Companies accumulate vast amounts of data, including the traces of our digital activities, and use this to develop platforms that have become increasingly central in our lives. On the 15th June 2026, the UK became the latest in a series of countries to announce a social media ban for teenagers under the age of 16.
Social media platforms present real opportunities, such as online spaces for creativity, sharing ideas, and building valued communities. However, awareness of the serious systemic and individual risks they generate has been rising only slowly - and platforms are often opaque, evading the checks and balances deemed essential by experts and advocates, such as our CoCoDa team. Particularly resulting from the recommendation systems that the platforms are created from, risks include generating misinformation, manipulating attention and creating dependencies on the platforms, normalising polar viewpoints via continued exposure to material such as unrealistic body standards, misogyny, or violent material, risks to emotional and mental wellbeing, and more. Unsurprisingly, among the wider population navigating these challenges, young people often face magnified risks.
Social media bans have met a mixed response from advocacy groups, researchers, those caring for young people, and young people themselves. For example, those countering the UK’s policy highlight concerns including: that the ban ‘won’t work’ to keep all young people off platforms, a view officially taken by the Estonian government and legitimised by the lacklustre implementation of a similar approach in Australia; issues undermining enforcement such as young peoples’ rights including privacy and freedom from discrimination - and concerns that efforts to circumnavigate bans through technologies such as VPNs may undermine their important use in protecting user privacy; loss of platform benefits to young people, including access to news and; a potential distraction the bans might create from platform scrutiny, with campaigners suggesting these “risk letting tech companies off the hook”. By contrast, those in favour identify immediate benefits of decreasing young people’s dependencies on platforms and mitigating some risks in the short-term, and suggest the ban helps to communicate and legitimise the seriousness of the risks. There is a suggestion of support among parents, with 9 in 10 UK parents who responded to the recent government consultation registering their support of the ban.
Broadly, we seek to amplify one resounding message from across experts and the public. Regardless of this ban, we need to maintain substantial focus on scrutinising current and emerging platform design and generating the mechanisms necessary to ensure platform accountability.
On that note, this blog shares three recommendations for better accountability of social media platforms in order to protect both young people and other users.
These recommendations are based on research by the CoCoDa project and our findings from an expert workshop held in collaboration with King’s College London (KCL) researchers exploring methods for monitoring young people’s experiences with generative AI. They were submitted to the UK government as part of the recent Growing up in an online world consultation. A summary version of our response is shared at the bottom of this page and contains links to our cited research.
Recommendations
1) Social media platforms are still not safe, even if young teenagers are officially banned; we need to continue to identify their risks and associated design features, and follow through with mitigations and accountability mechanisms to protect populations.
Social media bans are far from perfect interventions. Systemic risks that social media platforms generate affect the wider population and will continue to impact teenagers. Despite the differences in how adults and young people navigate online platforms (see recommendation 2), a ban for children and teenagers should not be taken to imply that these platforms are otherwise safe or that the main part of the work has been done. Further, the bans only apply to a specific set of platforms; some, including a YouTube spokesperson, have argued that the policy could push young people towards other, less regulated platforms.
Specifically, social media platforms are often designed in ways that encode ‘dark patterns’ and addictive design, similar to other online systems like chatbots and search engines. This includes: keeping people’s constant engagement and capturing attention for long periods of time; subtly preventing people from capturing, verifying and monitoring their content, as demonstrated in recent CoCoDa research; and providing content that focuses on monetisation and attention capture over social benefit. There is a huge breadth of platforms that create such patterns by design and as part of their fundamental monetisation approach. As the CoCoDa team argues in this blog, we need scrutiny and mitigations that are consistently targeted at platforms’ design features and company behaviours, and for platforms to be held accountable for their failure to mitigate negative impacts. For example, recent fines were levied at X under the EU’s Digital Service Act for design and transparency issues, and a recent landmark court case in the US on the addictive design of Instagram and Youtube contributing to mental health distress of a young woman, alongside similar cases settled out of court; it is these kinds of actions that should be maintained and extended.
Further, we do not yet have sufficient evidence to understand the impact of the bans, and we will need to carefully monitor their implementation for both positive and negative effects.
2) Investigations of social media platforms and the mechanisms to protect young people need to centre on young people’s rights and be calibrated to their real needs.
Currently, the opacity of platforms hinders the assessment and prevention of severe risks. These risks rapidly evolve, and the personalised experience of each individual creates additional challenges for understanding the role of platforms. Meanwhile, we must recognise that banning platform use, and the secrecy this could create, may provide additional barriers for assessing young people’s experiences. For a long time, researchers have urgently asked companies and governments for better mechanisms to be able to study the impacts of platforms, given that companies often provide limited and unreliable sources of data for research.
Researchers have resorted to creative methods to generate valuable data for studying platforms. For example, members of the CoCoDa team have developed sockpuppet auditing mechanisms to identify how quickly (young) people will enter filter bubbles. The results demonstrate that across a diverse range of topics, there is a concerningly rapid concentration of posts, particularly for young people.
Researchers are also part of ongoing campaigns to develop regulatory mechanisms to enforce data provision for investigating online platforms, such as continued campaigns for effective implementation of the Digital Services Act’s Article 40 in the EU. This Article aims to provide researchers access to data for studying key systemic risks, but it is currently providing researchers with an uphill battle to access this data in practice, as we demonstrated in recent research. We follow others, such as the Social Platforms Data Access Taskforce, in advocating for robust regulatory interventions to help researchers in the UK study online platforms, and expect news from the Department of Science, Innovation and Technology soon on what will be put in place.
In terms of mechanisms for intervention, those caring for young people need to be able to understand their experiences online without violating young people’s privacy, which risks limiting trust and exacerbating secrecy. Also, in order to draw nuanced conclusions, researchers need to be able to draw from different types of data that represent young people’s experiences. CoCoDa researchers have advocated for the ‘digital window’ mechanism to give concrete insights into individual young people’s experiences online, for example, giving summaries of topics where they see high volumes of posts. It aims to balance children's privacy with oversight mechanisms for carers and researchers, and should be calibrated based on age and needs.
Forthcoming CoCoDa research, alongside research by campaign organisations, also demonstrates that platform mechanisms currently designed to protect young people, such as options to block certain kinds of content, do not work. Young people often possess more limited digital literacy to help them navigate online platforms and to exercise control over the content they are shown, and how their data is used. 42 child and online safety groups, as well as researchers and families, signed a joint statement arguing the need for online platforms for young people that are safe for them to use, with the NSPCC’s CEO arguing that children have a “fundamental right to safely participate in the digital world”. Suitable products for young people would give controls over what they see, that are calibrated to their actual ability to use these.
Any mechanisms to protect people that do not work risk ‘washing’ away the impetus for those that do.
3) New systems, such as generative AI chatbots, are increasingly impacting populations, and we need to continuously monitor their impact.
The platforms that we rely upon continue to evolve. Systemic risks of increasingly influential technologies such as generative AI chatbots are also rising, encoding similar design patterns such as emotional dependencies and perspective manipulation.
We are in a position to do things differently than we have with social media platforms. In particular, we must not wait for systematic data provision by platforms themselves, and must identify mechanisms for nuanced investigations of platforms and their impacts.
In exploring potential solutions, we hosted a workshop with colleagues from King’s College London to examine how to produce useful and rich data about young people’s experiences - and proper, independent stewardship of this data. In the workshop, together with experts working in the fields of AI and youth protection policy, we discussed the practicalities of implementing the micro-narratives method for collecting group sentiments in a more real-time fashion. Such methods help us to generate rich data on a short timescale. This could help minimise the harms that we have observed emerging from fast-paced technological adoption in societies, when evidence to inform policy and other interventions has been outpaced.
Conclusions
This blog has urged continued, more ambitious research, policy, and platform development in the wake of social media bans for teens. Our CoCoDa team will continue to develop mechanisms that centre young people and other populations’ real needs in relation to dominant online platforms such as social media, in collaboration with a wide and international network of others seeking to do the same. We are also launching a continuous monitoring initiative examining young people’s experiences with online platforms, including generative AI tools. Get in touch if you are interested in collaborating or supporting us with this work through funding.
Please see a summarised version of our response to the Growing up in an online world consultation here.
If you would like to keep following our work on platform accountability as part of the CoCoDa project, follow the project on LinkedIn for regular posts about our work and perspectives.
Members of the CoCoDa team, who are also contributors to this blog post, are: Simon Mayer (Professor at the University of St Gallen), Luka Bekavac (Doctoral Student at the University of St Gallen), Aurelia Tamò-Larrieux (Professor at the University of Lausanne), Alice Palmieri (Doctoral Student at the University of Lausanne), Konrad Kollnig (Assistant Professor at the University of Maastricht), Henry Tari (Doctoral Student at the University of Maastricht) and Jake Stein (Open Data Institute). We collaborated with Petr Slovak and his team at King’s College London on examining how to steward data for monitoring young people’s experiences with generative AI.
