On 21 May 2026, the UK regulator Ofcom told the public which big platforms had agreed to make children safer, and which had not. Snap, Meta and Roblox have committed to specific changes. TikTok and YouTube refused to change how their feeds push content to children, claiming their algorithms are already safe.
This is the report Ofcom said was coming when it wrote to the six biggest platforms used by children back in March. It is the first time a regulator has published a platform-by-platform scorecard of what each company will and will not do to protect children, and it gives parents a clearer picture than they have ever had of which apps are moving and which are digging in.
What the regulator asked for
In March, Ofcom wrote to Facebook, Instagram, Roblox, Snap, TikTok and YouTube with four demands. Stop letting children under the official minimum age use the service. Add stronger protections against grooming. Make the algorithmic feeds that push content to children safer. And stop testing new features on children before checking the risks.
The deadline for responses was the end of April. The regulator has now published what each platform said it would do.
What each platform agreed to, and what they refused
Snap agreed to the most concrete set of changes. Adults will be blocked from contacting children on Snapchat by default. Children will no longer be prompted to add people they do not know to expand their friendship groups. Snap also told the regulator it will roll out age assurance to all UK users over the summer, so the protections apply to every under-18 account, not just the ones where the platform already knows the user is a child. Snap has been identified in research over several years as one of the highest-risk services for grooming, which is why the default setting change matters: it shifts the burden away from the child needing to block strangers and onto the platform needing to let strangers through.
Meta agreed to use AI to detect sexualised conversations between adults and teenagers in Instagram direct messages, and to report flagged accounts to the US National Center for Missing & Exploited Children. It will hide teenagers’ followers and following lists by default, so adults cannot use those lists to find and contact children they do not know. Meta is also extending its “13+ movie style” sensitive content control from Instagram to Facebook, limiting what teenagers see to age-appropriate posts. What Meta did not agree to was Ofcom’s measure that would have stopped Instagram from suggesting connections between children and people they do not know. The regulator wanted Instagram to switch off “people you may know” prompts for children entirely. Meta declined.
Roblox confirmed that parents will be able to switch off direct messaging entirely for children under 16. The platform is also rolling out the new Roblox Kids and Roblox Select accounts in June. Children aged 5 to 8 will get Roblox Kids accounts with all communications features off by default, and 9-to-15-year-olds will get Roblox Select accounts with stricter chat and content settings depending on age. These changes were already announced in April, but Ofcom has now secured the commitment in writing.
TikTok and YouTube refused. Ofcom asked both platforms to set out how they would make their algorithmic feeds, the For You feed on TikTok and the recommendations on YouTube, safer for children. The deadline was the end of April. Neither company committed to any significant change. Both insisted their feeds are already safe for children. Ofcom’s own research, published the same day, found that 73% of 11 to 17 year olds had encountered harmful content online in a single four-week window, and that the platforms children encounter the most harm on are the ones with the biggest reach. TikTok was the platform most frequently named. Ofcom said it is “deeply concerned” by the response and signalled it will use stronger powers under the Online Safety Act to force changes, including the power to require independent reviews of how each platform’s systems actually work.
What this means for you right now
The Snapchat default changes only protect children whose accounts the platform knows are under 18. If your child has a Snapchat account, check that the date of birth on the account is correct. Open Snapchat. Tap the profile icon (top-left). Tap the settings cog (top-right). Tap “Name, Birthday & Birthday Party”. If the year is wrong because your child signed up with a fake birthday, the new protections will not apply to them.
On Instagram, the new AI detection in DMs will run whether you do anything or not. But the followers and following lists will only be hidden if Meta switches the default for your child’s account. Check that your child’s Instagram account is set as a Teen Account. Open Instagram, go to Settings, then “Account type and tools”. If the account is not flagged as a Teen Account, the new protections will not apply.
On Roblox, the option to switch off direct messaging will need to be turned on by you through the Parental Controls. It will not switch on automatically for existing accounts. If your child is under 16 and you would rather they did not message strangers on Roblox, this is the setting to find. Settings → Parental Controls → Communication.
On TikTok and YouTube, nothing has changed and nothing is changing. The platforms have told the regulator their feeds are already safe. The regulator disagrees. The available family controls, TikTok’s Family Pairing and YouTube’s supervised account experience, remain the most useful tools you have, and both are worth setting up if you have not already.
The regulator has taken the unusual step of publishing exactly which platforms moved and which did not. That clarity is worth using.
Sources:
