Can We Save Kids from Social Media?
Listen and subscribe: Apple | Spotify | Wherever You ListenSign up for our daily newsletter to get the best of The New Yorker in your inbox.The poet and critic Randall Jarrell once said that a poet is someone who “manages, in a lifetime of standing out in thunderstorms, to be struck by lightning five or

Listen and subscribe: Apple | Spotify | Wherever You Listen
Sign up for our daily newsletter to get the best of The New Yorker in your inbox.
The poet and critic Randall Jarrell once said that a poet is someone who “manages, in a lifetime of standing out in thunderstorms, to be struck by lightning five or six times.” Jonathan Haidt, in his career as a wide-ranging social psychologist, has been flagrantly right about at least one big thing. His book “The Anxious Generation: How the Great Rewiring of Childhood Is Causing an Epidemic of Mental Illness” was a best-seller when it was published two years ago and its influence has hardly waned. As both a researcher and as a popularizer, Haidt argues that social-media companies have unleashed what is essentially a “mind-altering” danger in our lives—particularly in the lives of kids.
At first, Haidt’s work met with some critical eye-rolling, partly because a few in the field believed that he had not quite nailed one assertion or another, but more commonly because he seemed to some an alarmist, a techno-Luddite whinging about “kids these days and their devices.” But as the evidence of the harms accumulated—of social disconnection, of a sharp decline in mental health among young people—Haidt’s book became, for so many, essential.
I last met with Haidt for The New Yorker Radio Hour when his book was published. I wanted to catch up with him to review the political and social ramifications of his work from here to Australia and his own turn from scholar to activist. Our conversation has been edited for clarity and length.
Jonathan, I really wanted to have you back. We had a wonderful conversation a couple years ago, and you’ve done a lot of work since then. I’ve done a lot of thinking about it, too, I must admit, and I’m not alone. Your book has been cited as part of the inspiration for some new laws that are trying to shield children from social media. As we all know, there’s a major law passed in Australia, which we’ll get to. And a trial has just started, in the state of California, against social-media companies, and that seems really significant. Talk me through what’s going on in California.
That’s right. These companies were given near-blanket immunity [from liability] for their actions back in the nineties. Section 23src, the Communications Decency Act, said that we can’t sue Meta or TikTok because of what someone else posted on Meta or TikTok.
As a First Amendment idea?
It was actually done to incentivize the companies to moderate. The companies were afraid that if they take anything down, they become responsible for every single [piece of content]. So Congress said, “Go ahead and take down porn, and don’t worry—no one can sue you if you leave something up.” They wanted to give them more freedom of action.
It was a good idea originally, but the courts have interpreted it so widely. So you have all these parents with dead kids, and, in many cases, it’s just crystal clear. I mean, the kid got sextorted on Snapchat and was dead that night. That wasn’t a correlation. That was causation. You have a happy eleven-year-old girl, she gets on Instagram, and a few weeks later she’s developing an eating disorder.
So you have all these parents whose kids have been killed or damaged, and not one has ever gotten justice. Not one has ever even been able to face Meta in court. Meta has never faced a jury. None of these companies have ever faced a jury, because they keep saying, “Section 23src, you can’t touch us.”
Now, how insane is it that the makers of one of the largest consumer products in the world—that is the one that most children use, that seems to be harming and killing a lot of them—can never be held responsible for their actions?
Do you have any numbers for this?
We know from Snap that they were getting ten thousand reports of sextortion from their users in 2src22. And that wasn’t ten thousand a year, that was ten thousand a month. And, as they said themselves, this is probably the tip of the iceberg, because most people don’t report. And the kids—the boys who kill themselves—they don’t report, either. And with A.I. automating sextortion it’s going to go way up.
When we look at harms to mental health, we tend to find twenty to thirty per cent of the girls are saying, “It harmed my mental health.” The direct harms and the indirect harms are at such a scale that this could plausibly have caused the big increases, in 2src12, of mental illness.
So what’s happening in California?
The thousands of cases of parents who are suing can’t be combined into a class-action suit, because class-action suit requires that all the plaintiffs have been harmed in the same way. And in this case the stories are all a little different. So they’ve created what’s called multi-district litigation, in which several thousand cases will be heard by a single judge, a single court, in California.
Now, of course, that’s impossible. So the idea is the two sides argue about which cases to consider. They pick bellwether cases. Those cases go to trial in front of a jury. And then based on what those jury trials are it’ll kind of be clear which way everything has to go. So that’s where we are.
What’s the desirable outcome?
The desirable outcome is that a jury, which decides questions of fact, decides that social media is addictive and it was designed to maximize engagement. They use various tricks to, basically, addict kids.
When you say “tricks,” what do you mean?
Oh, you ever notice that on an iPhone, when you pull down, like, you want to check your e-mail, it kind of bounces up and you get new ones?
Yeah.
That was literally copied from slot machines. Literally.
I work with a lot of people in their twenties and thirties, and when I brought up your book a couple of years ago it seemed to some of them that it was rather censorious—yet another version of “kids these days with their loud rock-and-roll music,” or “it was all better before.” The same thing that you might’ve heard about television by people who had grown up on radio.
Yeah, the main argument that I get is, Oh, this is just another moral panic about whatever technology the kids are using. And as you and I talked about last time that’s a perfectly legitimate argument. I have to show why this time is different.
The big difference I’ve come to see is this. Screens have been around for a long time. Screens are good ways of presenting stories. Screens have a role in education. If you watch a movie with your kid, that’s great. A long story on a screen, across the room, that’s wonderful.
That’s not this.
That’s not this. The difference here is behaviorist conditioning. So, with the television, there’s no stimulus-response reinforcement loop. You watch, you’re entertained, you might get into the story. It’s a very pleasant state, to be into a great fictional story. That’s what art does—it takes us out of ourselves into an imaginary world. That’s great. But, when you give your child a touchscreen device, what the child quickly figures out is that if they touch something, they get something, and then they learn to touch to optimize the getting. And the getting is dopamine—quick dopamine.
So it’s not just the glow of the screen; it’s the reactivity of it.
That’s right. So if you just played thirty-minute videos on a video player, it would be pretty much like television was. But instead what you have is a Skinner box. B. F. Skinner was an important psychologist from about the nineteen-twenties or thirties through about the sixties. He would create these boxes—he would put a pigeon or a rat in it. By giving them reinforcement on a variable ratio reward schedule he could very quickly take control over their behavior and make them dance, make them learn to play Ping-Pong. I mean, he could do amazing things. And when you give your kid a smartphone it is a behaviorist-conditioning machine.
So anyone who says, “Oh, this is just like [the moral panic over] comic books”—no, this is really, really different from comic books. I just read you a bunch of surveys where the kids themselves say, “This is harming us.” They say, “We wish it never existed.” Half of them wish TikTok had never been invented. Nobody was saying that about comic books. So I understand people assuming that this is just another old man shaking his fist at the clouds, but this time’s really, really different.
I met with some leaders of Apple, and I raised a couple of your main points.
What’d they say?
“Turn it off.” You know, ration your time, be more logical about how you use it. It’s a great machine, you just have to, you know… They were, I have to say, pretty blasé about it.
That’s right. As a social psychologist, my rule is that, if one person is doing something bad or stupid, that person might be bad or stupid. But if all of us are doing something that seems bad or stupid it’s probably a bad situation that’s making us all behave this way.
We have a lot of experience with addictive products. We know a lot about gambling and how it ruins people’s lives. Not everyone is susceptible, but a lot are. Same with alcohol. Same with cigarettes. And rule No. 1 of addictive substances is: We don’t let companies give them to kids. We say, “Adults, we’re going to trust you to self-regulate, and ten percent of you will be severely damaged, but that’s your choice.” My God, we don’t say that about kids.
Do you think your subject—your obsession of late—is related to something that concerns me very much, which is the decline of reading? We see all this information, these statistics, about the number of people who have or have not read one book in the past year.
Yes. This, actually, I now believe, is the biggest damage. When I was writing “The Anxious Generation,” I focussed on the mental-health damage, because that’s where the evidence was best. And I mention attention fragmentation, I mention addiction, but I don’t have a lot on it. The book comes out and everyone begins talking about how they can’t pay attention anymore. And it’s not just kids. Adults are beginning to say that they can’t pay attention. And then we start hearing—
I’m telling you, it’s hard for me! I’m a professional reader.
It’s everyone I talk to.
And, as I told you before we went on the air, I have to take my phone and put it in the kitchen so that I can, in the other room, read manuscripts, read a book. And it strikes me that the rise of the phone and all it implies is the greatest experiment in human consciousness, in a sense, that hasn’t been thought through.
Absolutely.
It’s just speeding sixty miles an hour into our lives and carrying us along.
With effects far beyond what we can imagine.
The key neurotransmitter here is dopamine. Dopamine is wonderful, and we want our kids to experience a lot of slow dopamine. Slow dopamine is: your kid is trying to build a tree house, and at first he fails, and then he makes some progress, which feels really good. And so he’s motivated to work harder, and then he fails again. And over time he eventually finishes, and boy, what satisfaction that is. So that’s how you raise an adult: you give them a lot of experiences of slow dopamine. They learn to set goals and pursue them.
Here’s how you undermine that: Make available to every child, from the age of two, an iPad. And what the child will quickly learn unconsciously is they’re looking at something, and within eight seconds, they will know, this is kind of interesting, but it’s not—swipe. Oh, wow, this is so funny. Oh, this is great. Quick dopamine, quick dopamine. They go someplace else.
This is the experience that young people have had since birth, now. Kids are given iPads routinely when they’re in their strollers.
So this is not just crabby college professors whining about their students not reading “Middlemarch” in a week.
No, this is the subversion of the ability to pay attention on a species-wide level. And as one of my students said, because I showed her that Atlantic article about how students aren’t reading books anymore—she said, “Yeah, I pick up a book, I read a sentence, I get bored, I go to TikTok.” Because, again, you’ve been on this book for eight seconds and it’s not that interesting, but the thing in my pocket is a lot more interesting—quick dopamine, quick dopamine.
So this is what we’ve done. And this is even worse for the boys, because for the boys, it’s video games, it’s porn, it’s vaping, it’s gambling, it’s sports betting. So for boys, it’s open season on their dopamine systems. And this is going to make it very hard for them to develop executive function, follow goals, be useful as employees or spouses.
You came in today and put in front of me a paper that you’ve said is the most important research you’ve done. It’s called “Social Media Is Harming Young People at a Scale Large Enough to Cause Changes at the Population Level.” At The New Yorker, we wouldn’t call that a good print title—it would be a good S.E.O. title. But tell me what this report is all about, that you and your co-author, Zachary Rausch, have published this year.
Here’s a quote from Mark Zuckerberg, when he was questioned under oath in the U.S. Senate, January 31, 2src24. He says: “Mental health is a complex issue and the existing body of scientific work has not shown a causal link between using social media and young people having worse mental health outcomes.” His claim, in multiple places, is that it’s just a correlation and you can’t prove that there’s causation.
Zach and I have laid out seven different lines of evidence. We’re trying to reframe the argument as one that can actually be solved. We do have ways of knowing if A caused B in the law and in social science. So we have lots of studies, surveys of young people. What do they say? Do young people think that social media is great for their mental health? Absolutely not. So this is line one: what the victims say. And in Exhibit A we present a bunch of surveys of young people.
And these are surveys done by whom?
Pew, Gallup, Common Sense Media, many international outlets as well. We cover international research as well. Pew, which is probably the main source of evidence here, in 2src24, found that one-quarter of girls say that social media harms their mental health. One-third say it makes them feel worse about their lives. Fifty per cent say it harms their sleep. It all comes back to this question of correlation versus causation. So line one is what the victims say.
Line two is what the witnesses say, and that’s the parents, the teachers, the psychologists, the psychiatrists. They have very negative views of this. They see it up close. They say, “This is causing anxiety disorders. This is harming my patients.”
And then the third line is what the perpetrators say. If a prosecutor is laying the case out in court and he says to the jury, “I have here texts from Frank, the alleged perpetrator, in which he says to a buddy, ‘I’m going to mug Carol at three o’clock.’ ” And then he has a text saying, “I mugged Carol at three o’clock. Look at all the loot I got”—would that be evidence of causation, or is that just a correlation?
We have quote after quote [showing knowledge of harm] from inside the companies. I’ll just read a couple very briefly.
From TikTok, an internal research report: “Compulsive usage correlates with a slew of negative mental health effects like loss of analytical skill, memory formation, contextual thinking, conversational depth, empathy,” and it correlates with increased anxiety, and it goes on and on. So they know that they are hurting kids. They say the product in itself has compulsive use baked into it. They designed it to be addictive, and it’s addictive.
Snap, I already gave you the ten thousand reports of sextortion a month. Meta: “There are reasons to worry about self-control and use of our products, and”—
Wait, who’s speaking here?
So this is a member of Meta’s core data-science team and a senior data scientist at Meta having a conversation: “Without providing much more value, how to keep someone returning over and over to the same behavior each day? Intermittent rewards are most effective. Think slot machines.” They talk about it. They maximize for engagement.
Let’s talk about another country where something has been done about this. Late last year, Australia enacted a new law requiring age verification for social-media users. I think that’s the first national law of its kind.
Yes, it was.
What does this verification look like, and how is it working?
The Australia bill was very carefully drafted. They commissioned a former chief justice of the Supreme Court, Robert French, to figure out how it would be done. And it specifically says that it’s up to the companies to do it. It’s their responsibility. And it specifically says the companies cannot only ask for a government I.D. They have to offer an alternate way [to authenticate]. And there already were dozens of companies that offer alternate ways.
Wait, how does it work?
So the idea is: you want to open an account on Instagram. You put in your birthday and it then kicks you over to a page that says, “Here are four ways that you can validate that you’re old enough.”
So it’s not an honor system. Yes, I’m eighteen, on we go.
That’s right. Until December 1srcth, which is when it went into effect, the world was on the honor system. Porn, “Are you eighteen?” “Yes.” “You’re in.” So that began to end on December 1srcth.
Why did this happen first in Australia?
It just so happens that the wife of the Premier of South Australia read “The Anxious Generation” soon after it came out. And she said to her husband, Peter—Peter Malinauskas—“You got to read this book, and then you’ve got to effing do something about it.” And he did. And he called up Robert French and said, “How could we do this?” And they did it.
It wasn’t an ideologically divided issue?
It’s always bipartisan. Always. In every country, the left and the right are working together.
How is it coming along in Australia?
Here’s what we know. Julie Inman Grant, their e-safety commissioner, put out a press release about three or four weeks ago. She said all ten of the covered platforms have complied. They took down 4.7 million accounts from the 2.5 million Australian kids in that age range. And, of course, some are getting around it with V.P.N.s. Although I heard from someone who’s studying it—V.P.N. usage went way up at first, but it came way back down, because the kids want to check their social media thirty times a day, and if you have to load up a V.P.N. it’s a bit of friction. So of course kids are still getting around it, but as Julie pointed out: We’re trying to change the norms of a nation, the norms of childhood. We won’t really know the full effect for ten or twenty years.
How does it affect schools and phones?
So, two things. One is locking up the phones in the morning, a phone-free school policy, and that has magical effects—transformative effects. Some schools don’t implement it well, they don’t enforce it well, and then there’s cheating. But in schools that enforce it reasonably well the results are always spectacular. The thing that you always hear is: We hear laughter in the hallways again. The lunchroom is so loud. Kids are laughing.
That sounds too good to be true. Do we know this to be the case?
First of all, it’s very hard to find an account anywhere of this backfiring. And that would be newsworthy. There are a lot of efforts to measure what’s going on. Angela Duckworth, at Penn, is doing a major assessment, and she showed me some preliminary data in which the schools that used special phone lockers—that really took the phones away for the day—got the best results, in terms of teacher reports, academic outcomes. And the ones that used Yondr pouches got good results—not as good. And the ones that use a backpack policy, which a lot of schools do, unfortunately—“Keep it in your backpack, don’t take it out.” Look, if you’re a cocaine addict, and you’re told, “You can keep your cocaine with you all day long.” . . . So, yeah, it does seem to be working incredibly well.
The biggest argument against the Australian policy, or bringing the Australian policy to the United States, is a First Amendment argument. Explain the First Amendment argument and why you disagree with it.
So of course the First Amendment is that Congress shall make no law restricting the freedom of speech. And the companies argue that any kind of regulation is going to stop somebody from speaking and therefore violates the First Amendment. But the law already says that you have to be thirteen to sign a contract. This isn’t about who can say what. It’s about contract law. And right now, the law says that you have to be thirteen before a company can take your data without your parents’ knowledge or consent. And the Australian law says, first of all, thirteen was too low. For a child to sign a contract, they have to be sixteen. Oh, and guess who has to enforce the age limit? It’s not the child. It’s the company. So I don’t see any First Amendment—
The way a liquor store needs to see a driver’s license.
That’s right. It can’t be up to the parents to keep their kids out of liquor stores and strip clubs. It has to be the person at the door.
Now, the people being harmed by these things are not just thirteen, fourteen, and fifteen year olds. It’s even geezers like us, Jonathan, arguably. Are we left to our own devices?
We pretty much are, at least as far as I’m concerned. What I mean by that is: I’m an American, I have some generally libertarian tendencies. I don’t want to tell adults what to do. Adults can make their own choices.
How do we maintain some of the real connection and community that young people do find online?
Thanks for asking that, because this is one of the main arguments I get. It’s: thank God for social media. How could they ever connect if they didn’t have social media? How could they find information? To which I say, Yeah, kids need to connect. And the best way to connect is in person. And the second best way is by telephone or Zoom or FaceTime. And the worst way to connect is by posting something and having it be public and having people comment on it. That seems to be counterproductive. That seems to cause anxiety. That does not make people feel connected.
What I have found from talking to my students and to young people is that they’re afraid that they will be seen as a freak if they don’t have Instagram—because everyone has Instagram and Snapchat—but in fact the ones who go without it, other kids, when they talk to them, almost always say, “Wow, I wish I could do that. Wow, what’s it like?”
We’ve discussed what’s happening in Australia. What’s it going to take for anything like that to come to the United States? What’s the position of the Administration?
Two things. First, as soon as my book came out, mothers jumped into action, pressed for political action. We got huge amounts of reform in the states. Many states have taken action on phone-free schools, on regulating social media. Here in New York, our governor, Kathy Hochul, has been great on all these issues. So there’s been a huge amount of action at the state level, a huge amount of action around the world.
There’s only one place that I know of where nothing is happening, and that’s Congress. Now, what’s the role of the Administration? Because the tech moguls have been buddying up with President Trump, many people assume—and I saw this all over Europe. People are afraid to regulate social media because they think that Trump will come after them or put tariffs on them. But here’s the thing that I want everyone to notice about this. Yes, Donald Trump and Elon Musk will be very upset if you try to do content moderation and say what counts as hate speech. But if you’re protecting kids they actually have shown a lot of signs of support. The only thing America has ever done to protect kids [from social media] is the Take It Down Act, which was pushed by Melania. And the Kids Online Safety Act, the only act that ever almost made it to law a year or two ago, Donald Trump, Jr., tweeted support of it. Linda Yaccarino, the C.E.O. of X, tweeted support of it, and Elon Musk amplified her tweet. So I think the people in the Trump order—
Those seem like baby steps, though.
Well, passing KOSA would be huge, because we’ve never done much to protect kids ever. And so if we could do something in the U.S. Congress—
So you see the potential of a coherent, Trump Administration-led piece of legislation, analogous to what’s going on in Australia?
Raising the age would be a bigger step. And so that might take a while longer.
Have you ever talked to anybody in the Administration?
I’ve talked not directly with Trump, but with people in the office of the Vice-President, and . . . We have some contacts with people in or near the Administration.
And what are those conversations like?
Well, they’re interested in it, because again, everyone has kids. Everyone sees the threat. Parents everywhere see this as the biggest threat.
You have the same kind of conversations on the Democratic side?
Largely, yes. The Democrats are—
You’re saying “largely.”
Yeah. So what happens is that Meta puts out a set of talking points to inflame the right, and that is censorship, censorship, censorship. And they have a set of talking points to inflame the left. And that is that social media is a lifeline for L.G.B.T.Q. kids. And that is not true. The internet was a lifeline for them. Kids who were isolated, often in rural areas, when the internet came in—now they could find information, they could find others. There are all kinds of ways they were not isolated. The internet’s amazing.
Social media is just a small part of the internet, and it’s an incredibly toxic part. So Zach Rausch and I have an article in The Atlantic with Lennon Torres, who’s a trans activist, and Lennon talks about what happened to her when she was transitioning. And we have data showing that L.G.B.T.Q. kids do use social media more than any other group, but they’re also much more likely to report having been harmed by it. Social media is not a lifeline for L.G.B.T.Q. kids. The internet is.
Who’s your best political ally?
I would say the first one who stepped forward was Sarah Huckabee Sanders. She contacted me right away after the book came out. And then also Kathy Hochul, around the same time. So it’s female governors, or the First Lady of Virginia. Mothers are quicker off the mark. They were desperate to do something, and a lot of them have.
Interesting.
I would say female politicians, female governors—but lots of male governors, heads of state, as well.
Do you have any allies in the tech world? In other words, is there anybody unexpected who leads YouTube or TikTok and they say, “You know what? You’re right. I want to work with you to make this better and not just try to combat you.”
There’s one so far—Bill Ready, the C.E.O. of Pinterest. He reached out to me weeks after the book came out. He himself had, when he took over at Pinterest, just cut off the social features for everyone under sixteen [because of safety concerns]. And he just said, “No more of that.” So Bill Ready has been great.
You’ve not heard from Tim Cook, at Apple, or Elon Musk, or . . . ?
No, none of them. I had two meetings with Mark Zuckerberg in 2src19 and 2src2src. We debated the issue of causality versus correlation, but they have not reached out to me.
How will A.I. affect not only your work but all of us, in terms of what we’ve been discussing?
Yeah. So leaving aside all the existential risks, all the ways that it could lead to human extinction. Let’s just talk about human development.
Social media hacked our attention and took most of it. Young people, nearly half of them say that they’re online almost constantly. So social media made off with most human attention, not just for kids but for older people as well, but not as much. And that should be the crime of the century, and that is debilitating.
But A.I. is going to be much worse, because A.I. is going to hack our attachments. We have an attachment system in the parents and the child, and the infant must, with repeated turn-taking, serve and return interactions. You do something, the kid does something back, you do something back. That is what develops the brain. That is what develops the internal working models of attachment. If you have secure attachment, you’re well set up to have adult romantic relationships in which you are stable, not incredibly difficult to be with.
Now that we have A.I. in Teddy bears and in every social-media product—
What is A.I. in Teddy bears?
Oh, it’s a Teddy bear that you can talk with. It’s a chatbot in a Teddy bear. It’ll become your best friend. It’ll be supportive.
Great. That’s fantastic.
Right. So chatbots already have a death toll. We already know about kids who were talked into suicide, or at least encouraged to kill themselves and hide it. And we’re now going to raise our children with it. It’s going to be really good at entertaining them. We’re all busy because we have so much to do on our phones.
I sense another book coming from you, Jon.
I think things are moving too fast for another book. I may not write another book. I’ve got to just write articles.
Interestingly, you’ve gone from subject to subject over your adult and academic life, but this seems like now the work of a lifetime.
Yeah. This is what I’m going to do for the rest of my life, because this is the biggest—
What does that mean?
I have a contract to write a book on democracy, called “Life After Babble: Adapting to a World We May Never Again Share,” about what social media is doing to liberal democracy, how it may be incompatible with it. I’d love to write that book, and I might still write it. But by the time I write it everything could be so radically different in our country.
How would you summarize it?
I would summarize it by saying that democracy is a conversation, and when that conversation was in the Agora, in Greece, they had one kind of democracy. And when that conversation was during the Gutenberg era, which took place in print and in places like The New Yorker, and CBS News, it was a different kind of conversation. And now we’re out of the Gutenberg era—we’re into the network era. We will never again know what’s true. It’ll never be possible to have a shared reality.
So that’s not the fault of Donald Trump. He’s a symptom of it, you’re saying.
He is the first person who knew how to navigate the new world and to create some reality.
And exploit it.
That’s right. If not for Twitter, he could not have become President. But just as it is said that both John Kennedy and Ronald Reagan were extremely adept at the age of television. And Neil Postman writes about this, the great twentieth-century media theorist.
Exactly. And so in the same way, when that conversation moves on to Twitter, what happens to it? Read Federalist No. 1src, where the Founding Fathers worried about people’s ability to get pulled off into nonsense and craziness, and the ability of a demagogue to inflame the passions. They tried to design safeguards for it, but, in the social-media age, those safeguards are gone.
So that genie can’t go back in the bottle, either.
Can’t go back in the bottle. And that’s why the subtitle of the book is “adapting to a world we may never again share.” ♦
The New Yorker Radio Hour is a co-production of WNYC Studios and The New Yorker.

0 comments