Social Media Restriction - Should AI be Next?
On the 14th of July 2026, The Oxford Cambridge Forum for Flourishing, Learning, AI & Research (OXFFLARE), a collaboration of Generation AI & LOGOS hosted a special session to explore the role of bans in regulating social media & AI.
We are grateful to Dr Ayça Atabey, Sarah-Jayne Blakemore, Natalie Boll, Maxime Le Bourgeois & Kim Sylwander, PhD for their invaluable contributions to the discussion. Thanks too to Ann Kristin Glenster, Ph.D., FRSA & Megan Ennion, our OXFFLARE collaborators, for co-convening.
For more than a decade, concerns have been raised regarding the impact that social media is having (and has had) on young people. Challenges around bullying, exposure to harmful content and the potential for grooming have gone largely unaddressed in that time. Now, after years of sustained concern from parents, researchers and young people themselves, governments are starting to act. Their solution, banning and restricting children from accessing social media. In the UK, this applies to under-16-year-olds and also extends to features such as live streaming, communication with strangers and AI chatbots providing sexually explicit interactions to under-18s. Early evidence from Australia, which implemented similar restrictions last year, indicates that a majority of children have retained their accounts, raising questions of whether bans are an effective tool.
Alongside the harms it causes, social media currently acts as a space for young people to connect, consider their developmental needs and explore their identity. Young people are strongly motivated by their friendships and peer groups, and will work hard not to be socially excluded. As such, so long as their social circles are not complying with a ban, neither will they - raising questions around the enforceability of such restrictions. This is exacerbated when alternative spaces for connection are few and far between. As such, instead of punishing children for the failings of big tech, would it not be better to create and innovate online spaces which are safe for young people? At the very least, if we are going to take digital spaces away, we must ensure there are realised measures in place to replace these. The appearance and prevalence of social media, at least in the UK, has gone hand in hand with the emergence of an increasingly hostile built environment. Without spaces to interact, whether digital or physical, we risk young people becoming increasingly isolated.
AI is not the same as social media, nor is it a monolith, having many forms and facets. In the case of chatbots, they are far more private, personalised and do not facilitate human interaction. However, there are considerable parallels to be drawn between them insofar as they both gather large quantities of personal data, leverage this to personalise the way in which the user interacts with them, and are centred around engagement. Why then are governments restricting social media whilst simultaneously opening the door for wholesale integration of AI technologies into every aspect of our children's lives? Much like social media, AI chatbots are an emergent technology which children have rapidly adopted, and which early evidence suggests may pose significant harms. So we must ask whether we should wait for these harms to become embedded before taking action as we have with social media, or act now to prevent them? There is recognition of AI's potential for harm at a government level, with Norway banning AI in elementary schools, suggesting AI restriction may be on its way elsewhere. Yet, in the UK, children spend 85% of their waking hours outside of school, so simply regulating use in an educational setting does little to protect young people from harm - particularly as it is the ‘off-the-shelf’ tools which are most frequently used by children and are often most unsafe.
The harms AI poses, whilst not necessarily new, are embedded in a new technology with a reach far more extensive. Engaging with social media at least requires accessing a specific platform, but AI has been foisted upon young people without their input or consent into the way they obtain information, learn, and engage with the world. Beyond the question of whether AI chatbots should be banned, their pervasiveness raises the question of how could they be?
Whilst bans may violate the rights of the child to participation, expression and access to information, as enshrined under the UN convention on the Rights of the Child, so too does unchecked deployment of unsafe AI systems. Instead of permitting the culture of ‘move fast and break things’ that technology companies operate under and which created the exact conditions where banning social media has felt like a necessary step, regulation must shift toward a safety by design approach. This reflects the opinions of young people themselves who see the value of technologies like social media in providing connection and community, yet are also highly aware of the negative consequences it can have. Perhaps instead of user-side banning and restriction, developers should be made to take responsibility for the tools they create, the design choices behind them, and the harms they enable.
Fundamentally, AI systems are being deployed and iterated faster than any developmental science can generate evidence, making it incredibly difficult to assess risk and make decisions at a policy level. Regulatory efforts face the dilemma of taking measures now which may prove ineffective or waiting for evidence and exposing society to harm. However, decades of child and adolescent developmental research place us in a position to understand the kinds of design features which can cause harm. As such, there is real potential to establish benchmarks and audit platforms against these.
Are bans the right solution for regulating social media and AI? Only time will tell whether they have a real impact on child wellbeing. What is clear, is that bans alone are not sufficient and may in fact be harmful if not combined with a child-rights focused approach to technology design, and above all a willingness to hold big tech accountable when it comes to the safety of the platforms they develop.