91ÌÒÉ«

As governments across the world race to regulate social media and protect young users, the debate has moved from boardrooms to courtrooms.

In early 2026, juries in California and New Mexico found Meta liable for deliberately designing addictive platforms, awarding a combined $381 million in damages. Yet the picture is more nuanced than the headlines suggest. 91ÌÒÉ« also points to the potential of digital tools for early detection and mental health support. 

In this conversation, Professor , Professor of Child and Adolescent Psychiatry in the Department of Brain Sciences, offers some compelling insights into how we might find a more balanced path forward. 

There is growing political pressure to restrict or even ban social media access for under-16s –as we’ve seen in Australia, and some European countries, including the UK. Is an age-related ban a good idea? What does the research tell us? 

I can understand why people are concerned: the internet and social media were not designed with children and young people in mind, and they weren’t designed for them. We don’t have the right safeguards in place to prevent harm, as we do in most other areas of society. But an outright ban isn’t the right solution in my opinion. The more nuanced conversation is about where the harms come from and what the mechanisms are through which those harms arise. One thing that isn’t really part of the public debate is that not everyone is equally at risk. Certain types of exposure are riskier, but also certain types of people are more vulnerable. 

So, it’s not simply about how much time young people spend on social media? 

The weight of evidence is that it’s not so much about how much time you spend on social media, it’s about whether that time displaces other activities that are beneficial for mental health. What would be harmful is if you had no time to interact with other humans in real life, no time to sleep, no time to exercise. That’s the first category of harm: displacement of health-beneficial activities. There’s also a chicken-and-egg problem: sometimes people are on social media because they can’t sleep, rather than the other way round. Showing cause and effect is quite tricky in some of these associations. 

And what about the content itself – is that where the real risk lies? 

There are clearly certain types of content that can be harmful, either directly or indirectly. Indirect harms are things that fuel existing anxieties or distorted perceptions. In my clinical area of eating disorders, for example, we know that if you tend to social comparison and perfectionism, social media can hugely amplify those effects, because people present themselves in the best light when they post. That’s always been true of magazines. What’s different now is the accessibility and the feeling that it is not some remote, unattainable ideal on a screen or in a magazine, but someone who seems just like you. 

The other risk factor is cyberbullying, where social media is used to exclude or harm people. It is particularly toxic for people sensitive to social rejection. We’ve done studies on this, and what we found is that some of the more extreme personality traits you’d associate with bullying in real life weren’t as pronounced in young people who were cyber-bullying. In other words, young people with a fairly normal personality profile were doing things online that others experienced as bullying. There’s something about social media that allows people to behave in ways they wouldn't behave face to face – perhaps because you’re less directly accountable. 

 
You've also looked at profiles and sites related to eating disorders. 

That’s the more sinister end. There are sites where people are actively teaching others how to engage in harmful behaviours, and there’s absolutely no doubt in my mind that regulation is needed there. But I think the broader area of concern is more like school bullying that happens to occur online, and therefore it can follow you into your own bedroom – a place that used to be a place of safety. I absolutely understand why that’s alarming. Our research shows that parenting behaviour is hugely important in protecting young people against those risks. But we also know there are things schools can do. Creating environments where cyberbullying doesn’t happen as much isn’t far-fetched. Saying “all schools have bullies, so don't go to school” – that would be ridiculous. We need to make sure all young people are equipped to manage power imbalances in relationships as well as their own feelings and behaviour.  

We need to make sure all young people are equipped to manage power imbalances in relationships as well as their own feelings and behaviour.   Professor Dasha Nicholls Professor of Child and Adolescent Psychiatry

You’ve argued that social media and smartphones can also be used for good. Can you tell us about that research? 

One of the things I've been working on with a colleague who specialises in body image and eating disorder prevention is harnessing the influence of social media for positive benefit. Because young people do live in a social media world, we asked: can we get people who have influence to use it positively? She led a small study looking at whether social media influencers could deliver micro body image interventions, and the results suggest that if you can get influencers to do this, you can reach a very large audience in potentially beneficial ways. The problem is that the news tends to want to amplify fear. It’s very easy to find stories about harmful social media influence, and much harder to find stories about positive social media influence. 

What about using smartphones more directly for mental health monitoring and intervention? 

This is increasingly challenging, frankly, because of the widespread belief that all social media is harmful. But we and many others have been looking at how smartphones, which we carry with us all the time, can support mental health. One approach is to use them to seek help in a crisis.  There are interventions like Shout, designed for young people to reach out in the middle of the night, connecting them immediately to a support network. But there are also many more targeted interventions based on well-researched psychological models – cognitive behavioural therapy, cognitive dissonance – delivered remotely. You could have a virtual therapist, who could be a real person or, increasingly, not. 

And then there’s an in-between space for people who aren’t yet unwell but are beginning to experience distress: maybe disengaging from friends, sleeping poorly, losing their appetite. Your phone can recognise that this is a change from your usual pattern and deliver what's known as a ‘just-in-time' intervention: “We’ve noticed something’s a bit different. Are you OK?”  The sort of thing your mother might do, but delivered in real time. 

This links to your research on digital phenotyping. Can you explain what that means in practice?

My PhD student, , is currently running a randomised controlled trial to test whether personalised nudges  (“I’ve noticed you’re doing XYZ, have you thought about talking to someone?”)  are more effective than generic nudges like reminders to hit your step count, or than simply monitoring your own mental health without any feedback. We’re looking at whether there’s a meaningful difference between those three groups. We designed our own app, called Mindcraft, to enable this kind of research because we can tweak it ourselves based on feedback from our users. 

The point about phenotypes is that different people will respond in different ways. What AI and big data processing allow us to do is to say: this type of person responds to this type of intervention, rather than taking a one-size-fits-all approach. We’re beginning to move towards personalised interventions – not just that the content is personalised, but that we have a better idea of the profile of a person who might benefit from a particular kind of support. If you’re quite introverted, what might work for you is going to be very different from someone who’s naturally more extroverted, for example. 

You mentioned in your research that adolescents are particularly vulnerable to mental disorders, with over 75% of lifetime cases emerging before the age of 25. Young people's mental health deteriorated markedly during the COVID-19 pandemic and has not fully recovered since.  Are we seeing any shifts in the prevalence of conditions like depression, anxiety, or eating disorders since the pandemic? 

That’s a very big question, and it’s also the subject of the independent review into mental health conditions, ADHD and autism. The review is trying to answer whether the apparent increase in prevalence is real, or whether we've tipped into over-diagnosing mental disorder for what is normative distress. The findings are very mixed. We’ve just published a commentary showing that, if you look at something like ADHD, there's as much evidence of under-diagnosis as there is of over-diagnosis. There's also a massive social gradient in who does and doesn’t get diagnosed and treated – people with ADHD who do get diagnosed and treated are much more likely to be white and middle class. The overwhelming evidence is that we're under-recognising ADHD in other populations. 

There’s also the question of the boundary between reporting symptoms and the level of distress or functional impairment associated with them. One argument – and I don't pretend to have the answer – is that the reporting of symptoms is higher than before, but the levels of distress and functional impairment haven’t increased correspondingly. But that’s a very patchy story, and there are some groups where there are undoubtedly higher levels of distress. 

What can schools and teachers practically do to support young people’s mental health, and is that a realistic expectation given everything else they face? 

 I've been involved in research showing that there are things about the school environment that can reduce bullying and promote well-being. Creating environments where harmful behaviour is less likely is possible. But I think the more important point is that adolescence is a time when young people are navigating hugely heightened emotions, often for the first time. The first time you experience rejection, it’s the end of the world - the second or third time, it’s not. They're learning to process these feelings, put them in perspective, recognise their own feelings and those of others. Expecting young people to automatically know how to do that without being well-scaffolded and supported by those around them, including in the digital world, is the biggest challenge. Schools have a role, but so do parents, and so does the digital environment itself.  

Young people today are growing up saturated with distressing news – war, economic insecurity, climate anxiety. How do we realistically promote good mental health in that environment? 

 I’m not sure I would say social media is the primary harm there - the mainstream news now is, let’s be honest, as harmful to mental health as anything on social media. So, limiting how often you access the news is one very practical step. The second thing is recognising that social media and news are not the same thing. Social media has been filtered through other people’s lenses in ways that amplify anxiety. Both can be harmful, but in different ways. And the third is recognising your own tendencies. If you know you’re someone who gets obsessive about things, or who gets very anxious when exposed to catastrophic news, that’s what you need to manage, and as a parent, what you need to help your child manage. But that’s very individual – not every child is at risk in the same way, which is why blanket policy rules are so difficult.