select search filters
briefings
roundups & rapid reactions
Fiona fox's blog
A study published in the BMJ assesses the early effects of Australia’s social media ban on adolescents’ social media use.
Prof Alan Woodward, Professor of Cybersecurity, University of Surrey, said:
“The study, although involving a relatively small number of children, appears to be a sound, scientific finding. It is consistent with otter studies conducted in Australia, all of which point towards the same conclusion: a simple ban is ineffective.
“The study mentions two primary means of circumvention: fake accounts and private browser. The study suggests VPN is not the main form of circumvention. However they do talk about ‘private browsers’ which is a term that is typically a catch-all for when the browser itself has a form of VPN (or a subtle variation known as a proxy) built in, or as a freely available add-on, or the browser is using the web via anonymising network such as Tor (The Onion Router) browser, which is how many access the Dark Web anonymously. The Tor browser is a good example of something that can be freely downloaded and it would be practically impossible to prevent. This latter point is troubling, assuming that’s what the report means, as it means children are being exposed to not just the social media sites most of us use but potentially much darker content. It’s precisely this darker content that appears to have reportedly been involved in some previous tragedies.
“I really hope that governments, including the UK, take note of this latest evidence. Many have used the Australian approach as the example to follow in order to keep under 16s safe online, and what this suggests, along with other similar studies published since Australia introduced their ban, is that a simple ban will not keep children safe.
“Frankly, it was obvious it wouldn’t work which is why so many spoke out against it. This study is evidence not just that simple bans will be circumvented but that simplistic fixes fail the children, whose safety has to be the primary objective. Government policy has to be evidence based, not merely sentiment. Concern does not equal evidence. Simplistic fixes are doomed to fail, and that means failing to protect the children.
“This is a nuanced issue which requires action, but this study demonstrates that it has to be properly thought through. We need to understand how children are harmed and tackle the specific causes. It may be that policing the platforms and their algorithms is part of the answer; it may be that modifying social media behaviour towards different age ranges is also part of the answer; but what this study shows is that a blanket ban for under 16s is not the answer.”
Prof Dennis Ougrin, Professor of Child and Adolescent Psychiatry and Global Mental Health, and Consultant Child and Adolescent Psychiatrist, Queen Mary University of London, said:
“The findings should be interpreted cautiously. The sample was relatively small, drawn from a single Australian state, and relied on self-reported social media use, which may affect both accuracy and generalisability. The follow-up period was also only three months, which is unlikely to be sufficient to assess the full impact of a policy of this scale. Importantly, the study measured social media use rather than mental health, wellbeing, educational outcomes, or self-harm. Nevertheless, it provides useful early evidence that implementation challenges, particularly around age verification and circumvention, may limit the immediate impact of age-based restrictions and underscores the need for longer-term, larger-scale evaluations.
“These findings provide an important early reality check for policymakers. Three months after implementation, there is little evidence that Australia’s under-16 social media restrictions have substantially reduced social media use, largely because many young people continue to access platforms despite the ban and age-verification systems appear easy to circumvent. However, it is too early to conclude that the policy has failed. The key question is not simply whether use falls, but whether restrictions improve outcomes such as mental health, sleep, exposure to harmful content, and self-harm. These results highlight the importance of robust implementation and long-term independent evaluation before drawing firm conclusions about the effectiveness of social media bans.”
Prof Matt Williams, Director of HateLab and Professor of Criminology in the School of Social Sciences at Cardiff University, said:
“This is an important study because it moves the debate beyond speculation and provides some of the first real-world evidence on what happens after a social media ban is introduced. Methodologically, it is stronger than many previous studies in this area because it uses a quasi-experimental design that is well suited to evaluating policy interventions. That said, this should not be read as proof that age-based restrictions cannot work. The study only looks at the first three months after implementation, relies largely on self-reported behaviour, and the confidence intervals are wide enough that modest effects cannot be ruled out. In effect, the study is evaluating the policy as it was implemented in practice, not the hypothetical effects of a perfectly enforced ban. Perhaps the most important finding is that most under-16s were still accessing restricted platforms and many were bypassing age checks altogether. That suggests implementation and enforcement may be just as important as the legislation itself. The headline here is not that the policy has succeeded or failed, but that early evidence from Australia shows how difficult it is to translate a legal restriction into meaningful changes in young people’s online behaviour.”
Dr Victoria Baines, Fellow of BCS, The Chartered Institute for IT, said:
“This article presents high quality research, which tells us that regulation alone does not protect children from online harm. It reminds us that there is often a gap between policy and implementation, and that implementation can take time to enforce. While it doesn’t explicitly tell us the causes of this poor implementation, it provides some evidence that age restrictions and age verification measures are being easily circumvented.
“The article challenges the prevailing rhetoric presenting social media age restrictions as silver bullets for “giving children their childhood back.” It aligns with emerging evidence on non-compliance with the ban in Australia and its causes, including a University of Chicago working paper analysing data from a larger cohort of 746 teenagers – https://bfi.uchicago.edu/working-papers/why-bans-fail-tipping-points-and-australias-social-media-ban/. It also chimes with and acknowledges the Australian eSafety Commissioner’s own update highlighting poor or ineffective implementation of restrictions by social media platforms – https://www.esafety.gov.au/sites/default/files/2026-03/SocialMediaMinimumAgeComplianceUpdateMarch2026.pdf?v=1774905032806. In highlighting the Australian government’s investigations under way (p8), the article correctly identifies that the true impact of the law may in future lie in its enforcement against non-compliant platforms, which in turn could result in lower levels of usage by u16s. The findings presented in Table 5 are noteworthy because they prompt further questions concerning the adequacy of platforms’ age verification measures.
“Helpfully, recruitment of participants before the ban (“the baseline”) means that the same cohort’s experiences can be tracked over time, rather than basing findings on generalised populations.
“The data concerns a small sample, both in terms of number and geography (confined to NSW). This is acknowledged by the authors, as is the fact that “The full impacts of the Act may not be evident for a decade.”
“The press release slightly overstates the findings. While this may sound rather like splitting hairs, the article’s conclusion that “little evidence was found of immediate substantive reductions” is more cautious than “There is little evidence that Australia’s Social Media Minimum Age Act has led to any immediate reductions”. The research has found reductions for the 14-15 age group in both daily usage and time spent on social media, but they are small.
“The press release also does not state the baseline, pre-ban, percentage of u16s using social media. This matters, because it’s important to understand how the 85% still accessing social media differs from the proportion doing so before the ban. The article *does* contain this information: Table 2 on p6 shows the percentage change from the baseline sample before the ban, highlighting a small reduction. This confirms that the baseline is not 100% usage – so the 85% reported access does not represent a 15% reduction across the age groups.
“On the use of VPNs as circumvention measures, the article correctly points to a large increase in UK usage following the coming into force of the Online Safety Act, but neglects to reference the UK Safer Internet Centre’s finding that this is not attributable to children – https://saferinternet.org.uk/blog/new-research-from-childnet-into-vpns. This does not undermine the findings of the study itself, but taken with the misidentification of VPNs as “virtual network providers” – rather than Virtual Private Networks – it does suggest that the authors may have been less familiar with the existing body of research on this particular aspect of the debate on the effectiveness of social media restrictions.”
‘Assessing early effects of Australia’s Social Media Minimum Age Act on adolescents’ social media use: observational study’ by Courtney Barnes et al. was published in the BMJ at 23:30 UK time on Wednesday 24 June 2026.
DOI: 10.1136/bmj-2026-363695
Declared interests
Prof Alan Woodward: “I have no conflicts of interest in commenting on this subject.”
Prof Dennis Ougrin: “None.”
Prof Matt Williams: “I am on the Ofcom (Wales) Advisory Board, and am a founder and director of nisien.ai, but this comment is in my capacity at Cardiff University.”
Dr Victoria Baines: “Funded by Google to conduct independent research into the EU’s regulatory landscape regarding information security (2021 & 2022).
Funded by UK digital certificate authority Aretiico to conduct independent research into the exercise of digital sovereignty through Public Key Infrastructure (2025).
Member, Snap Safety Advisory Board (unpaid, 2023-).
Member, Advisory Board, International Association of Internet Hotlines (unpaid, 2014-2023).
Member, Advisory Board, cybersecurity vendor Reliance Cyber (2018-).
Former Facebook Trust & Safety executive (2013-2017).
International Expert, Council of Europe Children’s Rights and Cybercrime Divisions (2019-).
International Expert, Safe Online (UNICEF, 2018-).
Expert Reviewer, European Union Horizon Europe Fund (2012-).”