Social media bans illustrated by a child at a crossroads between digital life and outdoor play with friends.
|

Social Media Bans Will Not End Online Safety Education

Social media bans raise important questions about children’s rights, digital inclusion and the future of online safety education. However, do children need to experience online risks to learn how to protect themselves? As gaming, messaging and artificial intelligence increasingly shape childhood, the debate must extend beyond conventional social media.

On 7 October 2026, the London School of Economics published the opinion article How will we teach social media safety to children banned from using it? by Savannah Hanson, Carys Evans and Zubeda Limbada of ConnectFutures. The authors raise legitimate concerns about practical media literacy education following age restrictions. Rather than dismissing those concerns, we should examine the assumptions behind them and consider how education can adapt.

Should children experience risks to understand them?

The authors describe an Ofcom-funded media literacy programme across twelve Birmingham secondary schools. School mobile phone restrictions disrupted practical workshops intended to help students create responsible digital content. This illustrates a genuine teaching challenge, but it does not establish that children need personal smartphones or public social media accounts to acquire digital literacy.

We teach road safety without requiring children to drive, gambling awareness without requiring them to place bets and fraud prevention without requiring them to become victims. Similarly, educators can use simulated feeds, fictional profiles, anonymised cases and supervised technology to teach online safety. Of course, practical experience has value, but a controlled educational environment differs fundamentally from an unrestricted commercial platform.

Creating responsible content can still be practised without publishing on a personal account. Once a post is live on a student’s own profile, visibility, screenshots and contact fall outside the classroom. School supervision ends when the session ends, the post does not. Shared devices, private drafts that are never published, simulated interfaces or teacher controlled accounts allow the same exercise in ethical creation while keeping the child out of the public or semi-public space the lesson is intended to help them navigate safely.

The analogy is imperfect. Online environments involve identity, belonging and social comparison in ways that road safety or fraud education do not. That difference does not require personal accounts; it does mean educators must design activities that practise those social dimensions in controlled settings.

Prevention and harm reduction can coexist

Furthermore, the authors draw parallels with drug and vaping education, questioning approaches that simply tell children what not to do. Such criticism is reasonable because warnings alone rarely address the complexity of children’s lives. However, age restrictions and education can coexist, even when some children circumvent restrictions.

Education must prepare children for future experiences while supporting those who already encounter harm. Furthermore, educators should distinguish between a school phone policy that prevents a particular teaching activity and a national restriction on account ownership. The two policies create different challenges and require different responses.

If a ban proceeds, schools will also need clear guidance on what supervised educational use remains permitted. Without it, some may drop useful activities entirely rather than risk appearing to condone prohibited access.

Why social media bans cannot eliminate online harms

Today, the boundaries between gaming, messaging, video streaming and social media have become increasingly blurred. Children can share images, communicate with strangers, build friendships and experience social validation without opening an Instagram or Snapchat account. Ofcom’s 2025–2026 report illustrates the diversity of children’s online experiences.

Social media and other platform use among UK children aged 8 to 14
Source: Ofcom, Children and Parents Media Use and Attitudes Report 2025–2026. Use in the previous four weeks; YouTube excludes YouTube Kids.

In Ireland, CyberSafeKids’ 2026 annual report identified Roblox as the most popular platform among surveyed children aged eight to twelve and WhatsApp among those aged twelve to fifteen. In addition, 26% of surveyed children aged eight to twelve reported troubling content or unsolicited contact. These findings reinforce concerns about online gaming child safety gaps. Therefore, education must teach transferable skills rather than concentrate on particular apps.

Gaming, WhatsApp and online marketplaces

Modern gaming environments allow children to communicate, compete for status, purchase virtual goods and participate in communities. Meanwhile, WhatsApp groups can facilitate bullying, exclusion, image sharing and misinformation. Video platforms such as YouTube also provide opportunities to encounter misleading information, persuasive recommendations and social comparison.

Even marketplaces such as Vinted and eBay can create opportunities for inappropriate contact, impersonation and financial manipulation. A 2025 Guardian investigation into Vinted documented sexual harassment and misuse of sellers’ photographs. This does not establish the prevalence of abuse specifically targeting minors on either marketplace, and eBay requires members to be adults. Nevertheless, the examples illustrate why a platform’s intended purpose does not guarantee safety.

For example, a child who understands grooming, deception and emotional manipulation can apply that knowledge across different services. Similarly, lessons about privacy, digital identity, image misuse and commercial persuasion remain relevant wherever children interact online. Which specific learning outcomes actually require a personal account on a restricted platform?

Understanding risk does not eliminate vulnerability

Childhood and adolescence involve ongoing development of emotional regulation, impulse control and decision-making. A teenager may understand social comparison while continuing to experience its effects. Similarly, a child may recognise manipulation yet remain susceptible when seeking acceptance. The National Academies’ consensus report on Social Media and Adolescent Health addresses both benefits and harms and recommends action extending beyond education to platform design and accountability.

Importantly, the evidence does not establish that every child experiences the same effects, nor does it justify attributing all adolescent mental health difficulties to social media. Nevertheless, acquiring knowledge does not eliminate developmental vulnerabilities. Education cannot substitute entirely for appropriate protection.

The commercial environment

In practice, many commercial platforms generate revenue through advertising or purchases and design features that encourage engagement. Personalised recommendations, autoplay, public popularity metrics and behavioural profiling can influence users. Similar mechanisms exist in games and streaming services, so social media bans alone cannot address persuasive design.

Children should understand these mechanisms, but even informed adults can struggle to regulate their engagement. Therefore, responsibility must extend beyond children, parents and teachers to the companies that design digital environments. Free access does not necessarily mean that a service comes without a cost.

Could marginalised children lose essential support?

However, a strong argument against age restrictions concerns children who rely on online communities for acceptance and support. LGBTQ+ young people, children with disabilities, neurodivergent children and those facing geographical isolation, discrimination or abuse may find meaningful relationships online. Losing access to those relationships could have serious consequences, especially when alternatives are inaccessible.

A 2024 study, Social media: A double-edged sword for LGBTQ+ youth, examined 406 young people aged fourteen to eighteen. It identified experiences of connection and rejection, alongside associations between online discrimination and poorer mental health and substance use. These findings illustrate why we must recognise benefits and harms together, without assuming that all marginalised children share the same experiences.

Friendship, belonging and access to information are fundamental human needs, not inherently commercial products. However, removing access without providing meaningful alternatives could deepen isolation. Governments should consult affected children, preserve access to confidential support and invest in inclusive youth services and appropriately moderated communities. Digital inclusion and unrestricted commercial exposure are not synonymous.

Children’s rights and digital participation

The UN Committee on the Rights of the Child’s General Comment No. 25 recognises children’s rights in the digital environment. Access to information, participation, expression and privacy matter alongside protection from exploitation and violence. Policymakers must balance these rights and consider children’s evolving capacities rather than treat protection and participation as opposing objectives.

Social media bans are not internet bans

The UK government’s proposals envisage restrictions on social media while allowing standalone messaging services such as WhatsApp and Signal. Australia’s minimum age framework similarly excludes several categories, including services primarily intended for education, health support, messaging and gaming. Exclusion from a regulatory category does not establish that a service is safe.

What can Australia’s social media bans teach us?

The Molly Rose Foundation’s 2026 research found that 61% of previous users retained access to at least one restricted account, while 51% reported no change in online safety and 14% felt less safe. These findings raise questions about implementation and unintended consequences. However, the eSafety three-month evaluation provides additional context.

Social media bans in Australia: early changes in account ownership and social expectations
Source: Australian eSafety Commissioner three-month evaluation, 2026. Different measures have different survey denominators; the changes do not establish causation.

In addition, the evaluation identified reduced parental awareness of some children’s social media use and increased frequency of messaging and gaming. These are important potential unintended consequences, although the evidence does not establish that the restrictions caused them. Continued account ownership may reflect weaknesses in age assurance or enforcement as well as deliberate circumvention.

Ultimately, neither account removals nor early surveys can establish the long-term success or failure of social media bans. Governments must assess actual exposure to harm, wellbeing, social inclusion, compliance and displacement into other services. Effective enforcement and independent evaluation are essential components of any credible policy.

Artificial intelligence changes the debate

AI chatbots offer educational benefits but also introduce risks involving inaccurate information, emotional dependency and simulated relationships. CyberSafeKids’ 2026 annual report found that 55% of surveyed children aged eight to twelve and 76% aged twelve to fifteen had used AI chatbots. Consequently, online safety education must anticipate these experiences rather than wait until children encounter difficulties.

AI chatbot use among surveyed Irish children: 55% and 76%
Source: CyberSafeKids, The Year Little Changed, 2026. Surveyed Irish schoolchildren.

Should children first develop emotional relationships with AI companions before educators explain the potential risks? Clearly, education can precede exposure. Nevertheless, practical and supervised exploration of appropriate AI tools can help children understand their limitations.

News consumption and independent thinking

The LSE authors rightly raise concerns about losing incidental exposure to news. Social media can introduce young people to important events and perspectives that they would not otherwise encounter. However, algorithmic feeds may also amplify sensational content, repeated political narratives or misleading information.

Educators can help children independently locate news, compare sources, distinguish fact from opinion and evaluate competing political perspectives. The goal is not to prescribe a political viewpoint but to develop the ability to examine claims, evidence and arguments. Yet another technological change may already be transforming this process.

From social media echo chambers to AI answers

Increasingly, users turn to AI instead of traditional search engines and news sources. The Reuters Institute Generative AI and News Report 2026 found that weekly use of AI to find information increased from 11% in 2024 to 38% in 2026 across six surveyed countries. These figures concern general surveyed populations, not children specifically.

Weekly AI information use rose from 11% in 2024 to 38% in 2026
Source: Reuters Institute, University of Oxford, Generative AI and News Report 2026. Six-country survey trend, not child-specific.

Of course, search engines also rank and personalise information, but generative AI can select, interpret and synthesise sources into a single apparently authoritative answer. When a child asks about an election, immigration, religion or an international conflict, what determines which perspectives appear? Could earlier conversations or inferred preferences influence the answer?

A 2026 working paper, AI Pandering: Constructing Diverging Political Realities through Conversation, identified evidence of AI responses adapting to inferred political views, including differences in source recommendations and agreement with initial opinions. The paper has not undergone peer review. Moreover, context-driven adaptation does not necessarily mean that a model retrains itself around an individual user’s beliefs.

Political persuasion and AI personalisation

Peer-reviewed research in Science also found that techniques increasing political persuasion could reduce factual accuracy, with personalisation playing a smaller role than certain training and prompting techniques. Another study in PNAS examined AI influence on real-world political actions. Conversely, AI can help users understand complex issues and interrogate misinformation when used critically.

For example, training data, source selection, developer decisions, conversational context and the tendency to agree with users can all influence AI answers. These mechanisms differ, and none guarantees that a particular response will be politically biased. Nevertheless, an apparent agreement from AI could reinforce an existing worldview without offering the user sufficient opportunity to evaluate alternatives.

There is also a risk that convenience replaces curiosity. If children delegate the process of locating, comparing and interpreting evidence to AI, they may have fewer opportunities to practise those intellectual skills. Media literacy should therefore teach children to request original sources, challenge assumptions, investigate competing perspectives and recognise uncertainty.

Children must never fear reporting online harm

Nevertheless, perhaps the strongest point in the LSE article concerns a child reporting cyberbullying on a restricted platform. A child’s unauthorised access must never compromise the safeguarding response. The immediate priority is the child’s welfare, regardless of the service involved.

This reflects CODA’s #HowRUDoinOnline? initiative. Adults regularly ask children about school and friendships, but how often do they ask how online experiences made them feel? Children may conceal harmful experiences because they fear blame, punishment or losing their devices. The response to a first disclosure can influence whether they seek help again.

What should online safety education teach?

Effective education should progressively develop emotional awareness, critical thinking, an understanding of manipulation, privacy, commercial incentives and digital identity. Children also need to understand AI use and its limitations, social comparison and coercion. Above all, they must know where to seek help without fear.

In addition, practical workshops should remain part of the curriculum, using supervised tools and simulations where appropriate. Educators can also acknowledge experiences children already have without requiring others to open accounts. Learning outcomes should focus on transferable skills rather than proficiency in particular platforms.

Can social media bans help reclaim childhood?

The debate should also consider what children might gain from spending less time online. Childhood offers a finite period for developing friendships, physical skills, imagination and independence. Therefore, the potential benefits of age restrictions extend beyond reducing exposure to harmful content.

However, those benefits are not guaranteed. If children simply move from social media to gaming, messaging or AI companions, the balance between digital and everyday life may remain unchanged. Equally, some online experiences provide genuine opportunities for learning, creativity and connection.

Consequently, families, schools and policymakers should help children access appealing alternatives, including inclusive sports, outdoor play and opportunities to meet friends. The aim is not to remove technology from childhood altogether. Instead, it is to ensure that technology complements rather than displaces the experiences children need to develop.

Social media bans are not a complete solution

Children of the Digital Age supports a Smartphone-Free childhood, reduced recreational screen time and delaying access to social media. However, restrictions alone cannot eliminate online harms. Policy must address unintended consequences, protect access to meaningful support and hold technology companies accountable across the wider digital environment.

The LSE article raises a valuable question about adapting education after restrictions. The answer involves more than recreating existing teaching methods: it requires a broader understanding of the digital world, childhood development, commercial design and independent thought. We should be preparing children for the digital world, not preparing children for the business models of technology companies.

References and further reading

How will we teach social media safety to children banned from using it?
Ofcom’s 2025–2026 report
CyberSafeKids’ 2026 annual report
National Academies: Social Media and Adolescent Health
Social media: A double-edged sword for LGBTQ+ youth
eSafety three-month evaluation
Reuters Institute Generative AI and News Report 2026
AI Pandering (working paper)
Science: The levers of political persuasion with conversational AI
PNAS: AI can persuade people to take political actions

About Children of the Digital Age

Children of the Digital Age provides online safety education and training for schools, parents and organisations. Our guides support practical conversations about online behaviour and the choices people make. However, education and app settings do not replace safeguarding or professional support.

Send Us a Message

Contact Children of the Digital Age to enquire about online safety education, parent sessions or training for your school or organisation.

    Similar Posts

    Leave a Reply

    This site uses Akismet to reduce spam. Learn how your comment data is processed.