September 9, 2026 • 35 mins
Are kids finally going to be safer on social media? Jonathan Haidt thinks a landmark $17.1 billion settlement with Meta is a gigantic step. Not because of the money, but because of the restrictions the company agreed to put in place for young users in the US. And he believes the world will follow.
Haidt, author of The Anxious Generation, joins Oz to explain why he believes the evidence linking social media to the teen mental health crisis is now overwhelming. They debate the early results of Australia’s under-16 social media ban, the privacy concerns around age verification, and why Haidt thinks it’s too early to say the ban is a failure.
See omnystudio.com/listener for privacy information.
Listen
Watch
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:17):
Welcome to Tech Stuff. I'm Oz Veloshin, joined today by
Jonathan Haidt, one of the most prominent and arguably most
effective tech critics in the world. He's the author of
The Anxious Generation, a book that proposes new norms to
address the growing teen mental health crisis. In late August,
the New York Times ran a story with the headline,
A $ 17. 1 Billion Vote of Confidence That Jonathan Haidt Was Right.
(00:42):
The story describes Meta's landmark settlement with almost all US
states over claims its technology has harmed children. So we're
going to talk about that today. We're also going to
talk about the efficacy of social media bans more broadly.
In December 2025, Australia became the first country to ban
social media use for users under 16. It's been about
nine months since that ban went into effect. And a
(01:05):
lot of early assessments of the ban are focused on
the fact that many kids are finding workarounds and still
using social media. Some are even calling the bans ineffective. Jonathan,
I believe, begs to differ.
Speaker 2 (01:16):
Jonathan, welcome. Thanks so much, Oz. Great to be with you.
So let's start with.
Speaker 1 (01:19):
The META settlement. $ 17. 1 billion over allegations that Instagram and
Facebook were intentionally designed to addict children and harm teen
mental health. How did you first learn about this settlement
and what was your reaction?
Speaker 2 (01:32):
Well, of course, we'd known that the case was coming
for a long time. There are literally thousands of cases
against META. This case was the biggest of them all
because this was all of the attorneys general representing the
chief justice official of every state or almost all the states.
And so when I heard the news last Wednesday and
I saw the number, 17 billion, I was disappointed at first.
(01:55):
This is very much like the tobacco settlement. The tobacco companies,
they had a dangerous product. They knew it was dangerous.
They did internal research. They buried the research. They didn't
let their kids smoke. All of that is true for Meta.
All of it is true for the social media companies.
And that settlement in the 90s, I think, was around
$ 200 billion then. So that would be more like $ 400
billion in today's dollars. So I was actually disappointed in
(02:16):
the number because, really, they have to pay $ 12 billion
over 10 years. They're not going to notice that.
Speaker 1 (02:21):
De minimis for them.
Speaker 2 (02:22):
So I was disappointed. I was disappointed in them. But
I was surprised by the changes that they agreed to
make because they were pretty substantial. They weren't just cosmetic things.
It was, if you're under 18, you can't go on
between midnight and 6. So that's the biggest one. And
the heavy users are the ones who are most damaged.
And they're on it two, three, four in the morning.
(02:43):
So as I was reflecting on that, what I realized,
it took me a few days to realize this, is
that the rest of the world is going to do this.
That is, dozens of countries want to follow Australia and
are at various stages. And there are bills being drafted
right now. And the EU is committing to doing it.
So the world is going to follow Australia in various
forms and with innovations that we'll talk about. And it
(03:04):
suddenly dawned on me, like Thursday or Friday, like, wait
a second. all these countries that are going to do
the age limit thing, they're also going to require all
the social media companies, not just Meta, to have off periods.
So I realized this is like the biggest event we've
ever seen in terms of trying to limit these companies'
damage to kids. So I'm actually thrilled by what the
(03:26):
attorneys general achieved.
Speaker 1 (03:28):
So the dollar amount is more or less irrelevant, but
the framework created by this agreement is something that you
think could be replicable globally.
Speaker 2 (03:38):
Yes, absolutely. Of course, the money matters because these companies
only respond to financial incentives. They have not responded to
any kind of moral appeals. They've not responded to their
own findings of harm. So the money does matter. Now, fortunately,
there are still literally thousands of lawsuits pending, thousands and thousands.
So I am hoping that ultimately, Med will have to
pay out hundreds of billions of dollars.
Speaker 1 (03:58):
What were you happiest that was in this settlement? And
what do you think is missing to truly address the
harms that you've covered and researched.
Speaker 2 (04:08):
So last week, a lot of the talk was about like, well,
they're not doing anything about the central harm, which is
the algorithm, the central addictive property. And some people said like,
it's like saying you can only smoke 10 cigarettes a day,
but you haven't done anything to make cigarettes less harmful.
So that's, I think that's a fair description of it.
But again, by the end of the week, I was like,
(04:30):
wait a second. Like if the whole world were to
say that kids can only smoke 10 cigarettes a day,
You know, when I was a kid, a lot of
kids smoked a pack a day in high school. You know,
most kids aren't smoking a pack a day, but some
kids are smoking five packs a day. And now they
can only smoke a half a pack a day. So,
of course, more needs to be done. But, you know,
a journey of a thousand miles begins with a single step.
(04:51):
And this was a gigantic step.
Speaker 1 (04:54):
And what about this New York Times headline about the
$ 17. 1 billion vote of confidence that you were right? I mean,
what do you make of that?
Speaker 2 (05:02):
It's probably the favorite headline I've ever seen in my
entire life. I was very pleased. Because there has been
a debate. There's an academic debate. So when the book
came out in 2024, it was astonishing how little criticism
there was. I was expecting a lot. Nobody criticized anything
about the play side, the overprotection. Not a single person
(05:24):
has ever criticized that. And the academic debate has been, well,
is there evidence of causality? Is it correlation? Is it causation?
That's what the main academic debate has been. I hope
we'll talk about that because I think it is now
decisively answered. There are so many different kinds of evidence.
One of those kinds of evidence was Meta's internal research.
(05:45):
And that's what these cases hinged on. What they all
turned on was not so much the academic research. It
really was the direct statements made by their employees. And
so here, I'll read you a few. Meta did hire
a lot of social psychologists and clinical psychologists. And so
here's a conversation between a couple of them. One says, quote,
oh my gosh, y'all, Instagram is a drug. We're basically pushers.
(06:08):
We're causing reward deficit disorder because people are binging on
Instagram so much they can't feel reward anymore. Like the
reward tolerance is so high. Then he says, I know
Adam Mosseri, the head of Instagram. I know Adam doesn't
want to hear it. He freaked out when I talked
about dopamine and my teen fundamentals, but it's undeniable And
there's a lot more. It seems clear from what's presented
(06:30):
here that some of our users are addicted to our products.
I mean, these guys, they knew the science of what
they were doing to kids. If you design a product
to be maximally addictive, that's not free speech. That's not
Section 230. That is a defective product. And there are
just so many ways that Meta just didn't... I mean,
they felt like, we don't have to worry about the law.
And so they got nailed for a lot of things.
(06:51):
So COPPA, the Children's Online Privacy Protection Act, says... how
old do you have to be before a company can
take your data without your parents' knowledge or permission? And Congress,
in its wisdom, said 13. It should have been 16.
The original bill was 16, but it got lowered in
negotiations in 1998, I think it was, to 13. Okay,
all right, it's 13 for the world. Well, all these companies,
(07:12):
especially Meta, they're going after the eight, nine, 10-year-olds, and
they have them, and they know how old they are.
And we just accepted that. Like, oh, you know, okay,
I guess they can just take data from anyone No
need to tell the parents. Like that's insane that this
is just what we accepted and that's over. So they
got nailed for taking data from children without the parents' knowledge.
So that's incredibly illegal and they were nailed for that.
(07:33):
They got nailed for making a defective product if you
have a defective consumer product and it's not about the content,
it's about the design and you know it and you've
known it for years and you have all your researchers
talking about them. Like this is insane that this went
on for so long. So this is all gonna change.
Like the world is realizing this is completely insane you
and I will talk about the Australia law and age verification,
(07:56):
but I just want to establish the status quo is
not an option. Like this has to change. It's just
a question of how we change it. It is not
an option to just say, leave it as it is.
Speaker 1 (08:06):
Talk about the publication of The Anxious Generation, because I
refer to as one of the most prominent and arguably
most effective tech critics in the world. And prominent is
one thing, but effective is another. What is it about
you and your work? I mean, as compared to Let's
just say, for example, send a few main technology, Tristan Harris,
you know, great guy. He's been on the podcast, et cetera.
But what is it about Jonathan Haidt and you and
(08:29):
your message that has essentially catalyzed this global conversation?
Speaker 2 (08:35):
One word, mothers. It's mothers. That's it. If there is
a raging forest fire bigger than anything we've ever seen,
nobody is saying, I wonder how big the match was.
Let's find out. What kind of a match could start
such a forest fire? It's not about the match. It's
about the dryness of the forest. And the world in
(08:57):
March of 2024, when The Anxious Generation came out, was
a world full of families. In every family, the kids
had electronics. In every family, if the kid was 9, 10,
11 years old, he had an iPad or an iPhone.
In every family, it was constant fighting over that. In
every family, the parents, especially the mothers, the mothers are
(09:18):
much more sensitive to their wonderful, playful, happy kid, which
we all had when they were little. And then you
give them a device and they get glued to it
and then they go down. And mothers really felt that.
And they were desperate. They knew something was wrong, but
they were like, oh, you know, it's just correlation. Like,
oh no, this is how kids connect today. Like, oh, there's,
you know, pluses and minuses. So there was a lot
(09:38):
of confusion about what to do. And so when the
anxious generation came out, you know, I knew that it
would be popular. I knew that people would buy it,
but The thing that I think I really did contribute
is as a social psychologist, I was able to see
the collective action problem. The other researchers were not social psychologists,
they were developmental psychologists. They were focused on development. They
(09:59):
didn't fully appreciate it. that each kid has to be
on because the other kids are on. And so I
laid out a series of collective action problems. And I said,
the way out of a collective action problem is to
act collectively. So Talk to the parents of your kids'
friends and agree. You're all going to do this. You're
all going to delay smartphones. You're all going to delay
social media. You're going to go to your school. You're
(10:19):
going to say, let's have phone-free schools. You're going to
all send your kids out to play together. So those
were the four norms right there. And because I suggested
four specific behavioral norms that are easy to do, they
cost zero dollars. They're free. You can do them today.
And because the mothers of the world were at their
wits end, they were desperate. So as soon as the
book came out proposing this, boom, they jumped into action.
(10:41):
And that's why it happened all over the world. Brazil's
a great example. I didn't do anything in Brazil for
the first year, but some others there got a copy
of the book in English before it was translated to Portuguese,
and they jumped into action. They got Brazil banned phones
in school by November of 2024. And the book only
came out in Portuguese in July. So it really was
(11:01):
the mothers.
Speaker 1 (11:03):
So, I mean, we've skirted around this controversy element in
terms of correlation and causation. Let's address it head on.
There was a a paper in Nature by Candice Hodges,
who leveled a very trenchant critique against your work, saying
that you had fundamentally misunderstood correlation and causation. And that
critique caught on, and other people began repeating it. Describe
(11:25):
that moment, and where is the argument today? I mean,
how have you defended your work?
Speaker 2 (11:30):
All right, sure. So two things about that critique. So
it came out in Nature a few weeks after the
book came out. And you're right to use the word repeat.
That critique has been repeated. But that's really the only one.
Everybody cites that one article. There are one or two
other small things, but that's the one that everybody cites.
But if you read that article, and I encourage people
to read it, just, you know, put in Candace Auger's nature,
(11:53):
nothing in what she wrote suggests that she read past
chapter one. Chapter one, I lay out the correlational evidence.
Chapter one, I let the historical correlation. Just, you know,
here are all these graphs, look what happened. That's chapter one.
And she says, she accuses me of not understanding the
difference between correlation and causation. Okay. Throughout the book, whenever
(12:14):
I present findings, I always say, so here's the correlation,
but correlation doesn't prove causation. Let's look at the experiment.
So I'm incredibly careful about this in the book. And
then when we get to chapter six, which is the
chapter on girls, it starts off with a long section.
The evidence that social media harms girls and includes the
section is social media cause or just a correlate. So
(12:37):
I've been as careful and as explicit as could be
in all the Google Docs since 2020, in the book,
in everything I write. So it was just unfair. I
think it was just really unfair of her to say
that I've confused the two or that I wasn't careful
about it. At the time, there were multiple kinds of evidence.
But what I've done since then, what my team, the
Tech and Society Lab at NYU Stern, what we've done
(12:57):
is we've aggregated the evidence into a major review paper
that is published in the World Happiness Report 2026. So
if you've just type in World Happiness Report, and then
my name, Haidt, H-A-I-D-T, you'll find it. And we lay
out these seven lines of evidence as if we were
prosecutors in a court case. And so the correlational studies,
(13:19):
those don't prove causation, but the experiments kind of do.
I mean, they push you in a Bayesian sense, much
more confidence in causation. The longitudinal studies mostly point that way.
And the natural experiments point that way. So there's four
bodies of academic research that all point that way. It's
a great majority of the studies and the meta-analyses point
(13:40):
towards causation. So that's the academic evidence. But then we say, well,
if a woman accuses a man of having assaulted her,
attacked her, punched her, and she says, he did it.
I know the guy, he did it. Is that evidence?
Of course it is. In a court, that's evidence. It's
not a correlation. She was there. She knows the guy
(14:00):
and she points to him. Well, that's what Gen Z says.
When you survey Gen Z around the world, there's very
little gratitude to social media. The majorities generally say that
it has hurt their generation. And between 20% to 40%, 20%
to 30%, depending on the question, say it hurt me,
like girls especially. It hurt my mental health. It damaged
my sleep. It damaged my body image. So if the
(14:20):
victim says he did it, that's the guy, it was
Instagram and TikTok and all those So that's evidence. And
then if you have eyewitnesses, people who saw like, yeah,
I saw him punch her. Like, is that evidence? Is
that correlation or causation? That's evidence. And we have that.
So you survey parents and teachers and therapists. And what's
(14:42):
so interesting to me is that missing from this whole
debate are parents defending social media or teachers or therapists
or Gen Z. You don't have organizations of young people saying, No, no,
don't raise the age. We love it. It's so good
for us. You don't have that. And so I think
the silence of the defenders of social media is deafening.
(15:04):
And then finally, the strongest kind of evidence you could
have is if you had emails and text messages from
the perpetrator saying, I'm going to mug this girl tomorrow
at three o'clock to his friends. And then he has
photographs of it. And then afterwards he says, yeah, I
mugged her at three o'clock yesterday. And if the prosecutor
has those and he presents it to the jury, is
that just a correlation? No, that is... proof positive of causation.
(15:27):
And we have that. That's the internal document. So with
seven different lines of evidence, including meta itself saying that
they're hurting kids, I don't know what's going on. I
don't know why some people still say, oh, there's no evidence.
How could you possibly tell correlation from causation?
Speaker 1 (15:43):
So what is your academic work today? Do you feel
like it's case closed or are you still pursuing the
smoking gun on the causation versus correlation?
Speaker 2 (15:53):
Yeah. So in science, it's never case closed because we
can't actually answer a question directly. We have to first
operationalize it. And so we have to first say, okay,
social media, we can't really measure it all. Let's just
take a proxy. How many hours a day do you
spend online? That's a terrible proxy, but that's what we use.
(16:16):
And then on the output side, we don't really know
what's in people's head, but we're going to ask them
a question on a scale of one to five, how
anxious are they? So So we have bad measures at
the beginning, bad measures at the end. And all we're
really looking at is time spent on the device. We're
not literally studying what's happening. We're just like, almost like
feeling into a box where you can't really see what's
going on, but you kind of get a sense from
(16:36):
your hands, but you could be wrong. So I would
never say case closed. What I would say is given
the way that we are now conceptualizing it, I think
the evidence is overwhelming in particular because it's from so
many different sources. I always think about this as a
Bayesian sort of situation, like Bayes' theorem, like certain things happen.
Should they cause you to adjust your belief in causality
(16:58):
up or down? And everything should cause people's belief to
go up. And the clincher for me is this. The critics,
they do point to studies that find no effect. So
of course, I mean, of course they have evidence. They're
not making this up. So they point, here's a study
that didn't find an effect. Here's a meta-analysis that didn't
find an effect. So my team, a lot of our
publications are, we go into those and what we always
(17:19):
find is that the studies that fail to find a
significant correlation or experimental effect, they blended lots of things together.
So the classic one was a study done in 2019
that showed that the correlation is about as large as
the correlation of mental illness with eating potatoes. It's so
close to zero. But in that study, what they did
(17:40):
was they mixed together boys and girls, they mixed together
social media use with everything else, including watching television. So
just like you look at everything and on the output side,
they put in everything and they got no real relationship.
But what we did, we took the same data sets
and we said, let's just look at girls and let's
(18:01):
just look at social media, not everything. And let's just
look at the questions about anxiety and depression, not everything.
And the effect size goes from a betas of around 0.04,
which truly is tiny. two betas of around 0.2, which
is quite large. I mean, you don't get, because of
the crappy measurement on both ends, you don't get like
correlations of 0.5. You don't get, you know, effect sizes of,
(18:24):
you know, of a full standard deviation or a half
standard deviation. So given the general range of effects, what
we always find is you show me a study that
failed to find an effect, I'll unblend it. We'll find, oh, look,
if we look at girls in social media, we find it.
Speaker 3 (18:42):
So talk about the australia ban and i mean this
year for you.
Speaker 2 (19:08):
In general?
Speaker 1 (19:08):
I mean, was it meant for you to see your
work in the world? And what do you hope for
from the next five years?
Speaker 2 (19:15):
So it's been thrilling to see the world going through
a tipping point. And here I draw on Steve Pinker
has a wonderful book called When Everyone Knows That Everyone Knows.
And it's about what happens in a society where people
might know something beforehand, like everybody knew that communism didn't work.
Everybody knew that the government in these countries was corrupt.
(19:36):
But if you spoke up, you'd be killed or abducted.
So nobody spoke up. But then what happens when somebody
speaks up, then suddenly everybody knows that everybody knows that
everybody knows that this is terrible. And that's where we
were with social media. Everybody could see this is really
hurting kids. But they didn't necessarily know that everybody knew
(19:57):
and that there was something they could do about it.
And so what happened when Australia's law went into effect
was December 10th. What happened when it went into effect
was there was a lot of coverage around the world,
a lot of commentary on social media. And a lot
of it was people saying, why can't we do that
in our country? Or, hey, what a great idea, let's
do it in our country. And when everybody saw that
everybody wanted to do it, that's when political leaders saw,
(20:21):
wait a second, this is not some radical thing that's
going to cause a blowback among my people. This is
what adults in my country want. And polling always shows that.
Large majorities of adults from every political party, this is
not a left-right issue. It's one of the only issues,
not left-right. If you're a parent, you hate social media,
you want your kids protected. And so Australia, God bless them,
(20:42):
they just did it straightforwardly, simply. The premier of South Australia,
Peter Malinowskis, his wife read The Anxious Generation, and she
turns to him, she's reading it in bed, she turns
to him one night and says, Peter, you've got to
read this book and you've got to fucking do something
about it. That is a direct quote. I don't know if, well, anyway,
he's used the, he sometimes says, bloody well do something
about it. So he contacted a former chief justice of
(21:05):
the Supreme Court and said, What's the legal mechanism by
which we could do this? And the clear answer is
the signing of the contract. Whenever you open a social
media account, you sign a contract, terms of service, which
you say, yep, you can take everything from me, everything
I post, every photo, you own it, you take it,
all my private, you take it. And that's insane that children,
(21:27):
that a nine-year-old can say she's 13 and then can
sign that contract. That's insane. So the Australia bill was
basically saying, you can't sign that contract until you're 16.
And you know who has the responsibility for enforcing it?
Not the parents. It's the company that's making the goddamn money.
They're the ones who have to enforce it. So that's
what's historic about Australia. The companies have never borne any responsibility.
(21:49):
They could make whatever crappy consumer product they wanted, and
they got off scot-free. And Australia said, no, no, no.
You actually have to enforce the age limit.
Speaker 1 (21:58):
When we first started talking about having this conversation, it
was before the meta-settlement and the thing that we agreed
to talk about was this Australia ban, and in particular,
your sense that a lot of the critics of it
have missed the point. And so I'm curious, what are
the most trenchant criticisms and why are they wrong? Can
they be wrong and trenchant?
Speaker 2 (22:17):
So the main criticism is that the rates of social
media use are still very high. They went down somewhat.
There's some studies that show it may have even come
back a lot. So Australia passed a law The use
of social media went down, but it wasn't like down 90%.
It was down, we have an article in Time Magazine.
We review five different studies that looked at Australia's use.
(22:39):
And the use of various accounts goes down. And so
people say, oh, I guess it failed. Most kids are
still on, must have failed. And my response to that
is the Australians were the first in the world to
do this. So they had a going kind of soft.
What they did is they said, Okay, companies, you've never
done anything to verify ages. None of you have ever
done anything to verify ages. You have to start. And
(23:01):
they did. They all started. And a lot of them
used YOTI, which is you do this face scan to
show that you're over a certain age. And it's easy
to get around that. But they did something. All of
them did something. And that's historic. Okay, so that's round one.
This is what the critics don't understand. Of course, they
expected that the kids would get around it. It's still
quite easy to get around it. That's round one. Now,
(23:23):
the regulator, the E-Safety Commission, Now they have the data showing, hey,
you know, hey, meta, kids are still on your platform.
Kids under 16, kids under 13 are still on your platform.
You have to do more. And they sent those letters
out recently, I forget, in the last month or so.
And now the companies are told you have to now
do better on the second round. And then there will
be a third round. So that's the first thing critics
(23:44):
don't understand, that there are multiple rounds and round one
was a success. It didn't get the rate way down,
but it went through. The sky didn't fall. All the
companies complied. They didn't comply well. Okay, that's round one.
Second thing the critics don't understand is that the law
is not written to keep kids off social media. It's
written to keep them from opening accounts. If you have
(24:05):
an account, you can post, your life gets taken over
by what you post, you're talking with strange men who
want sex or money from you, all that stuff happens
when you have an account. But they can go on
YouTube as much as they want. They can go on YouTube,
they can search for how do I kill myself, and
they'll get videos telling them how to kill themselves. But
what they won't get is an algorithm that knows them
over time and that feeds them videos about how to
(24:28):
kill yourself. So when you survey, have you been on
YouTube or Snap or TikTok? The answer is yeah. They've
almost all been on these platforms since the bill went
into effect. On YouTube, there's no restriction. They can watch
as much as they want. TikTok and Instagram, they restrict.
You can't just keep scrolling if you don't have an account.
But that's the second misunderstanding is that it's about accounts,
not about access. And the third misunderstanding is is that
(24:53):
we were stuck in a collective action problem where because
age verification isn't great and easy, people say, oh, well,
therefore we shouldn't do it. But how does age verification
get great or easy? When companies are incentivized to improve it.
And now that it's clear, the world is going to
do this. Like this is not a question of whether.
This is not a question of if. This is just
(25:14):
a question of how. The world is going to age
gate social media. Just as it age gates gambling and
alcohol sales, it's going to age gate social media. And now,
all of a sudden, Apple and Google are incentivized to
make it easy. And they have, just since the Australia
Bill was passed. Apple and Google have both made changes
to their wallet so that if you have a driver's license,
(25:35):
a passport, anything, and you put it in your wallet,
now Google and Apple can vouch for you. Which means
if you want to open an account, you don't have
to show your ID. You don't need to give your name.
You don't need to give your address. You don't need
to give your age or your birthdate. The platform just says,
is this person over 16? Yes or no. One bit
(25:56):
of information. Yes or no. That's it. That's as privacy
preserving as can be. Porn sites want to do this.
They want to use this. And I think they're right.
If you want to go to a porn site, if
you use the device-based verification, you don't have to give
your address. You don't have to give your name. You
don't have to give your birthday. If you go to
the site, the site will say, is this person over 18?
(26:18):
And if the answer is yes, then you can get in.
So the point is, what Australia did is it kicked
off a new market, which is a market for quick
privacy-preserving age verification, which is not, it does not require
you to give away your identity. It does not. And
that's a major misunderstanding. My close friend, Greg Lukianoff, we
love each other, but we disagree over age verification. And
(26:40):
Greg is afraid, understandably, he's afraid of what he calls
a papers-please internet. Oh, you want to go to Wikipedia?
Show me your driver's license. Oh, you want to look
at porn? Show me your passports. And I agree. Greg
is the foremost free speech proponent, so I can't say
I am the most, but we wrote The Collinear American
Mind together. I care a lot about free speech on campus,
intellectual inquiry. So I want this to be privacy preserving.
(27:04):
I don't want to put a chill on free speech,
but we have to change the status quo. The status
quo is completely destructive of a generation. Actually, too, we're
now up to Gen Alpha. Gen Alpha is currently being
damaged as well.
Speaker 1 (27:16):
Here's what the electronic Frontier Foundation has said.
Speaker 2 (27:20):
Oh, my favorite. My favorite organization.
Speaker 1 (27:23):
This applies to the Meta settlement, but also, you know,
I think applies by extension to the Australia ban. One,
the settlement embeds age assurance technology and age gates into
Meta's social media products and requires all users, minors, and
adults to undergo a rights-threatening age estimation process. We've already
addressed that one. You said there are technologies emerging.
Speaker 3 (27:42):
Yeah.
Speaker 2 (27:43):
Nobody has written into the law that you have to
use age... In fact, the face... It's creepy to me too,
but it was the first one because it's very easy
for them to do. The Australia bill was specifically written
to say, it's up to you companies how you do it.
The only rule is you can't only ask for a
government ID. You have to have an alternative for people
(28:03):
who don't want to show a driver's license or something
like that. And so that alternative, the easiest to implement
was the face scan. That's why a lot of them
did it. But EFF is just plain wrong, just completely wrong.
to say this forces you to do an age scan.
That's ridiculous.
Speaker 1 (28:19):
The settlement places severe restrictions on teens that can largely
only be modified by the teen's parents and only then
in exchange for giving their parents a ton of information
about their online community and usage.
Speaker 2 (28:31):
The issue here is because the settlement involves parents linked
to parents' accounts or something, I think.
Speaker 1 (28:39):
The criticism here is that it takes away children's rights
to have a private life online. And if they may be,
for example, LGBTQ and their parents don't know or don't
approve of it, that this would limit their ability to
find community online.
Speaker 3 (28:51):
Yeah.
Speaker 2 (28:52):
So Meta uses one set of arguments to inflame the right,
and that's about free speech and privacy. There's a different
set of arguments to inflame the left, and that's LGBTQ.
And usually, a word you'll hear is lifeline. I don't
know if EFF used the word, but if you use
the word lifeline, you know this is the meta-talking point.
And my response to that is, you know what was
(29:13):
a lifeline for LGBTQ? The internet. If you were a
gay kid growing up in a rural area, before the internet,
it was horrible. You were isolated. When the internet came
in the 90s, gay people found LGBTQ, they found incredible
ways to connect, to talk, to get information. So I
would never say... oh, you can't get on the internet
until you're 18. I would never say that. And then
(29:36):
the question is, okay, so they're on the internet. It
was glorious. We all loved the internet in the 90s.
Then it got enshitified. And part of the enshitification was
the algorithms and the advertising-based model. So what's the experience
of LGBTQ kids before Facebook versus after 2012? Entirely different.
I have an article in The Atlantic with Lennon Torres,
(29:58):
a trans woman, who talks about the horrible things that
were done to her when she was 12, 13, 14,
as she was discovering who she was and later transitioning.
I mean, you're dealing with LGBTQ people are far more
likely than anyone else to be sexually victimized on social media.
We give statistics on that. They're the most likely to
(30:19):
say that it harmed their emotional health. So the claim
that social media is a lifeline for LGBTQ is frankly bullshit.
It is just complete bullshit. The internet is a lifeline,
but Instagram? Now, of course, some kids have good experience.
I'm not saying this is for everyone. But if we
look at the harms, the most victimized group is LGBTQ.
(30:41):
And so my argument is, do they really need the
algorithm from a company that we know is aiming to
addict them? Would you trust this sleazy company that has
shown to be liars? Would you trust them to take
care of LGBTQ kids? So, I mean, sure, there are
privacy issues if you are going to require parents to
do something. But my view is kids shouldn't be on
(31:02):
these things. This is an adult activity. And so that's
why I don't tend to talk. I don't tend to
want parental permission laws. Keep it simple. Minimum age 16,
below 16, you're just not on. That's my preferred solution.
Then there are no privacy issues.
Speaker 1 (31:16):
So you've batted down the Electronic Frontier Foundation's criticisms. Final question, Jonathan,
for you is coming back to correlation and causation. What
studies can you design? Will you design? Are you planning?
Are in the works? So that on the other side
of the Australia ban, the other side of the meta-settlement,
you will be able to point to convincing and hard
(31:38):
data saying, this is for sure mitigating the harms that
I outlined in the anxious generation.
Speaker 2 (31:44):
So two things. One is, am I going to do
more am I going to collect new kinds of evidence
or new studies to show that it's causation, not correlation? No,
I'm not. Because what else is there? If critics, if
you know of, we've already gathered seven kinds of evidence,
lots and lots of examples of each, meta-analyses, review papers.
So if you know of another kind of evidence, tell
(32:06):
me and we'll get it. But I don't know of
any other kind of evidence. Like we've got seven already.
So I don't plan to do any more to prove
the causation issue. I put out my work. When papers
come out, claiming to show no effect, my team and I,
we will still go in and we'll find out, okay,
here's why they got no result. But if we correct
for that, we get an effect. So we'll still do that.
(32:27):
But here's the exciting thing. My team is, we're working
with Jeff Hancock's group at Stanford and we're collaborating on
ways to measure social media harms. That is, until now,
all we've had is like big national surveys that ask
like two questions about mental health. And we try to
track that. It's very, very rough data. And we're working
(32:51):
together to design a survey that could be given, let's say,
every month to thousands and thousands of teenagers. And here's
the key thing. It will ask, what platforms are you using? And, oh,
you're on Snapchat? Okay, did this happen to you on Snapchat?
Right now, we don't know any of that. All we
know is I was on social media for three hours today.
(33:13):
We don't know which platform and they all have different effects.
So what we get right currently is mush. The signal
is so degraded in the data we can collect now.
But what we're proposing here is a way to measure
many different aspects of outcomes and very specifically which platforms.
This is crucial as countries put in these new laws.
(33:33):
We want to know which platforms are still harmful and
then they can be held accountable. And which platforms have
actually cleaned up their act so that they're not particularly harmful?
And this is especially important because here's the third transformative
effect of Australia. It's that many other countries are copying Australia,
but they're improving on it. Like Indonesia, they're saying, we're
(33:54):
not just going to ban 10 platforms the way Australia did.
Indonesia and Canada and a few other countries are saying,
let's look at the features. If you have this feature,
if you're talking with anonymous strangers, that's bad. Why do
you need to do that? If you have addictive design,
that's bad. Why do you need to do that? And
so what's going to happen in all these other countries
is some platforms will get caught up in this because
(34:14):
they have dangerous features. Other platforms will say, well, how
about if we don't have you talking with anonymous strangers?
So it's incentivizing companies to actually make their platforms safer.
And so if we have large scale surveys of outcomes,
harmful outcomes for all the platforms we care about, then
we can track for the first time in history, we
(34:36):
can track whether the changes are actually benefiting kids or not.
Speaker 1 (34:41):
Jonathan Haidt, thank you.
Speaker 2 (34:42):
Paz, thanks for giving me the chance to talk about
these issues.
Speaker 1 (35:01):
For Tech Stuff, I'm Oz Veloshin. This episode was produced
by Eliza Dennis and Tyler Hill. It was executive produced
by me and Julian Nutter for Kaleidoscope and Katrina Norvell
for iHeart Podcasts. Jack Inslee mixed this episode and Kyle
Murdoch wrote our theme song. A special thank you to
all our listeners. Please rate, review and reach out to
us at techstuffpodcasts at gmail.com. We love hearing from you.
Welcome to Tech Stuff. I'm Oz Veloshin, joined today by
Jonathan Haidt, one of the most prominent and arguably most
effective tech critics in the world. He's the author of
The Anxious Generation, a book that proposes new norms to
address the growing teen mental health crisis. In late August,
the New York Times ran a story with the headline,
A $ 17. 1 Billion Vote of Confidence That Jonathan Haidt Was Right.
(00:42):
The story describes Meta's landmark settlement with almost all US
states over claims its technology has harmed children. So we're
going to talk about that today. We're also going to
talk about the efficacy of social media bans more broadly.
In December 2025, Australia became the first country to ban
social media use for users under 16. It's been about
nine months since that ban went into effect. And a
(01:05):
lot of early assessments of the ban are focused on
the fact that many kids are finding workarounds and still
using social media. Some are even calling the bans ineffective. Jonathan,
I believe, begs to differ.
Speaker 2 (01:16):
Jonathan, welcome. Thanks so much, Oz. Great to be with you.
So let's start with.
Speaker 1 (01:19):
The META settlement. $ 17. 1 billion over allegations that Instagram and
Facebook were intentionally designed to addict children and harm teen
mental health. How did you first learn about this settlement
and what was your reaction?
Speaker 2 (01:32):
Well, of course, we'd known that the case was coming
for a long time. There are literally thousands of cases
against META. This case was the biggest of them all
because this was all of the attorneys general representing the
chief justice official of every state or almost all the states.
And so when I heard the news last Wednesday and
I saw the number, 17 billion, I was disappointed at first.
(01:55):
This is very much like the tobacco settlement. The tobacco companies,
they had a dangerous product. They knew it was dangerous.
They did internal research. They buried the research. They didn't
let their kids smoke. All of that is true for Meta.
All of it is true for the social media companies.
And that settlement in the 90s, I think, was around
$ 200 billion then. So that would be more like $ 400
billion in today's dollars. So I was actually disappointed in
(02:16):
the number because, really, they have to pay $ 12 billion
over 10 years. They're not going to notice that.
Speaker 1 (02:21):
De minimis for them.
Speaker 2 (02:22):
So I was disappointed. I was disappointed in them. But
I was surprised by the changes that they agreed to
make because they were pretty substantial. They weren't just cosmetic things.
It was, if you're under 18, you can't go on
between midnight and 6. So that's the biggest one. And
the heavy users are the ones who are most damaged.
And they're on it two, three, four in the morning.
(02:43):
So as I was reflecting on that, what I realized,
it took me a few days to realize this, is
that the rest of the world is going to do this.
That is, dozens of countries want to follow Australia and
are at various stages. And there are bills being drafted
right now. And the EU is committing to doing it.
So the world is going to follow Australia in various
forms and with innovations that we'll talk about. And it
(03:04):
suddenly dawned on me, like Thursday or Friday, like, wait
a second. all these countries that are going to do
the age limit thing, they're also going to require all
the social media companies, not just Meta, to have off periods.
So I realized this is like the biggest event we've
ever seen in terms of trying to limit these companies'
damage to kids. So I'm actually thrilled by what the
(03:26):
attorneys general achieved.
Speaker 1 (03:28):
So the dollar amount is more or less irrelevant, but
the framework created by this agreement is something that you
think could be replicable globally.
Speaker 2 (03:38):
Yes, absolutely. Of course, the money matters because these companies
only respond to financial incentives. They have not responded to
any kind of moral appeals. They've not responded to their
own findings of harm. So the money does matter. Now, fortunately,
there are still literally thousands of lawsuits pending, thousands and thousands.
So I am hoping that ultimately, Med will have to
pay out hundreds of billions of dollars.
Speaker 1 (03:58):
What were you happiest that was in this settlement? And
what do you think is missing to truly address the
harms that you've covered and researched.
Speaker 2 (04:08):
So last week, a lot of the talk was about like, well,
they're not doing anything about the central harm, which is
the algorithm, the central addictive property. And some people said like,
it's like saying you can only smoke 10 cigarettes a day,
but you haven't done anything to make cigarettes less harmful.
So that's, I think that's a fair description of it.
But again, by the end of the week, I was like,
(04:30):
wait a second. Like if the whole world were to
say that kids can only smoke 10 cigarettes a day,
You know, when I was a kid, a lot of
kids smoked a pack a day in high school. You know,
most kids aren't smoking a pack a day, but some
kids are smoking five packs a day. And now they
can only smoke a half a pack a day. So,
of course, more needs to be done. But, you know,
a journey of a thousand miles begins with a single step.
(04:51):
And this was a gigantic step.
Speaker 1 (04:54):
And what about this New York Times headline about the
$ 17. 1 billion vote of confidence that you were right? I mean,
what do you make of that?
Speaker 2 (05:02):
It's probably the favorite headline I've ever seen in my
entire life. I was very pleased. Because there has been
a debate. There's an academic debate. So when the book
came out in 2024, it was astonishing how little criticism
there was. I was expecting a lot. Nobody criticized anything
about the play side, the overprotection. Not a single person
(05:24):
has ever criticized that. And the academic debate has been, well,
is there evidence of causality? Is it correlation? Is it causation?
That's what the main academic debate has been. I hope
we'll talk about that because I think it is now
decisively answered. There are so many different kinds of evidence.
One of those kinds of evidence was Meta's internal research.
(05:45):
And that's what these cases hinged on. What they all
turned on was not so much the academic research. It
really was the direct statements made by their employees. And
so here, I'll read you a few. Meta did hire
a lot of social psychologists and clinical psychologists. And so
here's a conversation between a couple of them. One says, quote,
oh my gosh, y'all, Instagram is a drug. We're basically pushers.
(06:08):
We're causing reward deficit disorder because people are binging on
Instagram so much they can't feel reward anymore. Like the
reward tolerance is so high. Then he says, I know
Adam Mosseri, the head of Instagram. I know Adam doesn't
want to hear it. He freaked out when I talked
about dopamine and my teen fundamentals, but it's undeniable And
there's a lot more. It seems clear from what's presented
(06:30):
here that some of our users are addicted to our products.
I mean, these guys, they knew the science of what
they were doing to kids. If you design a product
to be maximally addictive, that's not free speech. That's not
Section 230. That is a defective product. And there are
just so many ways that Meta just didn't... I mean,
they felt like, we don't have to worry about the law.
And so they got nailed for a lot of things.
(06:51):
So COPPA, the Children's Online Privacy Protection Act, says... how
old do you have to be before a company can
take your data without your parents' knowledge or permission? And Congress,
in its wisdom, said 13. It should have been 16.
The original bill was 16, but it got lowered in
negotiations in 1998, I think it was, to 13. Okay,
all right, it's 13 for the world. Well, all these companies,
(07:12):
especially Meta, they're going after the eight, nine, 10-year-olds, and
they have them, and they know how old they are.
And we just accepted that. Like, oh, you know, okay,
I guess they can just take data from anyone No
need to tell the parents. Like that's insane that this
is just what we accepted and that's over. So they
got nailed for taking data from children without the parents' knowledge.
So that's incredibly illegal and they were nailed for that.
(07:33):
They got nailed for making a defective product if you
have a defective consumer product and it's not about the content,
it's about the design and you know it and you've
known it for years and you have all your researchers
talking about them. Like this is insane that this went
on for so long. So this is all gonna change.
Like the world is realizing this is completely insane you
and I will talk about the Australia law and age verification,
(07:56):
but I just want to establish the status quo is
not an option. Like this has to change. It's just
a question of how we change it. It is not
an option to just say, leave it as it is.
Speaker 1 (08:06):
Talk about the publication of The Anxious Generation, because I
refer to as one of the most prominent and arguably
most effective tech critics in the world. And prominent is
one thing, but effective is another. What is it about
you and your work? I mean, as compared to Let's
just say, for example, send a few main technology, Tristan Harris,
you know, great guy. He's been on the podcast, et cetera.
But what is it about Jonathan Haidt and you and
(08:29):
your message that has essentially catalyzed this global conversation?
Speaker 2 (08:35):
One word, mothers. It's mothers. That's it. If there is
a raging forest fire bigger than anything we've ever seen,
nobody is saying, I wonder how big the match was.
Let's find out. What kind of a match could start
such a forest fire? It's not about the match. It's
about the dryness of the forest. And the world in
(08:57):
March of 2024, when The Anxious Generation came out, was
a world full of families. In every family, the kids
had electronics. In every family, if the kid was 9, 10,
11 years old, he had an iPad or an iPhone.
In every family, it was constant fighting over that. In
every family, the parents, especially the mothers, the mothers are
(09:18):
much more sensitive to their wonderful, playful, happy kid, which
we all had when they were little. And then you
give them a device and they get glued to it
and then they go down. And mothers really felt that.
And they were desperate. They knew something was wrong, but
they were like, oh, you know, it's just correlation. Like,
oh no, this is how kids connect today. Like, oh, there's,
you know, pluses and minuses. So there was a lot
(09:38):
of confusion about what to do. And so when the
anxious generation came out, you know, I knew that it
would be popular. I knew that people would buy it,
but The thing that I think I really did contribute
is as a social psychologist, I was able to see
the collective action problem. The other researchers were not social psychologists,
they were developmental psychologists. They were focused on development. They
(09:59):
didn't fully appreciate it. that each kid has to be
on because the other kids are on. And so I
laid out a series of collective action problems. And I said,
the way out of a collective action problem is to
act collectively. So Talk to the parents of your kids'
friends and agree. You're all going to do this. You're
all going to delay smartphones. You're all going to delay
social media. You're going to go to your school. You're
(10:19):
going to say, let's have phone-free schools. You're going to
all send your kids out to play together. So those
were the four norms right there. And because I suggested
four specific behavioral norms that are easy to do, they
cost zero dollars. They're free. You can do them today.
And because the mothers of the world were at their
wits end, they were desperate. So as soon as the
book came out proposing this, boom, they jumped into action.
(10:41):
And that's why it happened all over the world. Brazil's
a great example. I didn't do anything in Brazil for
the first year, but some others there got a copy
of the book in English before it was translated to Portuguese,
and they jumped into action. They got Brazil banned phones
in school by November of 2024. And the book only
came out in Portuguese in July. So it really was
(11:01):
the mothers.
Speaker 1 (11:03):
So, I mean, we've skirted around this controversy element in
terms of correlation and causation. Let's address it head on.
There was a a paper in Nature by Candice Hodges,
who leveled a very trenchant critique against your work, saying
that you had fundamentally misunderstood correlation and causation. And that
critique caught on, and other people began repeating it. Describe
(11:25):
that moment, and where is the argument today? I mean,
how have you defended your work?
Speaker 2 (11:30):
All right, sure. So two things about that critique. So
it came out in Nature a few weeks after the
book came out. And you're right to use the word repeat.
That critique has been repeated. But that's really the only one.
Everybody cites that one article. There are one or two
other small things, but that's the one that everybody cites.
But if you read that article, and I encourage people
to read it, just, you know, put in Candace Auger's nature,
(11:53):
nothing in what she wrote suggests that she read past
chapter one. Chapter one, I lay out the correlational evidence.
Chapter one, I let the historical correlation. Just, you know,
here are all these graphs, look what happened. That's chapter one.
And she says, she accuses me of not understanding the
difference between correlation and causation. Okay. Throughout the book, whenever
(12:14):
I present findings, I always say, so here's the correlation,
but correlation doesn't prove causation. Let's look at the experiment.
So I'm incredibly careful about this in the book. And
then when we get to chapter six, which is the
chapter on girls, it starts off with a long section.
The evidence that social media harms girls and includes the
section is social media cause or just a correlate. So
(12:37):
I've been as careful and as explicit as could be
in all the Google Docs since 2020, in the book,
in everything I write. So it was just unfair. I
think it was just really unfair of her to say
that I've confused the two or that I wasn't careful
about it. At the time, there were multiple kinds of evidence.
But what I've done since then, what my team, the
Tech and Society Lab at NYU Stern, what we've done
(12:57):
is we've aggregated the evidence into a major review paper
that is published in the World Happiness Report 2026. So
if you've just type in World Happiness Report, and then
my name, Haidt, H-A-I-D-T, you'll find it. And we lay
out these seven lines of evidence as if we were
prosecutors in a court case. And so the correlational studies,
(13:19):
those don't prove causation, but the experiments kind of do.
I mean, they push you in a Bayesian sense, much
more confidence in causation. The longitudinal studies mostly point that way.
And the natural experiments point that way. So there's four
bodies of academic research that all point that way. It's
a great majority of the studies and the meta-analyses point
(13:40):
towards causation. So that's the academic evidence. But then we say, well,
if a woman accuses a man of having assaulted her,
attacked her, punched her, and she says, he did it.
I know the guy, he did it. Is that evidence?
Of course it is. In a court, that's evidence. It's
not a correlation. She was there. She knows the guy
(14:00):
and she points to him. Well, that's what Gen Z says.
When you survey Gen Z around the world, there's very
little gratitude to social media. The majorities generally say that
it has hurt their generation. And between 20% to 40%, 20%
to 30%, depending on the question, say it hurt me,
like girls especially. It hurt my mental health. It damaged
my sleep. It damaged my body image. So if the
(14:20):
victim says he did it, that's the guy, it was
Instagram and TikTok and all those So that's evidence. And
then if you have eyewitnesses, people who saw like, yeah,
I saw him punch her. Like, is that evidence? Is
that correlation or causation? That's evidence. And we have that.
So you survey parents and teachers and therapists. And what's
(14:42):
so interesting to me is that missing from this whole
debate are parents defending social media or teachers or therapists
or Gen Z. You don't have organizations of young people saying, No, no,
don't raise the age. We love it. It's so good
for us. You don't have that. And so I think
the silence of the defenders of social media is deafening.
(15:04):
And then finally, the strongest kind of evidence you could
have is if you had emails and text messages from
the perpetrator saying, I'm going to mug this girl tomorrow
at three o'clock to his friends. And then he has
photographs of it. And then afterwards he says, yeah, I
mugged her at three o'clock yesterday. And if the prosecutor
has those and he presents it to the jury, is
that just a correlation? No, that is... proof positive of causation.
(15:27):
And we have that. That's the internal document. So with
seven different lines of evidence, including meta itself saying that
they're hurting kids, I don't know what's going on. I
don't know why some people still say, oh, there's no evidence.
How could you possibly tell correlation from causation?
Speaker 1 (15:43):
So what is your academic work today? Do you feel
like it's case closed or are you still pursuing the
smoking gun on the causation versus correlation?
Speaker 2 (15:53):
Yeah. So in science, it's never case closed because we
can't actually answer a question directly. We have to first
operationalize it. And so we have to first say, okay,
social media, we can't really measure it all. Let's just
take a proxy. How many hours a day do you
spend online? That's a terrible proxy, but that's what we use.
(16:16):
And then on the output side, we don't really know
what's in people's head, but we're going to ask them
a question on a scale of one to five, how
anxious are they? So So we have bad measures at
the beginning, bad measures at the end. And all we're
really looking at is time spent on the device. We're
not literally studying what's happening. We're just like, almost like
feeling into a box where you can't really see what's
going on, but you kind of get a sense from
(16:36):
your hands, but you could be wrong. So I would
never say case closed. What I would say is given
the way that we are now conceptualizing it, I think
the evidence is overwhelming in particular because it's from so
many different sources. I always think about this as a
Bayesian sort of situation, like Bayes' theorem, like certain things happen.
Should they cause you to adjust your belief in causality
(16:58):
up or down? And everything should cause people's belief to
go up. And the clincher for me is this. The critics,
they do point to studies that find no effect. So
of course, I mean, of course they have evidence. They're
not making this up. So they point, here's a study
that didn't find an effect. Here's a meta-analysis that didn't
find an effect. So my team, a lot of our
publications are, we go into those and what we always
(17:19):
find is that the studies that fail to find a
significant correlation or experimental effect, they blended lots of things together.
So the classic one was a study done in 2019
that showed that the correlation is about as large as
the correlation of mental illness with eating potatoes. It's so
close to zero. But in that study, what they did
(17:40):
was they mixed together boys and girls, they mixed together
social media use with everything else, including watching television. So
just like you look at everything and on the output side,
they put in everything and they got no real relationship.
But what we did, we took the same data sets
and we said, let's just look at girls and let's
(18:01):
just look at social media, not everything. And let's just
look at the questions about anxiety and depression, not everything.
And the effect size goes from a betas of around 0.04,
which truly is tiny. two betas of around 0.2, which
is quite large. I mean, you don't get, because of
the crappy measurement on both ends, you don't get like
correlations of 0.5. You don't get, you know, effect sizes of,
(18:24):
you know, of a full standard deviation or a half
standard deviation. So given the general range of effects, what
we always find is you show me a study that
failed to find an effect, I'll unblend it. We'll find, oh, look,
if we look at girls in social media, we find it.
Speaker 3 (18:42):
So talk about the australia ban and i mean this
year for you.
Speaker 2 (19:08):
In general?
Speaker 1 (19:08):
I mean, was it meant for you to see your
work in the world? And what do you hope for
from the next five years?
Speaker 2 (19:15):
So it's been thrilling to see the world going through
a tipping point. And here I draw on Steve Pinker
has a wonderful book called When Everyone Knows That Everyone Knows.
And it's about what happens in a society where people
might know something beforehand, like everybody knew that communism didn't work.
Everybody knew that the government in these countries was corrupt.
(19:36):
But if you spoke up, you'd be killed or abducted.
So nobody spoke up. But then what happens when somebody
speaks up, then suddenly everybody knows that everybody knows that
everybody knows that this is terrible. And that's where we
were with social media. Everybody could see this is really
hurting kids. But they didn't necessarily know that everybody knew
(19:57):
and that there was something they could do about it.
And so what happened when Australia's law went into effect
was December 10th. What happened when it went into effect
was there was a lot of coverage around the world,
a lot of commentary on social media. And a lot
of it was people saying, why can't we do that
in our country? Or, hey, what a great idea, let's
do it in our country. And when everybody saw that
everybody wanted to do it, that's when political leaders saw,
(20:21):
wait a second, this is not some radical thing that's
going to cause a blowback among my people. This is
what adults in my country want. And polling always shows that.
Large majorities of adults from every political party, this is
not a left-right issue. It's one of the only issues,
not left-right. If you're a parent, you hate social media,
you want your kids protected. And so Australia, God bless them,
(20:42):
they just did it straightforwardly, simply. The premier of South Australia,
Peter Malinowskis, his wife read The Anxious Generation, and she
turns to him, she's reading it in bed, she turns
to him one night and says, Peter, you've got to
read this book and you've got to fucking do something
about it. That is a direct quote. I don't know if, well, anyway,
he's used the, he sometimes says, bloody well do something
about it. So he contacted a former chief justice of
(21:05):
the Supreme Court and said, What's the legal mechanism by
which we could do this? And the clear answer is
the signing of the contract. Whenever you open a social
media account, you sign a contract, terms of service, which
you say, yep, you can take everything from me, everything
I post, every photo, you own it, you take it,
all my private, you take it. And that's insane that children,
(21:27):
that a nine-year-old can say she's 13 and then can
sign that contract. That's insane. So the Australia bill was
basically saying, you can't sign that contract until you're 16.
And you know who has the responsibility for enforcing it?
Not the parents. It's the company that's making the goddamn money.
They're the ones who have to enforce it. So that's
what's historic about Australia. The companies have never borne any responsibility.
(21:49):
They could make whatever crappy consumer product they wanted, and
they got off scot-free. And Australia said, no, no, no.
You actually have to enforce the age limit.
Speaker 1 (21:58):
When we first started talking about having this conversation, it
was before the meta-settlement and the thing that we agreed
to talk about was this Australia ban, and in particular,
your sense that a lot of the critics of it
have missed the point. And so I'm curious, what are
the most trenchant criticisms and why are they wrong? Can
they be wrong and trenchant?
Speaker 2 (22:17):
So the main criticism is that the rates of social
media use are still very high. They went down somewhat.
There's some studies that show it may have even come
back a lot. So Australia passed a law The use
of social media went down, but it wasn't like down 90%.
It was down, we have an article in Time Magazine.
We review five different studies that looked at Australia's use.
(22:39):
And the use of various accounts goes down. And so
people say, oh, I guess it failed. Most kids are
still on, must have failed. And my response to that
is the Australians were the first in the world to
do this. So they had a going kind of soft.
What they did is they said, Okay, companies, you've never
done anything to verify ages. None of you have ever
done anything to verify ages. You have to start. And
(23:01):
they did. They all started. And a lot of them
used YOTI, which is you do this face scan to
show that you're over a certain age. And it's easy
to get around that. But they did something. All of
them did something. And that's historic. Okay, so that's round one.
This is what the critics don't understand. Of course, they
expected that the kids would get around it. It's still
quite easy to get around it. That's round one. Now,
(23:23):
the regulator, the E-Safety Commission, Now they have the data showing, hey,
you know, hey, meta, kids are still on your platform.
Kids under 16, kids under 13 are still on your platform.
You have to do more. And they sent those letters
out recently, I forget, in the last month or so.
And now the companies are told you have to now
do better on the second round. And then there will
be a third round. So that's the first thing critics
(23:44):
don't understand, that there are multiple rounds and round one
was a success. It didn't get the rate way down,
but it went through. The sky didn't fall. All the
companies complied. They didn't comply well. Okay, that's round one.
Second thing the critics don't understand is that the law
is not written to keep kids off social media. It's
written to keep them from opening accounts. If you have
(24:05):
an account, you can post, your life gets taken over
by what you post, you're talking with strange men who
want sex or money from you, all that stuff happens
when you have an account. But they can go on
YouTube as much as they want. They can go on YouTube,
they can search for how do I kill myself, and
they'll get videos telling them how to kill themselves. But
what they won't get is an algorithm that knows them
over time and that feeds them videos about how to
(24:28):
kill yourself. So when you survey, have you been on
YouTube or Snap or TikTok? The answer is yeah. They've
almost all been on these platforms since the bill went
into effect. On YouTube, there's no restriction. They can watch
as much as they want. TikTok and Instagram, they restrict.
You can't just keep scrolling if you don't have an account.
But that's the second misunderstanding is that it's about accounts,
not about access. And the third misunderstanding is is that
(24:53):
we were stuck in a collective action problem where because
age verification isn't great and easy, people say, oh, well,
therefore we shouldn't do it. But how does age verification
get great or easy? When companies are incentivized to improve it.
And now that it's clear, the world is going to
do this. Like this is not a question of whether.
This is not a question of if. This is just
(25:14):
a question of how. The world is going to age
gate social media. Just as it age gates gambling and
alcohol sales, it's going to age gate social media. And now,
all of a sudden, Apple and Google are incentivized to
make it easy. And they have, just since the Australia
Bill was passed. Apple and Google have both made changes
to their wallet so that if you have a driver's license,
(25:35):
a passport, anything, and you put it in your wallet,
now Google and Apple can vouch for you. Which means
if you want to open an account, you don't have
to show your ID. You don't need to give your name.
You don't need to give your address. You don't need
to give your age or your birthdate. The platform just says,
is this person over 16? Yes or no. One bit
(25:56):
of information. Yes or no. That's it. That's as privacy
preserving as can be. Porn sites want to do this.
They want to use this. And I think they're right.
If you want to go to a porn site, if
you use the device-based verification, you don't have to give
your address. You don't have to give your name. You
don't have to give your birthday. If you go to
the site, the site will say, is this person over 18?
(26:18):
And if the answer is yes, then you can get in.
So the point is, what Australia did is it kicked
off a new market, which is a market for quick
privacy-preserving age verification, which is not, it does not require
you to give away your identity. It does not. And
that's a major misunderstanding. My close friend, Greg Lukianoff, we
love each other, but we disagree over age verification. And
(26:40):
Greg is afraid, understandably, he's afraid of what he calls
a papers-please internet. Oh, you want to go to Wikipedia?
Show me your driver's license. Oh, you want to look
at porn? Show me your passports. And I agree. Greg
is the foremost free speech proponent, so I can't say
I am the most, but we wrote The Collinear American
Mind together. I care a lot about free speech on campus,
intellectual inquiry. So I want this to be privacy preserving.
(27:04):
I don't want to put a chill on free speech,
but we have to change the status quo. The status
quo is completely destructive of a generation. Actually, too, we're
now up to Gen Alpha. Gen Alpha is currently being
damaged as well.
Speaker 1 (27:16):
Here's what the electronic Frontier Foundation has said.
Speaker 2 (27:20):
Oh, my favorite. My favorite organization.
Speaker 1 (27:23):
This applies to the Meta settlement, but also, you know,
I think applies by extension to the Australia ban. One,
the settlement embeds age assurance technology and age gates into
Meta's social media products and requires all users, minors, and
adults to undergo a rights-threatening age estimation process. We've already
addressed that one. You said there are technologies emerging.
Speaker 3 (27:42):
Yeah.
Speaker 2 (27:43):
Nobody has written into the law that you have to
use age... In fact, the face... It's creepy to me too,
but it was the first one because it's very easy
for them to do. The Australia bill was specifically written
to say, it's up to you companies how you do it.
The only rule is you can't only ask for a
government ID. You have to have an alternative for people
(28:03):
who don't want to show a driver's license or something
like that. And so that alternative, the easiest to implement
was the face scan. That's why a lot of them
did it. But EFF is just plain wrong, just completely wrong.
to say this forces you to do an age scan.
That's ridiculous.
Speaker 1 (28:19):
The settlement places severe restrictions on teens that can largely
only be modified by the teen's parents and only then
in exchange for giving their parents a ton of information
about their online community and usage.
Speaker 2 (28:31):
The issue here is because the settlement involves parents linked
to parents' accounts or something, I think.
Speaker 1 (28:39):
The criticism here is that it takes away children's rights
to have a private life online. And if they may be,
for example, LGBTQ and their parents don't know or don't
approve of it, that this would limit their ability to
find community online.
Speaker 3 (28:51):
Yeah.
Speaker 2 (28:52):
So Meta uses one set of arguments to inflame the right,
and that's about free speech and privacy. There's a different
set of arguments to inflame the left, and that's LGBTQ.
And usually, a word you'll hear is lifeline. I don't
know if EFF used the word, but if you use
the word lifeline, you know this is the meta-talking point.
And my response to that is, you know what was
(29:13):
a lifeline for LGBTQ? The internet. If you were a
gay kid growing up in a rural area, before the internet,
it was horrible. You were isolated. When the internet came
in the 90s, gay people found LGBTQ, they found incredible
ways to connect, to talk, to get information. So I
would never say... oh, you can't get on the internet
until you're 18. I would never say that. And then
(29:36):
the question is, okay, so they're on the internet. It
was glorious. We all loved the internet in the 90s.
Then it got enshitified. And part of the enshitification was
the algorithms and the advertising-based model. So what's the experience
of LGBTQ kids before Facebook versus after 2012? Entirely different.
I have an article in The Atlantic with Lennon Torres,
(29:58):
a trans woman, who talks about the horrible things that
were done to her when she was 12, 13, 14,
as she was discovering who she was and later transitioning.
I mean, you're dealing with LGBTQ people are far more
likely than anyone else to be sexually victimized on social media.
We give statistics on that. They're the most likely to
(30:19):
say that it harmed their emotional health. So the claim
that social media is a lifeline for LGBTQ is frankly bullshit.
It is just complete bullshit. The internet is a lifeline,
but Instagram? Now, of course, some kids have good experience.
I'm not saying this is for everyone. But if we
look at the harms, the most victimized group is LGBTQ.
(30:41):
And so my argument is, do they really need the
algorithm from a company that we know is aiming to
addict them? Would you trust this sleazy company that has
shown to be liars? Would you trust them to take
care of LGBTQ kids? So, I mean, sure, there are
privacy issues if you are going to require parents to
do something. But my view is kids shouldn't be on
(31:02):
these things. This is an adult activity. And so that's
why I don't tend to talk. I don't tend to
want parental permission laws. Keep it simple. Minimum age 16,
below 16, you're just not on. That's my preferred solution.
Then there are no privacy issues.
Speaker 1 (31:16):
So you've batted down the Electronic Frontier Foundation's criticisms. Final question, Jonathan,
for you is coming back to correlation and causation. What
studies can you design? Will you design? Are you planning?
Are in the works? So that on the other side
of the Australia ban, the other side of the meta-settlement,
you will be able to point to convincing and hard
(31:38):
data saying, this is for sure mitigating the harms that
I outlined in the anxious generation.
Speaker 2 (31:44):
So two things. One is, am I going to do
more am I going to collect new kinds of evidence
or new studies to show that it's causation, not correlation? No,
I'm not. Because what else is there? If critics, if
you know of, we've already gathered seven kinds of evidence,
lots and lots of examples of each, meta-analyses, review papers.
So if you know of another kind of evidence, tell
(32:06):
me and we'll get it. But I don't know of
any other kind of evidence. Like we've got seven already.
So I don't plan to do any more to prove
the causation issue. I put out my work. When papers
come out, claiming to show no effect, my team and I,
we will still go in and we'll find out, okay,
here's why they got no result. But if we correct
for that, we get an effect. So we'll still do that.
(32:27):
But here's the exciting thing. My team is, we're working
with Jeff Hancock's group at Stanford and we're collaborating on
ways to measure social media harms. That is, until now,
all we've had is like big national surveys that ask
like two questions about mental health. And we try to
track that. It's very, very rough data. And we're working
(32:51):
together to design a survey that could be given, let's say,
every month to thousands and thousands of teenagers. And here's
the key thing. It will ask, what platforms are you using? And, oh,
you're on Snapchat? Okay, did this happen to you on Snapchat?
Right now, we don't know any of that. All we
know is I was on social media for three hours today.
(33:13):
We don't know which platform and they all have different effects.
So what we get right currently is mush. The signal
is so degraded in the data we can collect now.
But what we're proposing here is a way to measure
many different aspects of outcomes and very specifically which platforms.
This is crucial as countries put in these new laws.
(33:33):
We want to know which platforms are still harmful and
then they can be held accountable. And which platforms have
actually cleaned up their act so that they're not particularly harmful?
And this is especially important because here's the third transformative
effect of Australia. It's that many other countries are copying Australia,
but they're improving on it. Like Indonesia, they're saying, we're
(33:54):
not just going to ban 10 platforms the way Australia did.
Indonesia and Canada and a few other countries are saying,
let's look at the features. If you have this feature,
if you're talking with anonymous strangers, that's bad. Why do
you need to do that? If you have addictive design,
that's bad. Why do you need to do that? And
so what's going to happen in all these other countries
is some platforms will get caught up in this because
(34:14):
they have dangerous features. Other platforms will say, well, how
about if we don't have you talking with anonymous strangers?
So it's incentivizing companies to actually make their platforms safer.
And so if we have large scale surveys of outcomes,
harmful outcomes for all the platforms we care about, then
we can track for the first time in history, we
(34:36):
can track whether the changes are actually benefiting kids or not.
Speaker 1 (34:41):
Jonathan Haidt, thank you.
Speaker 2 (34:42):
Paz, thanks for giving me the chance to talk about
these issues.
Speaker 1 (35:01):
For Tech Stuff, I'm Oz Veloshin. This episode was produced
by Eliza Dennis and Tyler Hill. It was executive produced
by me and Julian Nutter for Kaleidoscope and Katrina Norvell
for iHeart Podcasts. Jack Inslee mixed this episode and Kyle
Murdoch wrote our theme song. A special thank you to
all our listeners. Please rate, review and reach out to
us at techstuffpodcasts at gmail.com. We love hearing from you.
TechStuff News
Hosts And Creators
Oz Woloshyn
Karah Preiss
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
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
The Joe Rogan Experience
The official podcast of comedian Joe Rogan.