Science, Innovation and Technology Committee 

Oral evidence: Social media age restrictions, HC 1710

Wednesday 11 March 2026

Ordered by the House of Commons to be published on 11 March 2026.

Watch the meeting

Members present: Dame Chi Onwurah; Emily Darlington; Dr Allison Gardner; Kit Malthouse; Dr Lauren Sullivan; Adam Thompson; Freddie van Mierlo; Daniel Zeichner.

Questions 1 – 59

Witnesses

I: Dr Rebecca Foljambe, Founder, Health Professionals for Safer Screens, and GP Partner; Frank Young, Chief Executive, Parentkind.

II: Professor Lorna Woods OBE, Professor Emerita, School of Law, University of Essex, and advisor to the Online Safety Act Network; Dr Kim Sylwander, Research Manager and Researcher, Digital Futures for Children Centre, London School of Economics and Political Science; Baroness Kidron OBE, Cross-Bench Peer, House of Lords, and Founder and Chair, 5Rights.

III: Julie Inman Grant, eSafety Commissioner, Australia; Professor Amy Orben, Research Professor and Programme Leader, MRC Cognition and Brain Sciences Unit, University of Cambridge, and Fellow, St John’s College, University of Cambridge; Professor Jeff Hancock, Founding Director, Stanford Social Media Lab, Director, Stanford Cyber Policy Center, and Harry and Norman Chandler Professor of Communication, Stanford University.

Written evidence from witnesses:

• Health Professionals for Safer Screens
• Professor Lorna Woods
• Julie Inman Grant
• Professor Amy Orben
  

This evidence includes content that some may find distressing.

Examination of witnesses

Witnesses: Frank Young and Rebecca Foljambe.

Q1                Chair: Welcome to today’s meeting of the Science, Innovation and Technology Select Committee. The Government have launched a consultation on measures to keep children safe online, including a potential ban on under-16s having social media accounts. This has been the subject of a lot of debate in the media and around the country.

Today, we are going to examine the actual evidence for introducing a ban and compare it to some of the alternatives. Across the country, parents and experts are debating the need for a ban. We will hear from both sides of the debate, and we will also hear from Australia about the first few months of the restrictions that are in place there.

We have three panels this morning, but, before we come to our first set of witnesses, I really do want to place on record the Committee’s thanks and gratitude to those who have taken the time to contact us and provide written evidence to assist our work. We are especially grateful to bereaved families for contributing their views. We will do our best to reflect all your views in our questions this morning.

We are going to be speaking with the first set of witnesses today about the evidence and arguments in favour of introducing a ban, and I am going to ask each member of this panel to introduce themselves the first time they come to speak.

My first question is just a basic one. What is the evidence for under-16s being harmed by social media use? What are the biggest causes of harm—for example, attention-grabbing design, increasing screen time, et cetera? We will start with Frank Young.

Frank Young: Thank you very much. I am Frank Young. I am CEO of Parentkind. We represent the voice of parents, and we conduct regular polling of parents and publish parent voice reports.

That is important because when we poll parents—they are a broad cohort—normally they divide something like 30/30/30 or 40/40/20; they are equivocal. There is one issue where they are not equivocal, and that is social media bans and social media harms. Parents are overwhelmingly concerned about social media and the risk of their children’s access to social media. About 93% of parents tell us that they think social media is harmful.

You never see 93% in parent polling. It is just unheard of. This underlines just how seriously parents as a cohort take social media as a concern. They see the harm with their own eyes and from their own experience of being a parent.

We have also spoken to teenagers themselves. I will give you some headlines from 16 to 18-year-olds who we polled. About one in five told us that social media had at some point made them feel that life was not worth living. About one third are regularly receiving sexual imagery or messages from strangers. That rises to about half for teenage girls. Another third are receiving extremely violent or gory images regularly. The dangers and risks for children and for parents are very clear, and that is why they are concerned.

Q2                Chair: You have set out the views of parents and the risks, and we will come to look at that in more detail, but I did ask you about evidence—studies, scientific learning, et cetera. Is there actual evidence for the harms?

Frank Young: There is. We have conducted polling of young people themselves, asking them to tell us about the harms they are receiving as older teenagers and what their experiences are as teenagers going through adolescence and childhood now. That is nationally representative polling. We stand by it and we believe it has credibility.

One third tell us that they are regularly receiving sexual or violent imagery, which we simply would not tolerate ourselves. That is the evidence that underpins parents’ worries.

Chair: That is evidence of experience.

Frank Young: That is right. We speak for parents, too. We try to be their voice.

Dr Foljambe: Hello, everyone. I am Dr Becky Foljambe. I am a GP in North Yorkshire and founder of Health Professionals for Safer Screens. I am delighted to be here.

If we are looking at evidence, from a health professional perspective, we have found it quite frustrating that there is so little research to date that looks at the actual health outcomes in our children. You will have seen that the Academy of Medical Royal Colleges, which is made up of 22 medical colleges from different specialties, recently came out and declared this issue a public health emergency.

It is important to say, from our perspective as clinicians, that we consider what we see in practice as valid evidence of the harm that children are coming to on these platforms. We believe that this harm is increasing day by day, week by week, year by year, and it is showing no sign of decreasing.

To look at the mechanisms of harm, we will go from the first principle that frustrates us as doctors. It is our job first to do no harm. We believe that any product for children should be the same; it should do no harm. When we look at, from our perspective, how social media is so strongly implicated in a lot of the presentations we see in children, we become very worried.

We can look at health misinformation for young women. The uptake of hormonal contraception has gone down by over 50% in the last 15 years. Women’s health leaders in this country say that that has been predominantly fuelled by misinformation on TikTok, telling young women that hormonal contraception is going to make them fat, infertile and cause brain tumours.

This is really damaging for young women. We have seen a direct drop-off in their uptake of hormonal contraception. Teenage pregnancies are back on the rise after decades of work to bring them down, and we have the highest ever rate of abortions among our young women at the moment. This is driving inequality as well because we know that women from deprived backgrounds are three times more likely to have an abortion and 20 years less of health. We are seeing profound relationship damage, and we are experiencing that in consultations with young women and men.

We know that there is an absolute epidemic of CSAM on the internet: naked pictures of our children everywhere. The CSAM reported from the Internet Watch Foundation is the worst on record; it went up 7% just from last year. They found over 300,000 sites with child sexual abuse material. We know that 90% of that is self-generated and over half of the perpetrators are under 18. They have seen a 260-fold increase in AI-generated CSAM, 65% of which is the worst category of sexual abuse.

Is it any wonder that organisations such as the Institute for Addressing Strangulation have found that a third of our 16 to 24-year-olds are strangling or choking each other at least once during sex?

We have seen a direct increase in our children are reporting violence against animals and towards each other. We had a recent case of a child who watched animal cruelty on the smartphone of another child at school, then and went home and killed a litter of puppies. Then he came to us with severe nightmares. He had watched a video of how to do that on social media through someone at school.

It is no surprise that constant, highly visual image and video-based content is partially fuelling our increase in eating disorder symptoms.

It is strongly evidence-based that sleep has been directly impacted by smartphones for the past decade. There is a very strong evidence base. We know that children are late-night scrolling in huge numbers. Ofcom itself found that up to 24% of our 14 to 18-year-olds last year were found to be on their smartphones between 9 pm and 5 am. We know that it interrupts sleep, and we know that anything that interferes with sleep provokes impulsive behaviours, accidents, and difficulty controlling emotions and triggers, which worsens mental health symptoms.

They are just a few of the areas. I have some evidence for under-16s, if you would like that.

Chair: Thank you very much. That is very sobering evidence of the impact of specific types of content and misinformation.

Q3                Adam Thompson: Good morning, both. Becky, thank you for your really clear introduction. My understanding is that there is not really a consensus on a causal relationship between social media and negative outcomes. Given what we are discussing today—a potential ban on social media for under-16s—in your view is there enough hard evidence to justify such a ban?

Dr Foljambe: There is more than enough. We need to look at what we are considering to be evidence. University-based evidence, as important as it is, is only a small part of the evidence that we should be looking at. We would say that the evidence at the moment—there is so little that looks at the actual health outcomes, as I have said—looks through a very narrow lens, using traditional research scope and processes that take time.

What really troubles me is that it goes from a baseline of children being completely desensitised to this current level of digital surveillance, social comparison, pressure to conform and all these norms pushed by addictive algorithms. We know that there is alarming evidence coming up, even from universities using traditional research paradigms, about things like the addictive way these children are using these smartphones; predominantly watching social media is consistently being shown to show an increase in suicidal acts later down the line. We know that deliberate self-harm and attempts at suicide are grossly underestimated in the clinical world. We do not even know that a lot of it is going on. To see that coming to us in bigger numbers is a real concern.

When we are talking about cause in medicine, in public health we look at things being significant components that make up the chain that leads to the outcome. We know that mental health has a variety of outcomes. We would very strongly say, as doctors, that those mental health outcomes are being eroded significantly by this component of children’s use of social media.

We use the Bradford Hill criteria to find cause. That was really the only way we ever escaped the tobacco vice. When you look at things such as adverse childhood experiences, it is accepted that they are a causative agent in mental and physical health problems. It is strength of association, consistency, plausibility and temporality. All that is associative and correlational. We can never get experimental evidence on that. We have got to that place with social media. We know how much harm it is causing.

Q4                Adam Thompson: I accept that, Becky. It is very clearly argued. If we work under the assumption that there is plenty of evidence for the damage it does to children, there is a risk when these kids turn 17. What is the difference between a child who is 16 and 364 days old and one who is 17 years old? Is there a risk of the floodgates opening—essentially a cliff edge when a child turns 17 and the potential impact of their mental health as a result of that?

Dr Foljambe: I struggle with the cliff edge analogy. If we are associating the use of social media with cliff edges, then it is predominantly harmful, because we do not want children near cliff edges. It is also very misleading to think that just letting them be on it longer prepares them for it better. It is not; it just exposes them to a longer period of harm. Why are we so comfortable with 13 as a cliff edge, when we know that 40% of users are under 13? We have 800,000 three to five-year-olds using social media independently. The cliff edge argument is weak, and it is not an excuse not to legislate.

Q5                Adam Thompson: No, I think that is totally reasonable. If anything, I was thinking in the other direction. Is 16 too young? Could it go older?

Frank Young: Could I jump in on that? It is an important point, clearly, for parents too, who are raising the children. We have to recognise the developing brain and the difference between childhood and what you might call emerging adulthood. The nature of social media—being addictive by design, with the dopamine hits from likes and constant repetitive engagement with social media—is training that childhood brain in an addictive way that simply we would not want. That is why we would want to put an age on it.

Ultimately, politicians have to decide. Ultimately, there is a line that is drawn and an age that is chosen. Sixteen seems to be a sensible age. Some of the evidence suggests that it is the right age to draw the line. It is an age at which children are becoming young adults, and we draw that line in other places, but ultimately it is for Parliament to draw that line.

What we are talking about here is children—children with developing brains, both as young children and as adolescents, too, where the real harm from addictive devices starts to develop.

Dr Foljambe: In answer to your question, we have just conducted a survey of 340 health professionals from different specialties, and 34% of them said that 16 was not old enough. They felt that, if this is so harmful for children, childhood goes to 18 and that should be the cut-off. Sixteen can feel a bit arbitrary in the sense of some of the harms.

Frank Young: Parents support the age of 16. When we have tested the idea of different ages, 13 through to 18, generally parents settle on 16 as about the right age.

Q6                Emily Darlington: I just have a quick question. If we are to use a science-based approach, brain development finishes for women at about 20 and for men at about 25. Are you asking the Government to consider when brain development has reached maturity?

Dr Foljambe: Yes, but not that old because that just sounds a bit unachievable and ludicrous. I totally accept that argument because it is absolutely right that adolescence moves long into the 20s, and we know that now; it is an established piece of neuroscientific fact.

If it is helpful, looking at the 16 cut-off, which is what we seem to be working with, we need to focus on what we already know about that period of time, especially age 11 to 16, when we know as parents that children can rely on emotions rather than logic in making their decisions. It is the time that they are most emotionally sensitive and their peer group is the most important thing in their lives.

Looking at that period of time, we know, as clinicians, parents, police and teachers, that that time is especially fragile. It is incredibly wise to aim to protect children from social media exposure, especially in that age group.

Q7                Emily Darlington: Yes, but science would tell us 16 for girls and probably later for boys in that stage.

Dr Foljambe: I want to check this with Amy behind me.

Emily Darlington: I do want to get on to my next questions.

Chair: Perhaps you could write to us about that.

Emily Darlington: Yes, you could write to us about some of that evidence.

Dr Foljambe: There are studies, such as Amy’s longitudinal study, that have shown that the most detrimental effect windows are 11 to 13 for girls and 14 to 15 for boys. Massive international studies have found that the highest proportion of addictive and problematic social media use, at 23%, was for 13-year-old girls. That was the highest.

Kit Malthouse: That coincides with the maturing of the limbic system in the brain, which is the earlier bit of adolescence.

Q8                Emily Darlington: Yes. It is just something interesting for us to pick up in our next inquiry when we are talking about this. We are the Science, Innovation and Technology Committee, so we are really trying to look at evidence.

I want to come back to attitudes. I similarly did a survey of 500 14 to 16-year-olds in Milton Keynes and came up with very similar outcomes in terms of what they are doing. I went further and found out where these approaches were happening and where they were seeing damaging material, and they were not necessarily what we would consider social media, so I wanted to talk to you guys about that.

The platforms where they were getting approaches from strangers and seeing damaging material were overwhelmingly YouTube and Roblox, which are not necessarily considered within the definition of social media material. Are you suggesting that we ban those? I am trying to understand better what you mean by a social media ban and how you would define “social media”.

Frank Young: Australia clearly went through this process, perhaps imperfectly. Once they had decided that they wanted what they called a ban on social media, they then had to discuss exactly what they meant by social media and which platforms were in and which platforms were out.

Often WhatsApp is one of the most contentious platforms. It tends to fall outside of the remit of social media, but young people themselves—and I dare say Members of Parliament—would recognise the intensity of WhatsApp and the potential for harm and for strangers to contact young people through encrypted messages.

We would have to go through a process, as a country, of being really clear and understanding—hearing from both parents and young people themselves—about which platforms are causing them the most harm and the most addictive harm. Clearly, platforms such as Snapchat and YouTube need to be within the realms of a social media ban, along with messaging services such as WhatsApp.

Q9                Emily Darlington: I had the pleasure of meeting most of the parents of these children in a workshop that we did. What was really clear from the survey that we did of the children was that 14 to 16-year-olds absolutely recognised the harms but did not want a ban; they wanted more regulation. Similarly, with parents, those who felt that social media could not be regulated wanted a ban and those who felt that you could improve their child’s online experience wanted to improve their child’s online experience.

In the survey work that Parentkind has been doing, have you asked those kinds of questions about the preferred options to make the online space safe or about whether people feel that that is just not possible and there is no level of online safety that you could have for people under 16?

Frank Young: It is fair to say that there will always be differences of opinion among parents as a cohort. They are inevitably a large cohort, so there will always be nuance.

However, when we speak to parents and when we survey parents, about two-thirds, 67%, say they support a full ban on social media. They say that because it gives them clarity. It effectively makes you guys the bad guys, not them. It is the same with things like the age of consent. It gives parents the clarity to say, “No, I can’t do this. You can’t access social media because it is simply against the law. It is not me. It is Government and Parliament making that decision”.

Emily Darlington: That is great. I can be the bad guy at home with my children and the bad guy here.

Frank Young: Dare I say that the experience of being a parent is always being the bad guy. Seven out of 10 parents say that a clear social media ban would make their lives easier—that it would be easier for them to police social media.

Q10            Emily Darlington: Dr Foljambe, I am really interested in the pieces of work that you talked about in terms of harms and other aspects. I want to know your professional view on the harms of this not only to people under 16, but to those over 16. Could there be Government intervention on some of the things that you have quite rightly identified, such as eating disorder material, violent material or that horrific case that you talked about with the puppies? That was just horrific. Would you like the Government to clean up, for everyone, some of that material because of the harm that it can do to post-16s? This is similar to what Adam was talking about. If it is harmful for under-16s, is it not harmful to all of us?

Dr Foljambe: We cannot clean up everyone, but I would like to see the Government make a deliberate proactive move to safeguard children using a law, because people from all backgrounds understand that a ban means a law, and laws fundamentally are there to protect.

With regard to addictive uses of smartphones and social media, any psychiatrist or GP we talk to all say the same: we are so close to this reaching criteria as a diagnosis of behavioural addiction. That is serious. We cannot have hundreds and thousands of our children at very young ages behaviourally addicted to these platforms. It is not just the fact they are using them for hours on end but the content. Gambling is behaviourally addictive and has laws.

The Government should feel confident and have the freedom to make those parallels because the way these children have been collectively pulled on to these platforms in this way is grotesque, really. I have been a doctor for 25 years, and the changes that we have seen since this stuff came in are so far reaching that I cannot list them. The list is too long.

We know that this is a significant component of why children are the way they are now. I am without doubt that it is a huge part of the mental health crisis we are seeing in them as well. First and foremost, they are not getting enough sleep; 70% of our children are not getting enough sleep. It does require a law to educate and lay down a line to support parents. It is just so important. It is a safeguarding responsibility.

Q11            Daniel Zeichner: Do you believe it is likely that young people would circumvent any such ban? There is talk about VPNs and so on, but there is also this definitional question of simply moving to something else that we have not proscribed. How risky is that?

Frank Young: There is imperfection built into a ban. A ban provides clarity to parents, but clearly there will still be a cat and mouse process going on. The argument for a ban from charities such as Parentkind and from parents is that it starts to change the culture.

This is what we are seeing in Australia now, and you might hear from witnesses later on about this. We have spoken to academics from Australia who are part of the process of introducing their social media ban, and what they are principally talking about is changing the culture around social media. It is less permissible for children to be on social media, to be active on those platforms.

You have a scenario where social media is now ubiquitous and almost all teenagers have a smartphone device that allows them access to social media apps. You are starting to change that culture; it becomes less and less acceptable once Parliament has led that culture change through clarity and law. But I do accept that the imperfection of a ban.

Q12            Daniel Zeichner: On that culture question, though, many of us can remember things that were banned in our youth and banning them made them more, not less, attractive.

Kit Malthouse: It still did reduce usage.

Frank Young: There will be less of something when Parliament enacts the clarity of a ban. Look at something like wearing seatbelts in cars. There are often analogies made with previous Acts of Parliament. Sure, it might make something more enticing or alluring for a teenager, but it also helps parents to say, “This is simply not allowed”. Schools will not allow it any more, and it is not allowed outside of the school gates. There is now a law. It is just something that you are not allowed to do. That is an example of Parliament leading and changing culture.

Q13            Daniel Zeichner: What assessment have you made of alternatives to banning? No one could have heard the evidence you have put this morning without being concerned about it, but is a ban the only way? Are there restrictions that might be more effective in achieving the outcomes that we want to see?

Frank Young: On behalf of parents—we speak for parents—we support the ban because of the clarity it gives parents. There are clearly other options presented in the Department’s consultation, such as curfews or further regulation around social media. We would be troubled by those alternatives because they do not give parents in particular, who are ultimately put in the position of policing these things, the clarity that they need to do the job that we want them to do.

Q14            Daniel Zeichner: I just have one other question. There clearly are some young people who benefit from access to social media. Is that your experience? How would you try to trade off the potential benefits against the harm that you have described?

Dr Foljambe: No, I would not say that the benefits are of huge interest to us at the moment, in light of the harms. The harms are not just small incidences for a few children. It is affecting their sleep. They are showing addictive behaviours. There is an increase in eating disorder symptoms, cruelty to animals, cruelty to each other, sexploitation, violent acts and increased criminalisation of children. Teachers, police and doctors are incredibly worried that a lot of this has been provoked or directly caused by social media. In that context, I am less interested in the benefits because there are other ways for children to communicate.

It could be a case of telling these companies, “You cannot have access to our children any more. We are more than willing to look at the benefits, if you remove the harms”.

Q15            Daniel Zeichner: So you do not see any benefits?

Dr Foljambe: I am not interested in the benefits in light of where we are now, no. I find it quite troubling that it is stalling us. I have a house full of teenagers. I know how they use social media, and I know the fun that can be had on family WhatsApp chats. WhatsApp probably would escape any restrictions or bans. If it is shown to be benign and useful to children, that is great. But TikTok, Instagram and the like are dreadful, and they need to clean up their act.

Can I just make one comment on the ban? You made the comment about how, when something was banned, we did it more because we were at that thrill-seeking stage. I agree, but the laws around, for example, smoking and alcohol have meant that there are now much fewer children abusing or misusing alcohol and smoking underage than there are misusing social media and smartphones. In that sense, bans—laws—are there as a public health message and as protection. I appreciate what you are saying. Children will circumvent, but many more children will be protected.

On VPNs, the amendment does have a broad scope to include making VPNs inaccessible to under-18s. I think there are many ways to make this work.

Q16            Daniel Zeichner: I will go back to Frank on that point about whether there are benefits and whether there will be groups that will be disadvantaged by a ban.

Frank Young: We polled parents on this, just as much as we asked parents, “What harms are you seeing? What are your concerns?” Only 5% of parents said they felt social media helped their children access support for personal issues. Only 8% mentioned positive influences from social media. That reflects the very low incidence of parents saying that there are benefits to social media versus perhaps the wider online experience.

Q17            Daniel Zeichner: I hear you about parents, but what about young people themselves?

Frank Young: We have to be very careful when we are talking about potentially vulnerable young people or young people who are exploring their sexuality. We have to be clear that we are not proposing to ban the internet or places where young people can receive advice. We need to make sure those places where they receive advice are either moderated—we remove any sort of wild west danger to those places—or carefully controlled by groups who we would respect as offering good advice.

I will just make a personal point about culture change. My son goes to school in the fine city of Cambridge. He goes to a school that is absolutely clear that there are no smartphones. Children are not allowed to bring smartphones into the school, and that has helped to change the culture around the parent community. It is a great school in Cambridge.

Daniel Zeichner: That is true of every school I go to in Cambridge; it is not unique.

Frank Young: Yes, it is a fine city.

Q18            Chair: Just picking up on the point that Daniel made about examples of harmful products and services in the past, every generation has a new technological challenge and the opportunity to meet it. I can remember the controversy around video gaming a couple of decades ago. The decision as to whether to ban or to regulate better is a trade-off with the benefits that it can bring.

You have both basically said that the harms are so great that there are no benefits. Yet I think this Committee would feel that, as adults, in our lives, we need to be able to use social media. Social media is an essential part of so much of society and technology—we have not talked really about ChatGPT—and will be even more so in the future. Are you absolutely convinced that this is a case of banning, not better regulation?

Dr Foljambe: I am not convinced of that, no. We need both. At the moment, the regulations are not exercising enough authority to catch these companies at the point of harming the children. We need to introduce laws that make them increase their level of accountability and responsibility to do that properly.

Chair: That would be in addition to a ban, though.

Dr Foljambe: Yes. We always need regulations, digital literacy and that sort of thing.

Frank Young: We would agree that a ban would sit within public health messaging, along with education and support for parents. It is never just one thing.

Dr Foljambe: It is not that we are saying that social media has no benefits at all, but there is very little robust evidence of its benefits.

Q19            Dr Gardner: Just very quickly, I am very concerned that we have a lack of definition of social media. You are looking to put in a ban but you have not defined exactly what you are banning, and that is really concerning to me. If you do it from a named platform point of view, you are dealing with a constantly shifting background. Are you talking about user-to-user functionality? Would gaming come under that? What about ChatGPT? There are huge issues with the use of ChatGPT. I went to a Parentkind drop-in session, and there was an example raised there by a colleague: a young girl logged on to a static website that had suicide content on it, and then she was bombarded with emails. Sadly, the story did not end well. Again, that is not defined as social media. That is one question that I really want to pin you down on.

Secondly, I am concerned about your answer about parenting responsibility and the state doing that instead of parents. A survey said that 50% of parents would not comply with this ban, so you would create an in-group and an out-group—those who have access and those who do not—so there would be huge issues with bullying there. What is your opinion on parents putting their children on social media, i.e. influencers using their children for that purpose? If a child cannot access social media, should parents be allowed to put their children on social media?

Frank Young: There were a lot of questions there. Just to be very quick, on the issue of parents putting their own children on social media, clearly, there has to be freedom for parents to make their own decisions. I might make a different decision from another parent.

On the point of parents not complying with any ban, that data was taken at a point of time when a ban has not been enacted. Polling will only ever tell you what people think right now. It is a snapshot. Once you change the culture and expectations around social media, you will start to see a change in behaviour. There is plenty of evidence from history of culture change being led by legislative change. There are plenty of social reforms where that has demonstrably been the case, and I suspect this will be another one.

Q20            Dr Gardner: When you present polling as evidence for a ban, you cannot then negate polling as evidence where a ban might not work.

Frank Young: Anybody who conducts polling for a living, as a mechanism, will understand the parameters within which it sits. This applies to voting intention polls, for example. The same conversation often happens.

Q21            Dr Gardner: On the social media definition, can I pin you down on this? This is a huge area. You have said you do not want to ban the internet. I have given some examples where services would not be classified as social media but could be classified as delivering harms. What do you mean by “social media”? How do you see a ban being legislated?

Chair: We are going to have to move on. Can you answer that quickly?

Dr Foljambe: It is more age-gating—raising the age—for appropriate reasons. That is the safeguarding positioning, I would say.

On the technicality of defining what constitutes social media and how AI or gaming fit in, it needs a broad scope and a list of definitions so that you can go in and target who is potentially in breach. Those decisions need to be made, but that can be made within the remit of raising the age. It can involve all user-to-user messaging. That might be through Fortnite, Call of Duty, Minecraft and all those sorts of things. It can be anything where strangers can access our children. That user-to-user definition does cross the divide over various AI, gaming and social media platforms.

I am a doctor, so this is out of my lane, but we must be able to figure that out. What we are really saying is that its current format is not safe for children. We are age-gating it, not banning it, like we do gambling, alcohol, smoking or sex. This framing is very frustrating. It is like we are spinning around in circles. I just feel that we can do it.

Q22            Freddie van Mierlo: You have spoken very eloquently, Rebecca, on the behavioural and mental harms of social media use. I just wonder whether, as a physician, you could speak very briefly to any physical harms that about their eyes. There are particularly high rates of myopia in countries where social media and screen use is very high. It is almost 90% in South Korea among high school children. I am also thinking about physical activity as well. I wonder whether you could just speak to that for a minute or so.

Dr Foljambe: Yes, of course. We have an obesity epidemic in children, which is not being made better by the fact so many children, especially from deprived backgrounds, are sitting for hours and hours on screens. We are actually seeing that in clinical practice. There are cases such as an eight-year-old coming in with peripheral oedema because they had not moved for 12 hours due to gaming. There are various physical manifestations of children sitting still for hours and hours when they should be moving. Usually, they are on screens. We are seeing that a lot.

Myopia—short-sightedness—is genetic, but there have been huge increases in rates of short-sightedness. Huge meta-analyses have confirmed that this cannot be down to genetics alone. They have found that for every extra hour of digital screen use, the risk of developing short-sightedness goes up by 20%.

We are not really joining the dots here. We need a public health campaign to educate and support families. It is not just that these children need glasses. They are at risk of cataracts, glaucoma and retinal detachment. In severe short-sightedness, they are at long-term risk of losing their sight completely.

When you bring that into a quite deprived rural environment, where I work, those children very often are not taken to their eye appointments. They are wearing out-of-date glasses. They are in chaotic families with low incomes, and they are often out of school. What are they doing? They are at home just sitting on screens. The impacts on speech and language are profound. There is very robust evidence coming out on the fact that we have completely lost the language-rich environment and in-context world that a child must have from birth in order to develop their language fully. That has been haemorrhaged on to screens that cannot perform that function. There is more to say, but I will stop.

Chair: That is very interesting. If that evidence has not already been submitted to the Committee, it would be useful for this session and for our inquiry into digital devices and childhood.

Q23            Kit Malthouse: There is now a growing or very significant body of correlative evidence. That is now translating, as you said, into causal evidence in research. I just wonder whether, as a doctor, you can confirm something about the reason why 16 is a good age. My understanding is that there are two key functions that develop in the brain prior to 16. One is the limbic system, which is about emotional response and particularly a reaction to peer approbation or otherwise. The other is the reward system and the response to dopamine, which develops between 13 and 16 most strongly, albeit, as you said, the development happens afterwards. That is why that particular age is particularly susceptible to this combination of peer review and a sort of trigger reward. Is that your understanding of the science?

Dr Foljambe: Yes, that is absolutely spot on. They are particularly susceptible, so it is even more tragic that they are being exploited and exposed to these harms at the time when they do the most harm.

To your point about the part of their brains that is developing in that period, it is what child neuroscientists, of which I am not one, describe as “use it or lose it”. When they are doing whatever they are doing in their environment—the spectrum of activities in their day—they are building, myelinating and strengthening neuropathways and neuro-lingual pathways.

Kit Malthouse: I am just conscious of the time. We all just voted not long ago to ban vapes for kids, and the argument was put that it will make them more attractive, so I get that. The other thing I wanted to ask is about an emerging area of concern and whether you had heard this from parents as well. There is research now emerging around what they call techno-ference, which is about the interference of these platforms in the relationships between parents and children.

We are now talking about a generation who have been through this who will themselves become parents. The parental obsession with social media is interfering with early attachments with their children. Parents of current teenagers are finding that the conflict caused by the discussion over social media is causing relationship breakdown. Lots of parents tell me that the only source of conflict in their house is screen usage and social media.

Emily Darlington: I wish.

Q24            Kit Malthouse: That is pushing adolescents more towards their peers online because of this breakdown. Is that an area of concern to parents as well?

Frank Young: It is. There are two points here. We have evidence from our polling of parents that it is screen time and being on social media that is causing conflict within the family. As somebody who leads a parenting charity, I would also say that there is a responsibility on parents, too, including myself, to model good behaviour. We have to leave space within this conversation, particularly those of us who have jobs where we might be on our phones quite a bit—such as a Member of Parliament, for example—to model that good behaviour. There should be space for saying that that is a responsibility, too.

Chair: Yes, absolutely. I am sorry. We are going to have to leave it there. We are over time, but it has been absolutely fascinating and wide-ranging. We have talked not only about social media but screen time and the other impacts of digital devices. Thank you very much for your evidence this morning, Frank Young and Dr Foljambe.

Examination of witnesses

Witnesses: Lorna Woods, Baroness Kidron and Kim Sylwander.

Chair: Welcome to the second panel in today’s session on social media and age restrictions. I am going to ask the witnesses again to introduce themselves in response to the first question, which is going to be put by Allison.

Q25            Dr Gardner: Welcome. We just had a very interesting session talking about why a ban is a good thing to do. I have some concerns that our definition of social media is still very diffuse, and with that lack of definition may be risks from this very well-meaning plan to ban under-16s—there might be unintended consequences and negative impacts. What are the possible negative consequences of a ban on social media for under-16s?

Baroness Kidron: Thank you very much for inviting me. I should say at the outset that I am an adviser to the Online Safety Act Network and chair of the Digital Futures Commission. I did not want that to go missing.

I was very struck in the last session by the Chair asking, “better regulation or ban?”. What I heard in all the questions is the failure of enforcement of the Online Safety Act. That is the context for anything we say.

While we are talking about definitions, I want to say that it is normal for products to have to be safe to access markets, citizens and consumers. We have to be careful about whether we are talking about a ban of children from something or of something to children—and I agree with your problem about definition. Why are we letting these avaricious companies, which are exposing our children to and creating harm, access to our children?

Having made both of those broader points, if products are not safe, they should not have access to children. I liked the questions that other people asked about the age going up to 18. I think ages 16 and 17 are very vulnerable ages; it is when you have least supervision and most peer pressure and, indeed, you are exploring broadly.

The key point that you make is, “What are we banning?” Are we going to ban last year’s problem, or are we going to start looking at this year’s problem, and what the hell are we going to do about that Online Safety Act?

Professor Woods: I am Professor Lorna Woods, professor emerita at the University of Essex and, as Baroness Kidron indicated, an adviser to the Online Safety Act Network too.

I am a lawyer by training and I love definitions. You are right: the devil is going to be in the detail in terms of scope, because scope affects both the effectiveness and the side effects of any measure that is taken. We are not just talking about bans here, but any regulatory regime.

The Australian regime, as I understood it, excluded services such as WhatsApp and quite a lot of gaming. This has an unfortunate effect. In a desire to be proportionate in speech terms, they may have undermined the regime’s effectiveness. This is not a question about children going off and having to search the dark web for obscure sites. They are going to where they were already, where the problems already were.

There is a real question about definition. We do have one. You have used the language “user to user”, which is the terminology used in the Online Safety Act, which does focus on functionality. As you have also remarked, it excludes email. Email on its own is not a problem, but email as a supportive tool can add to the issues.

Q26            Dr Gardner: To bring you back to the question, you have outlined how difficult it will be to manage the scope. Even looking at functionality, we can create arguments as to where that goes. Assuming that a ban, or a delay, as Australia calls it, of some sort comes into action, perhaps using named platforms, as Australia has done, what are the possible negative consequences of that? If it happens, how could we mitigate against them? I am sorry to interrupt, but I am conscious of time. Dr Sylwander, do you to pick up on that? People then can jump in.

Dr Sylwander: My name is Dr Kim Sylwander. I am from the Digital Futures for Children centre at the London School of Economics. I come here from the perspective of having worked for a long time researching children’s online experiences.

As I am sure you are all aware, there is a big group of researchers who are not on board with a ban, or what is often called a ban, at the age of 16. One of the reasons is that there is a risk of pushing children to unregulated spaces, both spaces that are more easily at hand and, of course, the darker spaces of the internet, rather than spaces that we can regulate more easily. That is one of the crucial arguments to the risk of implementing a ban.

As you say, the definition issue is a big issue. The other issue is, “Will it be effective?” Today, we essentially have a ban for under-13s. When we look at Ofcom’s research, we know that 82% of children aged 10 to 12 somehow access social media. If we implement a ban at 16, will there be anything that says that we will not have the same kind of figures for under-16s? It would introduce some kind of friction. We do not know the level at which it would actually protect children, because we have very limited evidence on how bans work.

Professor Woods: For me, there is a concern that, if you focus on a ban, you lose focus on enforcement of the Online Safety Act. Even if we have concerns about how far it goes, it is not helped by moving the regulatory focus elsewhere. My preference would be to try to make at least some products safer for all age groups.

Q27            Dr Gardner: You guessed my next question, which is fantastic. Thank you for that. I will cite some evidence. Some 50% of parents said they would not comply with a ban, which would create in-groups and out-groups. We have the cliff edge problem that the Molly Rose Foundation put forward. Is there evidence that a ban would prevent other interventions designed to promote safety for all users? We have had some talk about neuro-development up to 21 and the impact on women. This happens at all age ranges. In that sense, is there evidence that it could prevent other interventions to make companies produce safe products? Could a ban at least be an intermediate step to making them make those safe products?

Baroness Kidron: I am sorry to be difficult about the word “ban”, but I am going to continue to be difficult about it. If the Government do not do something, parents are going to be very disappointed. I really want to make this clear. You heard from a doctor. You heard from a parents group. I have evidence in front of me that I intend to leave with the Committee from the police. I get calls from lawyers. Teachers are incandescent. Everywhere where the adult world intersects with children, they are crying out for help.

This Government, and to a degree the previous Government, have failed to enforce and to fill the gaps of the Online Safety Act. Everybody has fallen out of faith with Government to do this. This is the problem. Let us not concentrate on whether the ban is the best way. The question is, “What do we do now?” The consultation is glaring evidence of government failure because it does not really deal with enforcement. It is managing the call for the ban rather than managing the call to make the internet safe for children. This is the problem.

I have never supported a ban in my life, but I have supported it in the Lords and I will support it again next Wednesday. I hope that we will bring back something that is more conditional; i.e. you may not have access to our children unless A, B, C or D.

I can give you the example of live streaming. Live streaming is on the risk register as a really important risk. It is not in Ofcom’s code, which means that any company in the world can have kids live streaming with impunity and still be compliant with the Online Safety Act. Last Monday, while the Commons was voting against a ban, I got some evidence from the police about live streaming on TikTok, which makes for such poor reading that my parliamentary assistant said he felt rather sick and asked to go home.

To conclude, just to make myself utterly clear, it is not about whether a ban is necessary now; it is that something is so necessary and so urgent. If the Government are not going to move but will keep on consulting and consulting and consulting, let us do something that is not tone deaf and lets parents, doctors, lawyers and police know that they are hearing them.

Dr Gardner: I am going to move on to age verification in a second, but I just want to clarify something. I hear what you are saying, but I also hear that there are real issues about whether it is going to be effective and about the scope. You raised this issue about enforcement, which I fully understand. How would you like to see enforcement work? Fines are fine, but when you are dealing with companies such as Facebook—

Chair: Allison, we cannot go into this. We have done a whole inquiry on social media and algorithms.

Q28            Dr Gardner: This is about the social media business model, not the algorithms. They factor fines into their business model so it does not really implement them.

Baroness Kidron: Fines do not work. I am happy to say that. We have to get rid of the safe harbour. We have to have business disruption measures and injunctions as a first port of call to bring them to the table. We have to have an individual redress system, at a minimum.

Q29            Dr Gardner: I will support the last one most certainly; I have called for that for many years. Very quickly, if we have this ban or this age limit, we need to have age verification. In a past life, as an academic, I audited Yoti’s algorithm for age verification, so I understand that, and I am chair of the digital ID APPG. What are the implications of age requirements for privacy and information security online? Is it possible to have age assurance measures that are both robust and privacy-preserving? Is there another approach where we could put age verification at the level of the app stores? That might simplify the process.

Professor Woods: I would like to jump in and add some detail to what Baroness Kidron said about live streaming. In Ofcom’s consultation on additional safety measures, it did add some controls around live streaming. For children, it consulted on whether audience interaction should be limited but not whether live streaming itself should be limited. It saw the harm coming from strangers interacting with children—giving them gifts and making comments.

On age verification, I am not a technologist so I am not going to step beyond my remit. It should be possible, but there is a need for clear standards so you can allow a clearer line of enforcement. At the moment, Ofcom has come up with principles for highly effective age verification, which seem relevant and legitimate but allow wriggle room. If we were referring to international standards, which pointed to the privacy and security aspects, we would at least have a baseline for that.

Baroness Kidron: Very briefly, the IEEE has a standard. The ICO is supposed to hold companies to account, but they are not being sent from Ofcom to the ICO. The ICO is not fully doing it.

The answer to your question is yes, yes, yes. You can have privacy-preserving age assurance. Age assurance is not perfect, but it is absolutely possible. You can double-check from the difficult ages of 16 and 17 up to 21 or 22.

Chair: I am sorry to move us on quickly, but we have so much interest here.

Q30            Dr Sullivan: We want to ask about the views of children. How have the views of children influenced your thoughts on where you are at the moment?

Dr Sylwander: They have absolutely influenced my position on this. Coming back to what was discussed in the first panel about the benefits of social media, one problem is that we have not adequately studied what the benefits are. There is a bias in research to looking at the negative, which means that, yes, it is true that we do not have robust evidence to say what the benefits are.

The evidence that we do have, based on surveys with children and a lot of qualitative studies, is that vulnerable groups of children, even though they are the most exposed to harms online, report the most benefit from social media: accessing friends they might not otherwise; accessing groups where they feel safe; or finding information that their parents perhaps do not let them look for. Social media are important spaces. Regardless of what we think about the negative impacts, these are realities for many children, and they want to keep that.

In our consultations with children around the world and all over the UK, we hear again and again that they want to be online, first of all, and they want the benefits of social media, but they do not want the harms. They want us to regulate away the harms.

To build on what Baroness Kidron was saying, if we know what the harms are, if we know which features are harmful, why do we not remove them for everyone? A previous speaker was asking, “Would we accept it for ourselves?” We should not for adults either. We need to look at what is age appropriate for children, definitely, but we also need to see how we can remove the harms that we know are there.

Q31            Dr Sullivan: Will the children who are most likely to experience those harms be in favour of a ban and preventing them from that space?

Dr Sylwander: If we take the Australian setting as an example, there were a lot of child and youth groups that came forward and said no. LGBTQIA+ groups and minority youth groups came forward and said that they were not in favour of a ban because they were afraid that their safe spaces online, which they have established through social media, will be dismantled and they will not have access to them.

Again, regardless of what the risks may be, these are realities that children are facing. They want those safe spaces. We also have to ask ourselves, “How can we create spaces that are safe for children and that are not commercially driven?”

Dr Sullivan: That is going on to regulation.

Baroness Kidron: Over the last 15 years, I have done so many workshops with children. I do them regularly and I love doing them. Whether they are in Pakistan, Kenya, America or here, I am always surprised at how similar their concerns are because they are using the same products.

I would agree with Kim. Children want to be online, but they want it to be safer, kinder, less intrusive and less demanding. Repeatedly, they say that, if their friends were off, they would be more comfortable being off. You have to be very careful about just hearing that they want to be on, because you also have to hear that they would prefer to meet in real life but do not feel they have that option.

It is maybe worth mentioning the age-appropriate design code, which was the first piece of legislation in this area, and that that was not properly and robustly enforced. It was about data privacy, and 13 of its 15 standards came out of the mouths of children. I can tell you the name of every child behind each thing. I am sorry to bleat, but let us not keep on passing laws if we are not going to implement them. Let us implement, forcefully and simply, the ones that we have.

Q32            Dr Sullivan: That leads me very beautifully on to the gaps that currently exist in the regulation space. Where are the current gaps in regulation and enforcement to make it safe?

Baroness Kidron: We are sitting here talking about a social media ban. Chatbots and friendship bots are the new emerging harm. I just find it very hard that Minister after Minister comes to the Dispatch Box and says, “We cannot anticipate”.

Meanwhile, I will say this—I am furious about this—that the Government had a three-line whip last Monday preventing risk assessing for AI child sexual abuse on generative AI. That is shocking. It is shameful. Even though we lost the vote because of the three-line whip, the Government should do something about that.

What I would say is safe harbour, duty of care. Be broader and simpler. Let us get all of these things within the remit and let us have proper powers of enforcement. This issue about privacy and data is the root of everything. The reason they are bugging kids is that they want their attention. They want their data. If you go to kids’ privacy, some of the things about content and addiction just fall away.

Finally from me, during the Online Safety Act, we thought we had provisions around addiction. The Lancet Commission, on which I sat, is about to publish, and that is around problematic internet use. That is evidence that you might like to use. I really would urge the Committee to understand that the behavioural problem is the source of everything. Most of the problems of the internet can be addresseds by switching off the intense demand rather than the content itself.

Chair: I just want to make the point that the Committee, as part of its report into social media and algorithms, did issue a number of recommendations, which included addressing some of the behavioural issues through algorithms, advertising and incentives. The Government did not accept our recommendations, but the Secretary of State has said that she will be taking another look at them. We will certainly continue to press for their implementation.

Q33            Dr Sullivan: It feels like we want simple solutions. As has been raised before, how do you define social media as? I spoke to some children in my patch. We asked, “Who is on Roblox? Who has heard something nasty being said on Roblox?” and lots of hands went up. When we asked, “Who does not want that anymore?”, they said yes. It comes back down to the intent with which some of these things have been designed. If the owner and designer of Roblox, as alleged, will not let his children play on Roblox, there is an issue here. There is something about the behavioural intent behind these programs.

Baroness Kidron: There is. I understand that this is a quick look at the ban problem, but software is a product.

Dr Sullivan: It is their choice.

Baroness Kidron: Software is a product. Why is it that something that is deliberately designed to optimise for certain outcomes is not subject to the law in an appropriate way? That is where we have to go, ultimately.

Coming back to the Chair’s point, it is about better regulation. Is a ban on the way to better regulation? I do not know whether it is, but, if the Government consultation has not dealt with enforcement, chatbots, AI and all these other things, a ban will be useless. If a ban is simply, as Rebecca said in the last session, a raising of the age for certain features, that is fine, but concentrating only on social media is a mistake. It should be about services that children use.

Professor Woods: If I can interject, bubbling under this is the idea of product safety and safety by design. There is a reference in the Online Safety Act to products being safe by design in section 1(3). That was the work of the Lords. Unfortunately, there is not a clear definition in there. At least putting down some principles as to what we mean by safe by design might be helpful.

I had a look at some of the other areas that require product safety. The first principle is to design to reduce the existence of the hazard in the first place. It is only when you have exhausted that as a possibility that you go on to mitigating and managing the hazard; remediation is the last stage. You have a range of ways of tackling the problem, but there is a hierarchy.

Chair: Emily has a question relevant to that.

Q34            Emily Darlington: You have hit a really good important point. What do we mean by safety by design? My colleague Tom Collins, who is a former BSI official, and I are doing additional work in that area, because there should be a set of standards about what that is. I talk about BSI a lot because it is based in my constituency and because I think it is amazing.

I want to talk about other solutions. Other solutions have been proposed not in isolation but as part of the way of making the online sphere safer for children and adults. One of the proposed solutions is quite an old solution, which is to give apps ratings like the BBFC ratings or the video game ratings that are known and understood by parents, children and adults alike across this country. Can I get a reaction from all of you about whether you think this could be part of the solution in terms of making sure content or apps are appropriate for the audiences that are currently consuming them? Kim, do you want to start?

Dr Sylwander: Every time I meet social media companies, I am struck by how they are the ones that are self-assessing their risks and they always find themselves to be low risk, surprisingly. Your example of the BBFC is a good example. There we have a system through which risk and age appropriateness are assessed, and there is also periodic consultation and review with the public to make sure that we are sticking to the times and understanding culture as it is now, not as it was 20 years ago.

There is a point to that model. How it could be applied in this case is definitely something that needs to be considered. Again, we should not leave it to companies to decide whether they are compliant or not. It needs to come from Government. When it comes to the best interests of the child, that is language that is co-opted by big tech companies. They feel like they are acting in the best interests of children, but it is not up to them to decide whether they are or not. Yes, I would say such a model is definitely interesting.

Q35            Emily Darlington: I have asked YouTube repeatedly what the rating is for YouTube Kids. I just want to know whether it is appropriate for a five-year-old, a 10-year-old, a 13-year-old or a 16-year-old? I do not know. They do not seem to know either, which is a bit of a concern and a worry when we are talking about potentially creating safe places for children. YouTube Kids is one that parents trust. They say, “Okay, I will let my child only watch that. That will be protected”.

Beyond YouTube Kids and the other children-only apps and spaces that most major social media companies and online companies are developing, how would you assess things such as addictiveness or feeding with an algorithm, one that generates what your next choice should be versus—maybe it is old school—picking your own next video to watch from a selection, or in-app purchases and product placement, where we have quite strict rules around what would be appropriate on streaming platforms and television at the moment?

Dr Sylwander: They need to be considered from an age perspective and from a developmental perspective. On personalised feeds, for example, there is no need for them. Considering that they are connected to commercial interests, having that aimed at children is just nonsensical. We know that captures their attention, keeps their attention focused and keeps them locked in. Those risks definitely need to be age assessed and limited for children under 18.

Baroness Kidron: It is a really interesting question. The bottom line is that anyone looking for a silver bullet should stop looking. It is very helpful for parents to have, “This is for 12”, “This is for 11”, “This is for 17”. I absolutely agree with Kim that it cannot be content-based. It has to be based on the experience. The biggest part of the experience is how compulsive or exposing it is. It is a tool in the toolshed.

I am very much more interested in some of the standardisation issues that you have put forward. There are a lot of things that we could standardise. Some cigarettes are stronger than others. Where the bar is I do not know, but there is a difference. I am not a smoker, but you can tell on the packet. There is something in there.

To the point that Kim made, we need to stop treating tech as if it was a responsible adult or—let me put it more strongly—a responsible parent to our children because it is not. Again, I am going to go back to the age appropriate design code. YouTube turned off autoplay for children as a result of the age appropriate design code. It did a number of very important things at the time, but we did not see the regulator go in and say, “That is our bottom line”. It let everybody do their own thing.

Q36            Emily Darlington: I wanted to just raise one point because it is really important. You introduced me the other week to the parents of a young man who had taken his own life after becoming addicted to a companion bot. This is something that has not really hit the UK in quite the same way as it has hit the US, but it is here and we absolutely need to be aware of it. This is a wider question about future-proofing. For those on the Committee who might not know, companion bots are bots that become your best friends. They are sycophantic. The most important thing they do is keep you online because they are collecting your data. To do so, they cut you off from everyone else. They use all the phrases that we understand from coercive control and grooming.

When it comes to a technology that is as corrosive as that, is there any amount of regulation that could prevent that being damaging? Is that an area that you would like the consultation to consider and potentially say, “No, absolutely not for under-18s” or whatever it might be?

Baroness Kidron: Let me briefly say that we have an amendment to stop it in the Lords next Wednesday. I would like the Government not to wait for a consultation. You say it is coming to the UK, but I have parents in touch with me who are worried about their children. It is an emergency.

Lorna has done more work on this, but future-proofing the Online Safety Act and raising the age more broadly across products that are unsuitable would really help us out. We cannot keep consulting. It took eight years to do the Online Safety Act. It has taken another two years to bring forward its codes. We cannot afford to wait. Some of future-proofing is about being bold and saying, “Sycophancy is not healthy for anyone, but we will not allow it for children”. They are parenting our children. That is the intervention.

Emily Darlington: She is very brave to mention sycophancy to a bunch of MPs.

Chair: I am afraid we are going to move on now because we are running out of time. Baroness Kidron, you talked about not anticipating future harms. All these harms have been anticipated. I worked for the regulator Ofcom before coming into Parliament. They have been anticipated. They can be anticipated. The point that Emily makes is that they are being anticipated and Government action can be pre-emptive rather than waiting for the market to address a failing that it does not address.

Q37            Adam Thompson: Good morning, all. I want to bring to the debate the idea of a social media ban versus raising the digital age of consent as two options that we might be looking at at the moment. What are your thoughts about the differences, in practical terms, between introducing a ban, as it were, and raising the digital age of consent?

Professor Woods: It depends on what you mean by “ban”. Baroness Kidron has said that we probably need to be a bit more precise. That is right. We could say something like, “Let’s require companies to enforce their current minimum ages”. That is one form of an age restriction. We can put a minimum age in ourselves through legislation or we can just say, “Let us take this product off the market until it is proven safe” or, “Let us just take this off the market”. There is a range of options there.

Q38            Adam Thompson: What are your thoughts, though, about this idea of changing the digital age of consent? I understand that it is currently 13. That could be moved to 16. That might be, on paper, different from a ban, but what would that look like practically?

Professor Woods: I am not sure what people mean when they say “raising the age of digital consent”, because it is usually related to data protection and the ability to consent to the processing of your data. In a way, it is coming from a different regime and addressing a slightly different problem. You may end up having one measure around minimum ages that does multiple things. You could use it to say, “This is about consent for data protection purposes, and this is also about your access to particular services”. You could use it as a multi-purpose tool.

Baroness Kidron: Lorna is absolutely right that those two ideas come from different regimes. What people who are talking about raising the digital age of consent are possibly trying to grasp at is that they do not feel that the current system is protecting children for long enough. You have heard evidence before—and I have lots of evidence here—that under-13s are not protected either.

If the data regime is not working properly because it is not being enforced, the online safety regime is not being enforced properly because we still have all these problems, and the terms and conditions of the companies are not being properly enforced, what we have is an enforcement issue.

What I like about the conceptual thought of raising the age is that we are saying, “What is suitable at what age?” The Online Safety Act does talk about evolving capacities and different ages and stages. It is Ofcom that decided to do 18 or under and nothing else. We have in law already the opportunity to consider as a society what we think is suitable at different ages and stages, and not just about content but the experience. That is where I would like to see us going.

Q39            Adam Thompson: On that point, with the previous panel we talked about the age specifically. Let us accept that 16 is a good age for us to bring in this threshold at, if we decide that we wish to. There are concerns, from some and maybe not others, about the potential for a cliff edge around 16 or 17 years of age. My understanding is that those cliff edge concerns are often tagged on to a ban but maybe not the digital age of consent. Lorna, do you have any thoughts on why those concerns are in one space and not in another?

Professor Woods: I am not entirely sure. If you are using digital age of consent as a threshold condition, you have a threshold, do you not? The underlying point is about why we have we created a cliff in the first place rather than an age-appropriate slope, if I can say that. We need to look at standards for adults as well. I do not accept that adults are bulletproof. If we have a generally better environment against which you then set higher protections for children, you are also potentially reducing this big difference in standards.

Adam Thompson: That is a really important point. It is common knowledge that, as MPs, we get a lot of online abuse. In my own experience, I am certainly not immune from the effects that we are talking about today. The point is very valid. Thank you all very much.

Q40            Chair: We are running out of time. I just want to ask one last question about children’s rights. We have talked about the child’s perspective. I have sat as a guest on the Joint Committee on Human Rights between the House of Commons and the House of Lords, which is conducting an inquiry into AI and its impact on human rights. The UN Convention on the Rights of the Child includes rights such as access to information, freedom of expression and privacy. Just briefly—it is an important question—what impact do you see an outright ban having on children’s rights? Is that a matter of concern for any of you? Who would like to respond on that?

Professor Woods: As the human rights lawyer, I will have a go. There is an impact on freedom of expression, whether you look at the European Convention, the ICCPR or the UNCRC. There is an unseen problem, though, with the justification. We all know freedom of expression is a qualified right. Interference can be justified on the basis of a legitimate interest, where it is based on the law and where it is proportionate.

The Australians drew the scope of their regime with a concern about going too far in terms of cutting out all forms of communication. There is another danger. If you have a regime that interferes with speech but is ineffective in its aim, it too is disproportionate. You can run into problems of, in a way, not going far enough.

Baroness Kidron: I had the privilege of chairing the drafting group of general comment No. 25, which explained the relevance of the convention to the digital world. Kim could talk more about it. I would just like to draw to the Committee’s attention to freedom of thought. That is the ultimate freedom for a child. If you are being nudged this way and that way, you are not free in your behaviour and you are not free in your thought. I would just challenge the idea that social media is the only way to get information. The digital world is far broader.

In my last sentence to the Committee, I would like to just say that people are desperate out there. Parents are desperate. Kids are desperate. There were seven hours of programming about online harms on LBC on Monday, and not one person rang in and said, “Let it be”. We need to deal with this.

Chair: Thank you very much, Baroness Kidron. You have made that point very clearly through this session. Thank you very much to Professor Lorna Woods and Dr Kim Sylwander for your evidence and for joining us this morning. You have given the Committee a lot to think about.

Examination of witnesses

Witnesses: Julie Inman Grant, Professor Jeff Hancock and Professor Amy Orben.

Q41            Chair: We come to our third set of witnesses in today’s session exploring social media restrictions. We are now looking at restrictions in Australia, how they are being evaluated, and to get an overview of the research evidence. We are especially grateful to Jeff, who has travelled from the United States, to Julie, who is joining us virtually from Australia, very late in the evening, and also, of course, to Professor Orben.

I will ask the members to introduce themselves in response to the first question, which I will put to Julie Inman Grant. Can you just give us an overview of how successful the delay—as the terminology in Australia is—to social media access has been so far?

Professor Orben: I would say that it has been very successful in its early days. We have about 2.5 million eight to 15-year-olds. Within the first two days, we had 4.7 million social media accounts of under-16s deactivated or restricted, depending on how the platform was configured.

Q42            Chair: Sorry, could you just introduce yourself?

Julie Inman Grant: Yes, I am sorry. I am Julie Inman Grant, eSafety Commissioner of Australia.

Q43            Chair: How many accounts were deactivated?

Professor Hancock: It was 4.7 million within the first two days, based on the legal notices provided to us. All of this is going to be contextual, you understand, Chair. We are now in the second phase of the process, which is probably one of the more difficult phases. This is where the platforms that are covered by the social media delay need to prevent circumvention, either through spoofing of their age assurance systems or through other forms of circumvention such as VPNs.

They are responsible for preventing that circumvention. Of course, like any age restriction or limit, whether it is alcohol, smoking, vaping, driving and speed limits, there are definitely people, particularly creative teenagers, doing a good job at getting around these. Again, the burden is on the platforms to keep under-16s off the social media sites that are covered.

We are going well. Some companies are doing better than others in terms of compliance, given the preponderance of under-16s on there. We are using a layered notice process and high-level talks with their different technical teams to make sure that their age inference tools are replicating throughout the system, and that they are implementing them effectively, particularly the third-party age-assurance technologies.

Chair: Could Members introduce themselves? I am not sure the commissioner can see.

Q44            Freddie van Mierlo: My name is Freddie van Mierlo. I am the MP for Henley and Thame in Oxfordshire. Thank you, Julia, for joining us at this time. I think it is around 10 pm where you are, so I really appreciate your joining us. In a moment I want to come to your view of what the benefit of this delay or ban in Australia is, but I did read in the brief we had beforehand that you previously worked at Twitter, now X. It is now your job to protect Australians online. What do you think they need protecting from, given your experience working in social media?

Julie Inman Grant: I did work in the technology sector for 22 years, and only two years at Twitter, when it was under different ownership. There are a range of harms. We have a pretty sophisticated research and evaluation team, which is working with both Professors Hancock and Orben on building that evidence base. Our recent work found that 70% of young people encountered various harms on social media, as well as across the broader internet on gaming and messaging platforms.

We also have four complaint schemes that you would have read about in the submission that I put forward to you. We take cyber-bullying reports from children, and we have for almost the past 11 years. We have a lot of data in terms of how that is manifesting and changing. We have an image‑based abuse scheme; we have a 98% success rate in terms of removing intimate images that are shared online without consent. That includes deepfake image-based abuse as well. We are the hotline, like IWF is. We collect child sexual abuse material reports, pro-terror content, and then adult cyber-abuse.

The threat surface is changing. In my first eight or nine years we saw image-based abuse, cyber-bullying, and child sexual exploitation and grooming. We have seen about a 1,300% increase in sexual extortion reports, and that is largely targeting young men between the ages of 16 and 24. We are also seeing young people encountering a lot more online hate and a lot more misogynist content. We are doing a lot of work around being a young man online, and what they are experiencing with the manosphere. We are also seeing more things such as violent fight videos. The threat surface is changing, but unfortunately the harms that have been there for the past 30 years that there has been an internet continue to thrive, sadly.

Q45            Freddie van Mierlo: Thank you for that. The harms you have identified there are really wide‑ranging, so I am really interested to know what regulatory and legal framework you have in Australia and the firepower that your office has that enables you to enforce this delay, but also to protect Australians online more widely.

Julie Inman Grant: I would say—and this was echoed in the last panel—that we do refer to it as a social media delay, because it is not an absolute ban. It is not an internet block of social media sites. It also has exclusions for messaging and gaming platforms. A primary reason for that is that some of our prior research showed that young people in particularly vulnerable communities like LGBTQI+, First Nations or those with a disability felt more themselves online than they did in the real world. We did not want to cut off their digital lifelines, effectively, or prevent them from creating and exploring the broader internet.

The other reason we call it a delay is that we are trying to prevent young people who are just forming their identity, their personalities, their impulse control and their critical reasoning skills from encountering the harmful and deceptive design features that we all know about, whether it is snap streaks, endless scroll, algorithms and the like.

The delay also comports with the way that parents speak to their kids about technology. We do not say to our kids, “You are never going to have a smartphone or be online”. We say, “No, you will be on social media someday, but when you are ready. The Government say that is the age of 16, so you will not be excluded, because this will apply to all of your friends”.

We need to flip the script from saying that we are trying to prevent children from accessing technology. Really, we are trying to prevent social media companies that have become quite predatory from accessing our children. I just wanted to get that framing in there for you and, because I went off script, I forgot your question.

Q46            Freddie van Mierlo: You have answered it in part, which was around the regulatory and legal framework that you are working within. One of the things we heard in previous panels was around the fact that we do have some regulations in the UK in the Online Safety Act, but we struggle to enforce them. Ofcom, our regulator, has faced a lot of criticism in recent weeks around its enforcement of that Act. Could you give us an idea of the regulations that you have that enable you to do your job, and maybe the size of your office and the firepower that you have as a commissioner to enforce the rules?

Julie Inman Grant: When you are looking at something like this, it is technical regulation meets social regulation, which is very, very complex. You are dealing with some of the most powerful and richest entities in the world, which will gladly use jurisdictional arbitrage, deep pockets and expensive lawyers to get around compliance. I have experienced quite a bit of that.

Under our Online Safety Act, we were established in July 2015. We started with 30 people. We have expanded about tenfold, to close to 300. We are still much smaller than Ofcom, but we take a holistic approach. No one solution, including the social media delay, is a panacea.

We have the complaint schemes. We have built quite a robust set of prevention and education initiatives that are all evidence-based over time, and co‑designed with the different communities that we are working with. That is really important to arm parents, carers, children, and women who might be experiencing technology-facilitated abuse, whether it is online harassment because they are in the public eye, or whether it is technology‑facilitated abuse as an extension of domestic and family violence. We have the complaint schemes, which are very effective for providing really immediate or short-term harms remediation.

Our systemic powers have been rolled out gradually. We have a co‑regulatory code scheme, which means that the industry writes them, but we do not allow companies to mark their own homework. They have to meet appropriate community safeguards. If they do, I register them. If they do not, then I move to a standard. We have had the unlawful standards in since June 2025, and we have used it to good effect against blocking nudifying sites, some of the most popular ones that were contributing to deepfake image‑based abuse. We have used them against chat roulette sites, including OmeTV based in Portugal. We have used our app store code; as a result, Apple has deplatformed more than 100 chat roulette sites, which pair predators with children through video chat.

We are using different gatekeeper services as well. Search engines now have to blur pornography or explicit violence when someone is searching in a logged-out state, or to refer people who are looking for explicit instructions in suicide to a helpline, rather than down a rabbit hole.

This week we deployed our age-restricted harmful codes. These also apply to AI companions and chatbots. No AI chatbot or companion company can serve suicidal ideation, self-harm, disordered eating, violent explicit content or any sexually explicit content to an under-18. Again, this applies up and down the stack.

Our codes are much broader. They cover everything from porn sites, to gaming sites, to search engines, to devices, to app stores, and also generative AI services.

We have transparency powers, and what I would call our proactive and systemic change. The safety by design work and the initiative we developed in 2018 is moving from strength to strength. We do technology future casting. This is how we are able to stay ahead of the technology curve in terms of getting these regulatory provisions in before something becomes a catastrophic harm.

Q47            Chair: Did you look at other methods of testing the impact of the delay, such as a sequential roll-out?

Julie Inman Grant: It is not really a sequential roll-out. What we tried to do in our regulatory guidance is to break down their various responsibilities and what the sequence is. First off, you need to identify the under-16s and deactivate the ones that you are aware of now. Again, we are continuing to measure their progress.

One thing that we did do—and we have reserved the right to update our regulatory guidance—is use our transparency powers, but also a youth survey, where we worked with young people and parents to assess the baseline numbers. As I had heard someone say in the last panel, they were not even enforcing their policies that they already had in place for 13-year-olds. We assessed that 84% of eight to 12-year-olds already had access to at least one social media site in 2024, so it was worse than just trying to target the 13 to 16-year olds. I had to be explicit that it had to be everyone under the age of 16.

We gave them a waterfall approach. We said, “You cannot just use one technology. If you are using age inference technologies, you need to supplement that with some form of age assurance, and there need to be triggers, whether it is asking for privacy-preserving ID or using another form of age assurance such as connecting anonymously to banks. You need to be constantly improving”.

I decided not to pick winners and losers. We tested about 60 age assurance technologies, many of which are UK-based. The ACCS from the UK ran that set of tests for the Australian Government, and found that these technologies are effective, robust, and can be accurate and privacy preserving, but I did not also specifically delineate accuracy rates for them. I wanted to give the companies the opportunity, because they know their user base and their technology infrastructure better, to implement the solutions. This is an outcomes-focused approach, but if this is not successful then we will be more prescriptive.

Q48            Chair: Thank you very much. It is very interesting, comparing your approach with Ofcom. I want to bring in Professor Hancock and Professor Orben to give their views on whether they are confident they will be able to establish cause and effect in terms of evaluating the ban. I will start with Professor Hancock. Could you introduce yourself, as you get over your jet lag?

Professor Hancock: My name is Professor Jeff Hancock. I am a professor at Stanford University, where I am also the founding director of the Stanford Social Media Lab. I am working with the commissioner and her e‑safety team to evaluate the SMMA. One of the great things about the legislation there and the e‑safety focus has been on evidence. We are designing a very robust evaluation, where we are looking at how young people and families, including parents, experience the delay.

We started a survey that took place before us, and we have some baselines. We will then be continuing five additional waves, where we will be doing the same survey, so we can look at changes over time for the children’s experience around whether they are adhering with the legislation. We will be looking at their digital experiences. Are they changing their behaviour around digital experiences? Does this then affect the outcomes that people are interested in, for example anxiety, depression, and exposure to harms, and then, ultimately, family conflict. As we heard in previous panels, this is one of the central concerns that every family raises—this is one of the main conflicts. As Baroness Kidron mentioned before, almost all adults who interface with kids are concerned about this issue.

We will be assessing all those from top to bottom. We have articulated a theory of change, which looks at some of the perceptual aspects. What do people think about these bans? What are the other parents thinking? What will be these norms that will then potentially change behaviour? Ultimately, does it lead to different kinds of intended and unintended consequences? One thing the commissioner made clear to us is that she is interested in understanding both those intended consequences—getting the kids back out on to the footy fields, as the Australians also say—and the unintended consequences. Is this problematic for some of the young people or their families? That is a sketch of the evaluation we are doing.

Q49            Chair: Just to clarify, how are you identifying the young people that you are going to be observing or watching?

Professor Hancock: We are working with the Social Research Centre in Australia, which is its most elite survey institute. It has a sample of 4,121—approximately—children and their parents. We will be following those children and their parents for the next several years.

Professor Orben: I do not have anything to add beyond that on the ban evaluation. Jeff has summarised that well. I am just an adviser; it is Jeff who has been leading it.

Q50            Chair: Who are you an adviser to? Could you introduce yourself?

Professor Orben: I am Professor Orben. I am a research professor at the University of Cambridge. I lead one of the UK’s most impactful teams researching social media and its impact on young people. I was a director of the DSIT commission on how to improve research in this evidence space. I am leading the trial up in Bradford, where we are looking at social media curfews.

Chair: Great. Very impressive backgrounds all round.

Q51            Dr Sullivan: It is about learning from other countries. We have heard about the work in Australia and the social media delay—not ban. What other countries are also looking at this style of restricting youngsters’ access to social media?

Professor Hancock: This is the conversation now taking place around the world. The Australian law has really triggered a lot of these conversations. We are now tracking at least 14 countries that are interested in this. I am meeting here today with you all to provide evidence, but I will be meeting later with the French. I have spoken to the Canadians, the Irish, those in Norway and Indonesia. Many of them have reached out because there is an interest. “How do we do an evidence‑based approach? How do we evaluate these, right off the bat?”

Professor Orben: It is interesting to note why there is such a focus on evidence, because we do not yet have that much evidence about the effectivity of what a ban might look like. It is very interesting to note how much the evidence has been part of this debate. It is something that has struck me over the last couple of months.

Q52            Dr Sullivan: Do you feel like the tech companies are forthcoming in helping supply evidence?

Professor Hancock: My view on that has changed over the years. In the past I had done research with Facebook—when it was called that—as an independent academic, but now that has just changed. I do not really see them as reliable partners any more. I believe that is one of the reasons why we are seeing increased calls for things such as these legislations, which sometimes can be very blunt instruments. I believe that parents in particular feel that they just cannot rely on the platforms to be reliable partners any more.

Q53            Dr Sullivan: It is very interesting framing it as social media companies accessing our children, and they are not a responsible parent. Over to the screen—beaming straight over to Australia—how important is it to have a co-ordinated international approach to online and specifically social media regulation?

Julie Inman Grant: It is essential. Again, we are talking about regulating some of the most powerful companies in the world. Just being a small middle power in the south Pacific, we can have some impact and we can have them comply with our safety laws. After being the only online safety regulator for the first seven years, having Ofcom, the Irish, the European Commission, Fiji, Singapore and others come on board has been amazing. You might know that we started the Global Online Safety Regulators Network with the UK, Ireland and Fiji in 2022, and that has been growing from strength to strength. My hope is that we will have regulatory coherence for countries that decide to follow this path.

Again, we are going first. One of the reasons that we wanted to engage in such a comprehensive evaluation is not only to evaluate this for ourselves, but to develop an evidence base that did not exist when our policymakers made some of these determinations. A common question is, “Is 16 the right age, or should it be 18, 15 or 13?” A lot of the policymakers who developed this were avid readers of Jonathan Haidt’s “The Anxious Generation”. On that book, I will defer to the other research scientists in the room, but that is one of the things that we can look at.

I do think it will be challenging if there is a patchwork of ages around the globe. It will be hard for the companies to comply. In fairness to them, because they have been testing out their technologies, including their age inference systems on those 13 and under and those 18 and over, and because of the different levels of adolescent maturation during that 13 to 15-year-old period, it is going to take a while for their classifiers to really catch up and be perfectly accurate. A lot of the age assurance technologies, particularly coming out of the UK, are getting more accurate, but that is why we need a multiple-layered approach.

Q54            Adam Thompson: I am Adam Thompson, Member of Parliament for Erewash. It is good to see you all this morning. Amy, you were talking about evidence generally and how it is sometimes hard to find. Sometimes the evidence is not very forthcoming. In your experience, Jeff and Amy, how strong is the evidence for social media’s impact on under‑16s, whether positive or negative?

Professor Orben: That is the killer question. We need to think about many different types of evidence. It is like asking what the impact of online life is. It is so diverse. What helps me try to understand this complex system is to first think about harms or benefits on an individual level versus on a population level.

On an individual level, we have seen very strong evidence of harms; we have also seen strong evidence of benefits. Children are using specific types of social media to organise social change and protests in their schools, for example, or working together to build communities that they care about, but we have also seen severe harm that, in other products, would be enough to change or recall that product. If a toy has killed four children, you would request a recall. You would not really care about, “Well, how much does spending time playing with this toy relate to the wellbeing of the larger population of children?”

We also need to think about population-level evidence where we average everything together, because the question here is not, “Are we banning that specific toy?” but, “Are we banning all toys?” How do we go about weighing up those risks and benefits?

On the population level, that is where things often get a bit harder, because we do not have the right access to data from companies. Ofcom did write a report on the Online Safety Act. We have seen delays in getting that into practice. For example, do you talk about polarisation? Do you talk about mental health? Do you talk about eyesight? It is very varied, but that is where things are a bit poorer, especially when you think that a ban will manipulate time spent online. We know that time spent online is a really bad measure of impact. It is probably what you see, what it displaces and how you interact with people, but the ban is quite crude, because the lever is time. There we see that the evidence is even more mixed.

Harms naturally exist, and we need change. The question is on what level, and what areas to be able to weigh up. I see this not as a question about where the evidence is or not, but about what philosophers of science would call inferential risk. What is the risk of acting now on incomplete evidence? What is the risk of waiting, when potential harms might accumulate in the meantime? Also, what is the risk of putting in place a policy that has less evidence than there is with other forms of policymaking?

Q55            Adam Thompson: Thank you so much, Amy. I was actually going to ask some follow-ups, but you covered them off. I have one thing I was curious about for the record. Obviously I assume that your background here is academic journal papers, et cetera, but where is the underlying source of the evidence that those academics are bringing it in from?

Professor Orben: I often do rely on academic journal-published evidence as the foundation. That is often population studies, where you question people over time. Because tech companies do not share their data properly, you often ask a young person, “How much time do you estimate you spend online?” and then you track how they develop over the next few years. Rebecca Foljambe did mention some studies of my own on UK data. It is inherently limited, because we do not have a good exposure measure. We are not measuring whether they are seeing really harmful content. Are they seeing the horrendous things that Rebecca was mentioning, such as the “kitten in a blender” videos? All these do happen, and we do not have that accuracy to measure that.

More and more, what my DSIT report focused on is that those longitudinal studies are not enough, and we need to become more agile in our evidence creation. We can do that by going into the field, changing things and seeing what happens. That is where that real evidence limit lies, because we have not done that much on the different interventions we are discussing. That is important for us to learn, as well, to do better on AI.

Q56            Chair: Could you write to us and say what requirements you would like to see placed on social media companies in order to give you the evidence, whether that be APIs or access to algorithms, et cetera?

Professor Orben: We are writing an open report on that at the moment.

Q57            Daniel Zeichner: I am Daniel Zeichner, Member of Parliament for Cambridge. I am interested in the difference between the evidence base that is being used in Australia, America and here. Are there any differences, and are they significant?

Professor Orben: I do not think there are. We often interchangeably use evidence across countries that we would call the global north, which includes Australia and the UK. We do not have a study at the moment in healthy under-18s that has reduced or limited social media to see what happens. There is one in Denmark that did so, which did not find an impact on life satisfaction, but then there was one in the US that did limit social media for under-18s who had pre‑existing mental health conditions, which did find a positive impact. I do not know of any experimental work that has gone on in the UK, even though a lot of the longitudinal population-level evidence provision is on UK data, because of the quality of our data resources.

Q58            Daniel Zeichner: There is more work to be done. How could effective evaluation be built into the design of any proposed ban? That seems to be a really important point to finish on.

Professor Hancock: The Australians did a great job on that point. We have already started an evaluation framework, which could be really valuable. We have worked with the eSafety Commissioner over the last year to develop that evaluation framework. One thing that is missing—and this is a quote from business that applies here—is “measure what matters to you”. That is what we are trying to do in Australia, but one of the problems that us academics have is that none of us agree on what is exactly the best thing to measure or how.

Q59            Daniel Zeichner: In one sentence, then, what would be the key test of whether this has worked in two years’ time in Australia?

Professor Hancock: There is no one key thing to measure. That is the problem, but it is okay. What we need to measure are the things that the people of the UK in particular care about. In Australia, it is about kids’ experiences online, about their exposure to harms and their ability to flourish, but also these family dynamics, which are often missed, although that is really what is driving a lot of this.

We can develop standard measures, which we have already begun in Australia. We are working with the UK Home Office to put the same evaluation framework we have in Australia here in the UK. That will take place later this month. We will also have that going out in New Zealand this month. If we can standardise those measures around those different things I said that we need to measure, that will allow us to understand how different policies around the world are affecting children, but also the families.

Professor Orben: This policy question is so broad: what people would like a social media ban to impact. You have conversations with the Home Office about radicalisation, and with the Department of Health and Social Care about health outcomes. You might be thinking about myopia. You might be thinking about conflict in families. While we will start seeing evidence coming out of Australia in the next six months, we should be ready for there to be conflicting evidence. There are about 15 evaluations I know of that are independently trying to evaluate Australia. What success looks like is not clear.

It is also not clear what that would be in the UK. The important thing for us to consider is what exactly we want a ban to produce, and then to measure that appropriately. The best is to compare very similar populations that did or did not get the ban. If you can do a staged approach where only one age group gets it at a time, that can improve the evidence base, but naturally, for example, we could also compare young people in Northern Ireland with those living in Ireland, just across the border, who will experience potentially different regimes. Those national experimental policy evaluations will be critical.

Chair: Thank you so much. I am afraid we have to leave it there. We could continue to discuss this critical, fascinating and, in many ways, contentious subject. We have learned a lot from hearing from you. We particularly appreciate your contribution, Commissioner Inman. We look forward to seeing the evaluation. Your input today will influence our input to the Government’s consultation that is ongoing here, so a huge thanks for your contribution. I am sure we will continue this discussion at another time.