Britain's media regulator has set out a new package of Online Safety Act duties aimed at child sexual abuse material, moving from a regime in which platforms assessed risk and responded to reports toward one in which they are expected to hunt for the material themselves. Ofcom announced the measures on Monday, 16 November 2026, and the government laid them before Parliament the same day, with the duties expected to take effect from early 2027 subject to parliamentary approval.
The package covers social media platforms, file-sharing services, pornography sites and other online services. Under the new duties, those services would have to ban users who generate, upload or share child sexual abuse material, extend the use of technology that detects and removes known illegal images, and in some cases deploy automated tools to identify content that has never been seen before.
The timing is deliberate. Britain announced in June that platforms such as TikTok, Instagram, Snapchat and X would be barred from providing services to children under 16, and that ban is set to come into force by spring. Ofcom's announcement lands ahead of it, and the regulator said that because many sites will soon be legally required to run age checks, it is not progressing with some measures it had previously explored.
The backdrop is a policy wave rather than a single British decision. Countries from Australia and Britain to France, Denmark and Turkey are moving to restrict children's access to online platforms and impose tougher obligations on technology companies over content moderation and child safety. Reuters reported on November 16 that the new rules are expected to take effect from early 2027, subject to parliamentary approval, and that Ofcom framed the measures as a response to the scale of the material online.
Key Facts
The first duty is removal of offenders, not just removal of files. Ofcom said the measures would require social media platforms, file-sharing services, pornography sites and other online services to ban users who generate, upload or share child sexual abuse material. The government's announcement put the intent plainly: services in scope of the Online Safety Act should boot off predators who share child sexual abuse material, and stop them from creeping back.
The second duty extends hash matching. Hash matching compares uploaded images and videos against databases of known illegal content; when there is a match, the platform can remove the material and stop it from spreading without anyone having to see it first. Under the new rules, a wide range of services, pornography sites among them, would have to use the technology to identify and remove known child sexual abuse material. The same hash-matching technology will also have to be used to stop terrorist material from all other platforms.
The third duty is proactive detection. Ofcom said services would in some cases have to deploy automated tools to identify previously unknown content. Resultsense reported on November 17 that every file-sharing service and the largest social networks must use automated tools that proactively detect and remove child sexual abuse material not yet on any list. The Independent reported on November 16 that the regulator is requiring platforms to go further because of the scale of the material online, which it says is made worse by AI, and its effect on victims.
The package also tightens grooming protections. Ofcom said it is setting new standards so that existing anti-grooming rules are implemented comprehensively and all accounts held by children are protected by default. Services covered by its children's codes must either run robust age verification to identify children's accounts or extend their anti-grooming safeguards to all users, and certain platforms must take further care before their algorithms recommend content to children.
Enforcement carries weight. IBTimes UK reported on November 16 that the measures could force technology giants to crack down or face fines worth up to 10% of their global revenue under the Online Safety Act. Oliver Griffiths, director of Ofcom's Online Safety Group, said: "While we have seen progress from industry on tackling child sexual abuse on their platforms, we are not satisfied that enough is being done." Online Safety Minister Ian Murray said the online world had been left to police itself for too long and that more platforms will now have to hunt down child sexual abuse material and terrorist content rather than wait for it to be reported.
Analysis
The numbers explain why the regulator moved. Research by Edinburgh-based Childlight, the Global Child Safety Institute, found that reports of AI-related child sexual exploitation increased by more than 8,400% in just two years, rising from approximately 4,700 reports in 2023 to more than 400,000 in 2025. The Internet Watch Foundation assessed 6,310 AI-generated images that met the legal definition of child sexual abuse during the first half of 2026, already exceeding the more than 4,500 it identified throughout 2025. Ofcom's own notes cite Internet Watch Foundation research from its AI CSAM Report 2026 recording a 26,385% annual rise in AI-made abuse videos online. The NSPCC put the UK figure at roughly 100 abuse-image crimes recorded by police every day.
What this really means is that the regulatory question has shifted from whether platforms respond to abuse reports to whether they are structurally organised to find material nobody has reported. Hash matching only works against known material held in a database, so extending it to pornography sites closes an obvious gap but does not touch the novel output that generative systems now produce at volume. That is why the third duty, proactive automated detection on file-sharing services and the largest social networks, is the one with the longest reach and the hardest implementation path.
The enforcement mechanics matter as much as the duties. A ban on a user who shares abuse material is only meaningful if the service can recognise the same person returning, which is why the government's language about predators being kept off sits alongside the removal obligation. Similarly, the 10% of global revenue penalty gives Ofcom a credible lever, but the regulator's own framing, that it is not satisfied enough is being done, signals that it regards voluntary progress as insufficient rather than as a foundation to build on.
The bigger picture here is that Britain is assembling a stack of interlocking obligations rather than a single ban. Age checks arrive through the under-16 restrictions by spring, grooming safeguards are being tightened so children's accounts are protected by default, and algorithm recommendations to children face extra precautions. Ofcom's decision not to progress with some previously explored measures, precisely because age checks will soon be mandatory, shows the regulator treating those strands as substitutes in places and complements elsewhere. The risk is timing: duties that take effect in early 2027 will be judged against a harm landscape that AI is changing faster than any parliamentary timetable.
Why It Matters
For families and victims, the shift to proactive detection changes what a platform owes before anything is reported. The government's announcement describes technology that spots material before it spreads, so that a platform can remove content without anyone having to see it first. Combined with the obligation to remove the people who share it, the package targets both the file and the distributor, which is the difference between cleaning a service repeatedly and denying offenders a place to return to.
For companies, the scope is unusually broad. Social media platforms, file-sharing services, pornography sites and other online services all fall within the measures, and the hash-matching duty for terrorist material extends to all other platforms. Firms that treat this as a compliance exercise in image filtering will miss the parts about user bans, default protections for children's accounts, and algorithmic caution before recommending content. The 10% of global revenue ceiling is the number boards will focus on, but the operational requirements are where the cost sits.
For policy, the package is a test of whether duties written before the worst of the synthetic material arrived can hold as it scales. The Independent reported on November 16 that Ofcom expects the measures to come into effect from early 2027, subject to parliamentary approval, and that the regulator wants existing anti-grooming rules implemented comprehensively. If the duties survive that process intact, Britain will have moved further than most jurisdictions in requiring services to find unseen abuse rather than wait for it, and the results will be watched closely by governments weighing similar restrictions.
Next Up
The measures now go through Parliament, and the scrutiny there will determine how much of the package survives unchanged. If approval comes, the duties are expected to apply from early 2027, which puts implementation work for platforms, pornography sites and file-sharing services on a clock that runs through next year. Ofcom has said it will expect serious consequences for firms that fail to act, and Griffiths has warned that companies which do not comply should expect to face exactly that.
The nearer milestone is the under-16 social media ban, due by spring, which will change the compliance landscape before the abuse duties even start. How Ofcom sequences age checks, grooming standards and proactive detection, and whether the largest social networks can show automated tools working against material that appears on no list, will decide whether this package is remembered as the moment platforms were made to hunt abuse or as another set of obligations that arrived after the harm had already scaled.
Comments (0)
Log in or sign up to leave a comment.
No comments yet. Be the first to share your thoughts.