, and he's put years into this project, but
they're no longer certain that textbooks are the optimal way
to carry information. So open Stacks is trying out many
different things. And by the way, we are in a
long period of discovery here, so some things will work
(33:23):
well and some things won't. One of the things they're
trying is this. Imagine that you need the students to
read chapter seven of this book. You click a button
to pass chapter seven to Google's Notebook LM, or even
better now Google Illuminate, which turns the book into a
podcast with two artificial podcasters having a conversation about the content.
Speaker 1 (33:49):
And you can set.
Speaker 2 (33:49):
All the parameters like which age group this is for,
and the level of difficulty, and so on. I'll play
a brief clip for you from one of Google Illuminates podcasts.
Here it's absorbed Charles Darwin's Origin of Species, and you
can learn about the book in a conversation.
Speaker 1 (34:07):
Let's talk about it. Absolutely.
Speaker 3 (34:10):
This is a classic foundational work in the field of
evolutionary biology. It's amazing to think that it was published
way back in eighteen fifty nine.
Speaker 1 (34:20):
Eighteen fifty nine, that's quite a while ago. Can you
tell us a bit about what makes this work so important?
Speaker 3 (34:24):
You bet, Darwin's theory of evolution by natural selection was
a total game changer.
Speaker 2 (34:30):
Now I'm only playing a bit of this, but it
goes on to mention the key ideas and to break
those down even further, such that in a short podcast
you get a really good summary of the key ideas. Now,
the thing I want to emphasize is that by this
clever trick of turning it into two people talking, it
becomes more interesting than a bulleted list of facts about
(34:55):
the book. Why because it gets closer to a story,
and we have a story shaped hole in our brain,
and that's what allows things to go in rather than
simply bullet points of facts. Now, it's not necessarily easy
to turn everything into a story. But this technology that
turns a book into a podcast gets us part way
(35:17):
there because instead of somebody writing, here's a fact, here's
a fact, here's a fact, now you have someone saying, wait,
I'm confused, and the other person says, ah, hang on,
this is amazing, and the first person says, wow, that
really blows my mind. And now suddenly you're recruiting parts
of the brain that we study in a subfield called
social neuroscience, which is just to say, when other people
(35:39):
are involved, we care more. And that's a big part
of why we care about story. So this quick trick
of converting a chapter into a podcast dialogue with a
click allows story to be approximated and gets you part
way there. And it's instant and free. And that's not all.
If you, as a teacher, want to emphasize some aspect
(36:03):
of the chapter in particular, you just steer the prompt
that way. So open Stax is experimenting with saying, here's
chapter seven of this textbook, but I really want my
kids to concentrate on this aspect of it and not
spend much time on.
Speaker 1 (36:15):
This other aspect.
Speaker 2 (36:16):
And even better than that, Open Stax wants to make
it so the teacher can add his or her own voice,
so that one of the voices on the podcast is
the teacher saying the words.
Speaker 1 (36:28):
All of this is.
Speaker 2 (36:29):
Easily done with the current technology, and it allows this
story shaped hole to be hit. It's not a Kloak
and Dagger mystery about Darwin, but it's a heck of
a lot better than.
Speaker 1 (36:40):
Reading a long list of facts.
Speaker 2 (36:43):
All this leads to the big question, which is what
is it we want our students to learn in school. Well,
one of the most important aspects of school is this
social education, learning all about other people and how to
deal with them, and that's obviously much more than simple
fact gathering. Beyond this, there are two fundamental things we
(37:03):
need them to learn and to get there, I'll step
back to the great German thinker Gerta, who in the
eighteenth century suggested that children need two things from their parents,
roots and wings. That is the ability to know what
has come before them that's the roots, and the courage
(37:24):
to move into the future and create something new that's
the wings. So what does this mean in the context
of our future world? We need to be thinking about
what this translates into for our students, because lots of
jobs will get outdated, just as always happens. Think about
buggy drivers and phonograph makers and telephone operators. There's always
(37:47):
been no point in teaching towards particular vocations, but that's
especially true now. In ten years, it is really difficult
to know what jobs are going to exist. Twenty years
from now, it's probably impossible. Our students of today are
going to have job titles we can't even imagine right now,
(38:08):
and the question is how do we prepare them for it. Well,
in this context, my interpretation of the Gerta quote about
roots and wings translates to this, for schools, everything is
going to change. So the roots we want to teach
our children that's critical thinking, and the wings we want
to give them that's creativity. And I'm going to argue that,
(38:31):
even as everything else changes rapidly around us, these are
the two finest gifts that we can bequeath to the
next generation. So let's concentrate on critical thinking first. The
idea of writing an end of term paper, or memorizing
a bunch of formulas or dates, all of that is
becoming less important because AI can take care of that instead,
(38:54):
the central job for us is to focus on how
to think, how to critically reason, how to make good decisions.
And let me tell you what I think is a
beautiful example of how AI can be leveraged.
Speaker 1 (39:08):
Here.
Speaker 2 (39:09):
One of the things I'm going to discuss with Saul
Kahan next week is the issue of teaching critical thinking
by debate. So, without giving that discussion away, I'll just
say we're all interested in teaching students not to pipe
off with opinions that they believe are incontrovertible, but instead
to teach them how to think by being challenged. Now,
(39:32):
somehow our schools have gotten less good.
Speaker 1 (39:34):
At that training.
Speaker 2 (39:36):
How can we make sure that our students are given
that challenge, that training such that they're able to understand
the other side of an argument well enough to steal
man it. Steel manning is the opposite of straw manning.
It's where you hear someone else's opposing argument and then
you try to reproduce it and possibly even strengthen it
(39:57):
to demonstrate your understanding of it. So, in this way
of using AI, a student takes a hot button topic
and debates it with the friendly but firm AI, and
the student is graded on how well she defended her
position with logical arguments as well as how well she
(40:17):
was able to steal man the computer's argument, and then
the sides are switched. Now, to my mind, this is
one of the most powerful approaches to how AI can
be leveraged in leveling up education. A student can get
the time and patience and practice to meaningfully learn how
to think this way. As I said, we're going to
(40:40):
come back to this in more detail next week, and
now let's turn to creativity in theory. The aim of
the National curriculum is to prepare children for life, right,
but many worry that it has drifted into the realm
of teaching for the next test, and we sometimes miss
the fact that the content of those tests isn't in
the right aim, because we're still training kids for what
(41:04):
was important in the last generation, not the upcoming one.
And I mentioned a moment ago the issue of jobs
that won't even exist anymore and the unimagined jobs that
are coming down the pike. So presumably the important thing
is not to teach rote knowledge, but instead to teach
the ability to learn to be cognitively flexible, to be creative.
(41:27):
And AI is extraordinary at helping with this, students and
all of us can throw the ball farther every time
by having a thought partner. Fundamentally, what's happening in the
brain is that creativity happens because we absorb information from
the world and then remix it by bending it, by
breaking it, by blending it. If you're interested in this topic,
(41:50):
please check out my book The Runaway Species or listen
to episode fifty eight for more on that. And what
AI gives is plenty of practice with this bending, breaking
and blending. It's so good at working with us to
do these remixes. And if we zoom out to AI
and our technology more generally, what it is extraordinary at
(42:12):
is giving us the ability to consume a larger diet,
which gives us the warehouse of materials that we can remix.
This is now something that students have the opportunity to
do at a whole new level, to have an incredibly
expansive storehouse of knowledge, bigger than any generation before them.
(42:34):
For many of us, when we were young, we'd get
our information by going down to the library and pulling
out the Encyclopedia Britannica.
Speaker 1 (42:42):
But now you can get the entirety.
Speaker 2 (42:44):
Of humankind's knowledge from a rectangle in your pocket. That
makes an enormous difference. And you had whatever homeroom teacher
you had in your little town. But now students get
to absorb a larger world due to technology. Our devices
have flattened the world and zero the distances. I run
(43:06):
into kids all the time who say something really smart,
and I.
Speaker 1 (43:10):
Say, how did you know that?
Speaker 2 (43:12):
And they say, oh, I learned that from a Ted
talk where you've got somebody giving the best talk of
their lives in fifteen minutes. To my mind, this is
one of the most beautiful things about AI and technology
in our lives. It's about consuming a larger diet. Children
now have an opportunity which we never had. They can
(43:32):
build whole new worlds with AI. They can figure out
how to implement anything not on a year's timescale, but
in a few dedicated hours. As a result of our
increasingly large diet, we have become a species with a
runaway imagination. Our innate cognitive software has produced a society
(43:54):
with increasingly faster innovation when that feeds upon its latest ideas,
and our students now are on the steepest part of
the curve that we've ever had in the history of
the species. There are now more materials than ever for
them to absorb and remix.
Speaker 1 (44:14):
And the question for us as educators is.
Speaker 2 (44:17):
How do we meet them there on their terms, with
current technology and with their method of learning.
Speaker 1 (44:23):
We're entering a.
Speaker 2 (44:24):
New world with AI in scientific discovery, AI and art
and music, AI and social sciences and engineering and legislation.
And there are patterns emerging that we can't conceive of
with our twenty twenty five minds, but will look obvious
enough in twenty thirty five. So to wrap up this thought,
(44:47):
few endeavors hold as much power for our students of
today as critical thinking and creativity. Everything is going to
look very different from the world we know today. There's
going to be new medical treatments, fl forms of communication,
new works of art, and the road to that future
begins in the classrooms of today. So let's do a
good job making sure that our students have the tools
(45:10):
to go out and build the next.
Speaker 1 (45:13):
Generation of our world.
Speaker 2 (45:15):
So please join me for the next episode, Part two,
where I'll be interviewing Saul Khan, founder of the kN Academy.
As a preview, he's not opposed to teaching, wrote knowledge,
and he'll argue why that is to some degree a
really important thing, and he'll talk about the incredible work
he's done with the con Academy over the years, and
also where things are going with our new world of technology.
(45:38):
So until next week, I'll just end by reminding us
that very little is going to remain status quo in
our society. There are many fronts on which we'll watch
and see what happens, but I suggest the most important
place for us to take action now is with the
education of our children. As the slope of change increases,
(45:59):
we need to keep in how we can make the
right moves now so that we can successfully bequeath to
our children roots and wings. Go to Eagleman dot com
slash podcast for more information and to find further reading.
Send me an email at podcast at eagleman dot com
(46:21):
with questions or discussion, and check out and subscribe to
Inner Cosmos on YouTube for videos of each episode and
to leave comments until next time. I'm David Eagleman, and
this is Inner Cosmos.
Host
David Eagleman
Popular Podcasts
40s and Free Agents: NFL Draft Season
Daniel Jeremiah of Move the Sticks and Gregg Rosenthal of NFL Daily join forces to break down every team's needs this offseason.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
Las Culturistas with Matt Rogers and Bowen Yang
Ding dong! Join your culture consultants, Matt Rogers and Bowen Yang, on an unforgettable journey into the beating heart of CULTURE. Alongside sizzling special guests, they GET INTO the hottest pop-culture moments of the day and the formative cultural experiences that turned them into Culturistas. Produced by the Big Money Players Network and iHeartRadio.