More than a year after the Online Safety Act's children's safety duties came into force, new findings from Ofcom suggest that children are still encountering harmful content online. In her latest blog, Polly explores the effectiveness of the legislation, the role of social media algorithms and why online safety education remains essential in protecting young people.

More than a year after the Online Safety Act’s (OSA) protection duties came into force in July 2025, I now want to revisit the effectiveness of the legislation following a concerning new finding from Ofcom.
In a report published on the 22nd October 2026, Ofcom have revealed that “there was no significant change in the proportion of children encountering harmful content over the first nine months after the children’s safety duties came into force”. This raises an important question; if social media companies are now legally required to prevent children from seeing harmful content, how are they still encountering it?
One of the most significant issues highlighted in Ofcom’s report is the role of social media algorithms in exposing children to inappropriate content. Platforms such as TikTok and Instagram rely on personalised video recommendation systems that analyse users’ viewing habits and interactions with posts (for example, what kind of topics do they watch videos on for the longest, which ones are they commenting on/sharing with friends, or which advertising links they are clicking on). Although these algorithms are designed to keep users engaged by showing relevant content to the user, they can still recommend harmful material, even if the user has not actively searched for it.
Unlike traditional search engines, social media algorithms do not rely solely on what users deliberately searched for. Instead, they rely on the behaviours previously outlined to predict which content is most likely going to keep the user engaged and using the app. Therefore, when users interact with controversial or shocking content out of pure curiosity, an algorithm will instead interpret this as a genuine interest in the topic at hand and resultantly continue to recommend videos with similar content. This can become particularly concerning when algorithms begin recommending increasingly extreme or inappropriate material based on similarities between appropriate and inappropriate topics. For example, a young person watching ordinary fitness videos may gradually be exposed to videos promoting completely unrealistic body standards or unhealthy attitudes towards appearance, despite never having actually sought out these kinds of videos. Furthermore, these algorithms are also designed to identify new topics that might attract a user’s attention, therefore will recommend content outside a user's established interests, which can include inappropriate material. Although platforms do provide automated content moderation settings, there are limitations. Users can bypass these settings using coded language and alternative spellings of restricted words that would normally have content flagged as being inappropriate and taken down, therefore allowing harmful material to be shared.
How effective has the Online Safety Act been?
With such harmful content continuing to reach young users, it is understandable to question how effective the OSA has truly been in protecting children. However, despite Ofcom’s findings, it would be inaccurate to suggest the report showed the legislation made no beneficial difference to children’s online safety. In fact, earlier research published by Ofcom in July 2026 demonstrated that the OSA had already produced measurable changes in how children are accessing online services. By June 2026, all 10 of the UK’s most visited pornography websites had implemented age checks. The impact of this implementation can be shown in Ofcom’s Use of Age Assurance Report, which states that the proportion of children encountering age-verification measures has increased from 25% to 43% between July 2025 and January 2026. More importantly, there is also evidence to suggest that these checks are influencing children’s online behaviour. Ofcom has found that 8% of underage users have attempted to access pornographic pages, but half only reached websites with age verification measures in place. Of these 8%, 87% of their visits were recorded as lasting less than 30 seconds, suggesting that some children may have encountered an age check and left the website rather than proceeding further. While this does not prove that every child was prevented from accessing such content, it does demonstrate that the introduction of age checks is in fact successfully preventing children from viewing inappropriate material. This can hopefully offer some reassurance that the OSA has made promising progress in reducing children's exposure to harmful online content.
How can our children be helped outside the Online Safety Act?
An equally important consideration is the extent to which children can realistically be expected to take responsibility for their own safety online. While social media platforms do provide features allowing users to report inappropriate material, these systems rely on those users to be able to recognise the harmful content in the first place. Identifying such material is not always straightforward, especially when the harmful material is disguised as ordinary content. As a result, simply encouraging children to report harmful content cannot be considered a sufficient solution when most do not have the skills to recognise the risks associated with what they are viewing.
One way to improve children's ability to identify these more subtle risks is to encourage them to question the content they encounter, rather than automatically accepting it as reliable or harmless. From my own experience growing up with the internet and being a frequent social media user, I recognise how easy it can be to place trust in someone online simply because they may be of similar age, share the same interests or just seem like a relatable person. When advice comes from someone who appears familiar, it can feel more genuine and trustworthy, even when there is little evidence to support what they are saying. However, children need reminding that being able to relate to someone does not automatically make them a reliable source of information. Ultimately, the people appearing on our social media feeds are complete strangers, and a short video provides very little insight into the credibility, intentions or the accuracy of the information they are sharing. This misplaced trust can be particularly dangerous when seemingly harmless videos go on to encourage young viewers to explore harmful topics, accounts or online communities, leading the viewer towards increasingly inappropriate material they would not have otherwise sought out.
There are of course many other essential skills beyond questioning the content they watch that the younger generation needs to develop to protect themselves online. However, these skills should not necessarily be developed independently. Parents and carers play an essential role in helping children recognise online risks, understand the importance of protecting their personal information and foster the confidence to respond appropriately when encountering something concerning. Encouraging open conversations about online experiences, rather than simply imposing restrictions, can help children feel comfortable discussing unfamiliar or upsetting content without fear of judgement or punishment. It is also important that adults familiarise themselves with the platforms their children are using, allowing them to offer more relevant and practical guidance as online trends and risks continue to evolve. For further advice on recognising potential dangers, developing safer online habits and supporting children as they navigate the digital world, explore the guidance and educational resources available through Every Child Online and Safe & Smart.
Looking for help with a device? Please note that Every Child Online does not provide devices directly to members of the public. *
We’d love to hear from you! Whether you’re interested in partnering with us, making a donation, or simply want to learn more about our work, please get in touch using the form.
*Every Child Online operates within the UK only. We provide support through trusted referral partners, including schools, charities, social services and community organisations. We do not supply devices directly to members of the public or outside the UK.