Following in Australia’s footsteps, policymakers around the world have been introducing their own age-based social media restrictions. Given the limitations of an outright ban, legislation should instead focus on making platforms safer by design, tackling both functionality and harmful content.
Concerns surrounding the negative impacts of social media on children have led policymakers to consider age-based social media bans, with Australia being the first to act. These blanket approaches don’t take into consideration the many benefits that social media can have for children.
Initial evidence from Australia suggests that blanket bans are ineffective, easily circumvented, and do not prevent children from experiencing harm online. They can also create an illusion of safety for parents, and result in a “cliff edge” when minors finally do gain access to social media.
The UK’s proposed ban claims to go further than the Australian model, targeting contact with strangers and livestreaming. These additional measures for 16-17-year-olds could be undermined if platforms decide these are not worth introducing and simply ban all under-18s.
Legislation should instead focus on making social media platforms safer by design, tackling both platform functionality and harmful content. Promisingly, online safety panels commissioned by the EC and Germany have recommended this, advocating for a more nuanced approach than bans.
Generally overshadowed by complete bans, some countries have begun to legislate for safety by design. For example, Brazil has enforced measures tackling addictive features such as autoplay, push notifications, and infinite scroll for under-16s.
For any of this to work, stricter enforcement of age-verification methods are needed as platforms can largely choose their own systems, which vary in terms of reliability and privacy. For Europe at least, the EC is well positioned to implement more harmonised rules.
While social media can expose children to severe harms, it also brings benefits such as connection and community
In recent years, concerns over the harms that social media can inflict on children have become an international talking point, putting pressure on policymakers to implement bans or restrictions. Australia was the first nation to act, passing a social media ban for under-16s in November 2024, which came into effect on 10 December 2025. This has opened the floodgates for countries around the world, with 24 nations now considering similar measures, according to our Age-Based Restrictions Benchmark in the Online Safety Tracker. Most recently, on 15 June 2026, the UK decided to follow in Australia’s footsteps, with former Prime Minister Keir Starmer announcing plans to legislate for a ban.
The main risks of social media are related to addictive design features and exposure to harmful content. Addictive features such as infinite scroll, autoplay and personalised algorithms can make it difficult for users to stop using social media, making them dependent while also reducing their attention spans. The amount of time that teenagers spend online dramatically increases with age, with those in the 15 to 17 age group spending more than twice as much time online as 8 to 9 year olds – see Figure 1.
For the oldest cohort, this equates to over six hours per day – roughly the same length as a whole school day. While this highlights the need for some form of restrictions, it also demonstrates the profound role the online world plays in children’s lives. A complete social media ban would create a major void in the lives of young people, with the most significant impact felt by older teenagers. While the Government has argued that this measure will “give kids their childhood back”, this is perhaps easier said than done: for children who have spent hours per day online for many years, the immediate transition to going completely offline may be difficult.
Australia’s confusing set of criteria for determining restricted platforms has been copied across to the UK
Of the 24 countries that are considering complete bans, only three (Australia, Malaysia, and Indonesia) are currently in force. The UK’s proposed ban is the only example of a government explicitly following Australia’s model, targeting user-to-user platforms that enable social interaction and allow users to post material. It has subsequently been labelled “Australia plus”, as it then goes further by targeting gaming sites and AI chatbots. However, this model has caused some confusion regarding which platforms are actually in scope. While 10 major platforms (Facebook, Instagram, Snapchat, Threads, TikTok, X, Youtube, Reddit, and streaming platforms Kick and Twitch) are often cited as being the key targets of Australia’s ban, in reality an extensive number of sites are affected, to the point that the eSafety Commissioner does not maintain an exhaustive list. Instead, platforms must meet all of a set of criteria to be restricted – see Figure 2.
Platforms would only be exempt from the ban when the sole or primary purpose of the service is enabling or supporting: communication, online games, information about products or services, professional networking or development, education, and health. As such, the eSafety Commissioner has noted that platforms such as Google Classroom, LinkedIn, Messenger, Pinterest, Roblox, WhatsApp, and YouTube Kids will not be impacted.
The UK’s ban goes further than Australia’s by targeting gaming sites and AI chatbots, but some of these measures have already been introduced by other countries
The UK’s proposal targets a wider pool of online services by restricting specific functions for various age brackets. For example, AI ‘romantic companion’ chatbots will have to enforce a minimum age of 18, and AI chatbots in general will have to force minors to take regular usage breaks. Features allowing under-16s to post livestreams, access disappearing messages, and have communications with adult strangers would also be prohibited, with the UK being the first country to restrict these features specifically. However, it is unclear whether children would still be able to be in livestreams from an adult’s account, or if platforms would have to detect when a child is present in every single livestream posted online. Both of these options would require very different monitoring mechanisms, with the latter being much more invasive. Restrictions on these functions will also be on by default for 16- and 17-year-olds, and the Government announced further measures including overnight curfews and restrictions on addictive features (autoplay video and personalised feeds) in July 2026. However, minors will be able to disable these, which could render them ineffective. Some of these features have already been restricted in other countries – see Table 1.
The UK Government has claimed that its proposed restrictions “go further than any country”. However, while it is true that no other country has banned children from posting livestreams or chatting to adult strangers, some of its other measures (such as restrictions on certain features for 16-17-year-olds) have already been implemented by other governments. For example, Brazil’s age-based restrictions tackle addictive features and require that accounts held by under-16s are linked to a parent’s (allowing them to set time limits or disable personalised recommender systems), while still allowing children to participate online. Portugal’s restrictions would also limit gamification aimed at extending usage on online services and systems for creating fake images and videos.
In terms of the legislative process, the UK Government already passed the Children’s Wellbeing and Schools Act on 29 April 2026. This requires the Government to introduce regulations imposing some form of age or functionality restrictions for children when using social media, and the ban will be passed under this legislation. By introducing these restrictions as secondary legislation (rather than having to pass a separate bill), the Government is hoping to streamline the process and aims to have the ban in effect by Spring 2027.
Age-appropriate restrictions would be more effective than an indiscriminate blanket ban, which forces a significant number of children off social media
While technically most social media platforms already have a minimum age of 13 – generally in order to comply with the age of consent for data processing – this is not meaningfully enforced. Instead, children can just lie about their date of birth to make an account, with around 38% of children aged 5-7 in the UK using social media, and 63% of children aged 8-11. This is a large proportion of young children who, according to platforms’ own rules, should not be there. One positive impact of a ban would therefore be that children who were never meant to be using these platforms would, in theory, no longer be able to do so.
Australia’s ban was on a much smaller scale due to its relatively small population, meaning that fewer children were impacted. Due to this, tech companies might only be prepared to take limited measures to ensure that minors cannot access their sites. The UK’s ban would impact up to three times more people (see Figure 3), which would give tech companies more incentive to spend time and resources on these measures.
These age groups are at very different developmental stages, but a blanket ban would essentially treat them all the same, without gradually giving more autonomy to older children. The number of people in the 16 to 17 age group (1.7m) is also significantly smaller than the number of children who would be impacted by blanket bans (around 9.1m). This calls into question whether platforms would find it worth spending time and money designing services to implement these restrictions for what is a relatively small group. Instead, they may decide just to block anyone under the age of 18.
Early evidence from Australia shows that blanket bans do not effectively safeguard those they’re designed to protect
Early evidence from Australia’s world-first ban has suggested that this may not be the most effective way to safeguard young people. It has been widely criticised as ineffective, with the majority of children still being able to access social media by circumventing age-assurance measures. In the eSafety Commissioner’s first compliance update after the ban came into force, it noted that around seven in ten parents reported their child still had a social media account. Bans can also push children onto more dangerous online spaces that are unregulated, thus exposing them to more harm. This can create an illusion of safety, where parents may lower their guards, assuming their children cannot access social media, but in reality they still may be exposed to harm.
While initially, the maximum penalty for failing to prevent underage users from accessing social media sites was set at A$49.5m (£25.4m), on 28 June 2026 the Australian Government announced it would be doubling this to A$99m (£51.4m). It will also strengthen the eSafety Commissioner’s information-gathering powers, allowing it to compel social media companies to provide evidence of what they have done to stop under-16s from getting an account. This suggests that the initial penalties did not fully incentivise tech companies to comply with the legislation – the revenue generated from having under-16s on their platforms probably outweighs what they would have to pay in fines. As of yet, the eSafety Commissioner has not issued any penalties, however, it is investigating five companies for potential non-compliance: Facebook, Instagram, Snapchat, TikTok and YouTube.
Suddenly gaining full access to social media at 16 could result in a “cliff edge” for young people, and not gradually developing digital literacy skills could make them more vulnerable to online risks. The UK Government has acknowledged this as a concern, and claims that its additional protections for 16-17-year-olds will prevent this. Nevertheless, so far it has only suggested overnight curfews and restrictions on infinite scroll, as well as on-by-default restrictions on contacting strangers and livestreaming. These limits would not sufficiently protect this age group from addictive design features or harmful content. Furthermore, if big tech companies decide to just ban all minors from their platforms, these measures would be rendered meaningless.
Complete bans also erase any potential benefits of social media. Children use these sites to connect with peers, learn about other cultures, or as a news source. Bans also target platforms such as YouTube, which is not as social as other platforms – to some extent it is more comparable to a streaming service – and can be a key educational resource for young people. As a result, bans may result in children feeling as if they are being punished and shift responsibility away from big tech companies who allow harms to occur on their platforms.
Legislation should instead focus on making platforms safer by design
More meaningful and proportionate legislation could require platforms to implement “safety-by-design” features. This proactive measure would ensure that safety is integrated into a platform’s initial infrastructure rather than treating it as an afterthought. Such an approach would allow policymakers to tackle the risks associated with both platform functionality and the content itself. With regards to content, legislation could force platforms to take down illegal material and prevent children from viewing violent, extreme, sexual, or generally age-inappropriate content, including material that promotes self-harm or suicide. This type of content is already targeted under the UK’s Online Safety Act, the EU’s Digital Services Act (DSA) and Canada’s proposed Digital Safety Act, but if minors are banned from social media sites, this legislation could become almost meaningless, as platforms will no longer need to protect children from viewing this content.
In terms of platform functionality, safety-by-design features could target addictiveness by restricting infinite scroll, autoplay, personalised algorithms, and push notifications. The EU’s Digital Fairness Act (DFA) has not been tabled yet, but is expected to cover manipulative online commercial practices such as dark patterns and addictive design and aim for consumers to have more control over these. Under the DFA, addictive design features would be switched off by default, with consumers being able to opt in if they wish, and certain features being completely prohibited for minors. This would go further than current online safety regulations by protecting not only minors but adults too.
Time limits or overnight curfews could also ensure that children still benefit from time offline while instilling digital hygiene into young people, which will be beneficial as they transition into adulthood. However, there still needs to be a balance between protecting children and giving them more autonomy as they transition into adulthood, and time limits in particular could be seen as overly strict. Hyperpersonalised algorithms, where platforms use personal data to create a custom experience for users, can be especially harmful, as they often push users into echo chambers or suggest divisive content to drive engagement. Algorithms can therefore harm individual well-being while also fueling broader social polarisation. To mitigate this, restrictions could prohibit companies from using minors’ personal data to suggest content, which would also limit the addictiveness of social media and the possibility of children encountering harmful content.
Some platforms have already taken steps to create safer spaces, such as Instagram’s “Teen Accounts” initiative launched in September 2024. This introduced safety features for minors, including default privacy settings, restrictions on livestreaming, and linking children’s accounts to a parent’s. In April 2025, Instagram claimed that there were 54m underage users with Teen Accounts and the platform states that it has 50+ safety tools to protect minors. However, the initiative has faced criticism, with some claiming that the safety features do not work as advertised and that children are still exposed to harmful content and addictive design features. TikTok has similarly introduced “Family Pairing”, which lets parents link to their child’s account and set daily time limits, although children can circumvent this by making a new account. The limitations of these initiatives highlight that meaningful legislation and enforcement is still needed, but the onus should be on platforms to demonstrate that they can make their services safer by design, rather than resorting to bans.
Various countries have shown that making platforms safer by design is a feasible approach
Various countries have restricted design features rather than excluding children from social media entirely. The main examples that have enforced or adopted this are Brazil, China, Vietnam, and Portugal. The restrictions vary between countries but the focus remains the same: safety by design rather than excluding children from online platforms altogether. In China, restrictions come primarily at a device level, with parents being able to turn on “minor mode” on their child’s devices, which includes features such as a daily usage cap and overnight curfews. Legislation in Portugal and Brazil targets specific design features associated with addiction, such as infinite scroll, non-essential notifications, and autoplay. Both countries also ban loot boxes on gaming sites, classifying them as a form of gambling.
This approach has also been advocated for by two expert online safety panels, one based in Germany and the other commissioned by the EC, which published reports in July 2026 recommending EU-wide partial restrictions rather than blanket bans. Both panels recommended a legal minimum age of 13 for independent social media use, with parental permission needed for under-13s, and further partial restrictions for those under the age of 18. The German panel also suggested a second option, where there would be no uniform age limit, but instead specific services and functions would be restricted based on a risk assessment. This demonstrates that there is scope for governments to implement measures which are more nuanced than blanket bans. While these may be more complicated to enforce, it is arguably worth it as a compromise between allowing children to experience the online world and protecting their safety.
Binding regulation is needed regarding which age-assurance methods should be used
For age-based restrictions to actually work, social media sites need to verify the ages of all users on their platforms, so they know which users should be restricted. Therefore, under complete social media bans, all users would have to verify their ages to prove they are over 16, not just children as some have assumed. This will be challenging for governments to enforce, as attempts at age verification have generated public backlash, mainly due to privacy concerns. It’s likely that this backlash will grow as bans come into force and users begin to encounter age checks, with many being reluctant to share personal details with platforms.
There are three main types of age-assurance methods, with varying degrees of accuracy:
Self-declaration: The user declares that they are over a certain age, usually by providing their date of birth;
Age estimation: The user’s approximate age is determined through methods such as facial or behavioural analysis; and
Age verification: The user provides personal data on a temporary basis to prove their age, such as an ID, bank details, or a SIM card.
Users can (and do) lie about their age, with the eSafety Commissioner finding that 80% of children aged between 8-12 used at least one social media platform in 2024, making self-declaration an unreliable method. Therefore, more reliable methods are needed in order to make bans or restrictions effective, with part of the reason Australia’s ban has been so ineffective thus far is because children have been able to circumvent age-assurance methods. While age-estimation methods (such as facial analysis) are more reliable to an extent, these can still fall short, with anecdotal evidence of children providing pictures of an adult relative or using makeup to circumvent online safety measures. Age-verification methods are generally more accurate, but come with additional privacy concerns.
Few countries with age-based restrictions have outlined which methods for age verification should be used aside from the fact that they must be reliable and effective. However, in Malaysia (which has enforced a ban for under-16s), age verification must be achieved through government-issued documents. Similarly, in Vietnam (which has enforced partial restrictions for under-16s), accounts must be verified through mobile phone numbers (SIM registration in Vietnam is linked to real identity) or a national ID number, with unverified accounts being unable to post, comment, or livestream. Elsewhere, platforms are largely able to decide which methods to implement for themselves, which is problematic as these methods differ considerably in terms of accuracy and privacy.
Due to the inherently global nature of social media sites, restrictions and age-verification methods would be stronger if they are harmonised internationally. The EU is therefore well positioned here, and both online safety panels recommended EU-wide measures as being more effective than fractured restrictions across Member States. The EU is currently trialling its digital ID wallet (EUDI), which uses zero-knowledge proof technology. This means that the online platform requesting age verification will not know the user’s identity or be provided any of the user’s data, and would also take the pressure off of individual sites to verify their user’s ages, although concerns still remain about whether user data could be compromised. The German panel recommended that the EU should establish binding regulations specifying which age-verification methods should be used, suggesting that the EUDI wallet should be used in the case of a complete ban, while restrictions on specific services and functions should primarily be enforced via parental controls on end devices. A binding, harmonised set of rules regarding which specific age-verification measures are appropriate could overall make it easier both for platforms to implement and for regulators to enforce.
