Something strange is happening with teenagers’ mental health. In Britain, the US, Australia and beyond, the same trend can be seen: around the middle of the last decade, the number of young people with anxiety, depression and even suicidal tendancies started to rise sharply. Jonathan Haidt, a psychology professor at New York University’s Stern School of Business, noticed a change when students who were brought up with smartphones started to arrive on campus. They were angrier. More fragile. More likely to take offence.

Social media, he concluded, was shaping their view that society is in permanent conflict, which in turn led to ideas about microaggressions and competitive victimhood. All this, he found, was damaging young people’s mental health. He is working on a book, due out next year, and is ready to share his thesis.

We meet in unusual circumstances: over a videolink and in front of an audience at a conference in London. What he says draws gasps and applause from those watching, though his message is quite horrifying.

He argues that the tools of social media are just too sharp for young minds. On digital platforms teens parade themselves, often to an audience of strangers, and this is leading to addiction, paranoia and despair. For girls, the effect is especially acute. ‘What we’re seeing is a very sharp, sudden change in girls’ mental health all around the Anglosphere and the Nordic countries,’ he says. A big change was evident from 2013, when physical friendship groups started to be supplanted by smartphones and online chat. ‘But you cannot grow up in networks. You have to grow up in communities.’

It is striking that boys who have religion in their lives seem to be less susceptible. ‘If you’re a kid who’s a religious conservative, on average, your mental health is not really much worse than it was ten years ago. But if you’re a secular liberal girl, you’re probably more than twice as likely to have a mental health problem.’ He cites a University of Michigan survey into ‘self-derogation’ – i.e., how likely teenagers are to say they are ‘no good’ or ‘can’t do anything right’. Figures had been stable for years but started rising sharply ten years ago – except for among boys who identified as conservative and said that religion was important to them.

Faith, it seems, does not help girls as much. Why not? One theory is that girls simply use social media more. But Professor Haidt also thinks they are more likely to buy into what he calls the ‘three great untruths’ of social media. The first is that they are fragile and can be harmed by speech and words. Next, that their emotions, and especially their anxieties, are reliable guides to reality. And finally, that society is one big battle between victims and oppressors. All this, he says, is the subtext to social media discourse.

‘It’s what I’ve been calling the phone-based child,’ he says. ‘For all human history, millions of years, all mammals play. Anyone who has had a puppy knows it’s all about play. So we had playdates in childhood, up until around 2010.’ In Britain, he says, the number of children who went on real-life playdates then fell sharply. Lockdown, needless to say, didn’t help.

‘I’m calling it the great rewiring of childhood… It hit the US, Canada, Britain, Australia, New Zealand in exactly the same way.’ Social media is a bit of a misnomer, he says. It’s no longer about connecting people, but ‘performing on a platform’. Perhaps this is fine for grown-ups, but not for children, ‘where they can say things in public, including to strangers, and then be publicly shamed by potentially millions of people… Children should not be on social networks. They should be playing in person. Social media platforms should never be accessed by children until they’re 18. It’s just insane that we let kids do these things.’

I ask if he thinks all platforms are equally dangerous; it’s hard to imagine anyone getting up to much mischief on LinkedIn, for example. ‘TikTok is probably the worst for their intellectual development. I think it literally reduces their ability to focus on anything while stuffing them with little bits of stuff that was selected by an algorithm for emotional arousal. Not for truth.’ And if you get your news from social media (which many people do – in the UK, Instagram has overtaken all newspapers as a news source), this can change your view of the world, especially as the algorithms tend to promote the most provocative views.

If asked to choose whether they side more with Israel or Hamas, ‘the great majority of Americans side with Israel, except for Gen Z, which is split 50-50’, Haidt says. ‘There was a Twitter thread recently showing how if you look at what people are saying on TikTok, you can understand why. TikTok and Twitter are incredibly dangerous for our democracy. I’d say they’re incompatible with the kind of liberal democracy that we’ve developed over the last few hundred years.’ He’s quite emphatic about all of this, almost evangelical. Which makes me think of his 2012 book, The Righteous Mind, in which he argued about the danger of getting too caught up in your own bubble, believing your own spin. Might he be guilty of that here? Might it just be the case, I ask, that there’s less of a stigma around mental health now, so teenagers are far more likely to admit that they have problems?

‘But why is it, then, that right around 2013 all these girls suddenly start checking into psychiatric inpatient units? Or suicide – they’re making many more suicide attempts. The level of self-harm goes up by 200 or 300 per cent, especially for the younger girls aged ten to 14. So no, the idea that it’s just a change in self-report doesn’t hold any water because we see very much the same curves, at the same time, for behaviour. Suicide, certainly, is not a self-report variable. This is real. This is the biggest mental health crisis in all of known history for kids.’

Haidt believes that since this crisis translates into suicide and self-harm, it should prompt a robust government response. ‘How many kids died from Covid? How many from polio? The increased number of suicides since 2010 is so large that I suspect this is among the largest public health threats to children since the major diseases were wiped out.’ In Britain, suicide rates started rising in 2014, up about 20 per cent for boys (to 420 a year) and 60 per cent for girls (to 160 a year).

What should parents do? They know that if they try to remove their teenager’s smartphone, their child will accuse them of destroying his or her social life. ‘That’s a perfect statement of what we call a collective action problem,’ he replies. ‘Any one person doing the right thing is in big trouble. But why do we ever let our kids on social media? It’s only down to the dynamic you just said.’ New norms are needed, he says. And his book will suggest four.

Rule one, he says: no smartphones before the age of 14. And what if they say they’ll be cut off from their friends? ‘Give them a flip phone. Millennials had flip phones. They texted each other, “See you at four at the mall.” And then they would meet at the mall. Just don’t give a smartphone to your ten-year-old. Wait until 14.’ Rule two: no social media before 16. ‘If half the kids are not on social media, they actually meet up after school and they do fun things. They’ll become the cool kids.’

His third rule: no phones in schools. ‘That does not mean keep it in your backpack,’ he says. ‘Otherwise, the kids are going to go to the bathroom. They’ll find ways to get their fix.’ And finally: more unsupervised play. ‘Both of our countries freaked out in the 1990s, locked up our kids because we lost trust in each other. We thought everyone was a child molester or a rapist.’ Children and teens could do with six or seven hours each day out of contact with their parents, he argues. Keeping them inside risks more harm than the outside world would pose.

Our conversation ends abruptly as the conference, the Alliance for Responsible Citizenship, breaks for lunch. The aim of that conference was to discuss important issues that aren’t on the political radar. The Online Safety Bill has just become law but has nothing to say on teen screen time. If Haidt’s theory is even halfway correct, regulation is unlikely to be far behind.