A child creates a TikTok account, enters a date of birth that makes them 16, and is immediately shown videos about self-harm, eating disorders, or pornography. The platform's technology—designed to guess how old they really are—didn't catch the lie. This is the scenario that has led the UK's media regulator, Ofcom, to launch a formal investigation into whether TikTok is doing enough to keep children off its platform.
At the heart of the investigation is a technique called “age inference.” Instead of requiring robust proof of age like a passport scan, some social media platforms estimate a user's age by analysing signals such as the videos they watch, the accounts they follow, their nickname or biography, and even their voice or facial features. Ofcom has “serious doubts” that this method is good enough to meet the legal standard of “highly effective” age assurance required under UK law.
“Ofcom investigates TikTok's use of age inference to check whether children are protected from harmful content under UK law.”
The investigation follows a review by Ofcom in May 2026 that criticised TikTok for not being “safe enough” for children. Almost a year earlier, in July 2025, the children’s protection codes of the Online Safety Act (OSA) had come into force. These codes require platforms that host content harmful to children—such as pornography, content promoting self-harm or suicide, and material about disordered eating—to use highly effective age checks to prevent under-18s from seeing it. Ofcom has already issued large fines against dozens of adult sites for non-compliance. Now scrutiny has turned to social media.
TikTok says it requires users to enter a date of birth when they sign up and uses age inference technology to look for “indicators that someone may not meet our minimum age requirement.” A company spokesperson said it has invested “billions” in online safety since launching in the UK eight years ago and is “confident that we meet our Online Safety Act obligations.” But Ofcom’s group director for strategy and research, Kate Davies, told the BBC that the regulator has “very serious questions about whether age inference can be highly effective.”
For UK readers, especially parents and guardians, this investigation matters because it affects how—and whether—children are protected on the platforms they use. The government has also announced plans to ban under-16s from a range of social media platforms entirely, starting early next year. If age inference is found to be failing, platforms could be forced to adopt stricter methods, such as face scans or official ID checks, which raises questions about privacy and access.
Here are the key questions answered:
Q: What is age inference? Age inference is a technology that estimates a user’s age based on signals from their activity and profile. These signals can include the content they view, their nickname, biography, voice, or facial features. It is different from verifying age with official documents or biometrics.
Q: What does the Online Safety Act require from platforms? The Act’s children’s codes, effective from July 2025, mandate that platforms must use “highly effective” age assurance to prevent children from encountering harmful content. Failure to comply can lead to fines of up to £18 million or 10% of global annual revenue, whichever is greater. Ofcom can also apply to have sites blocked in the UK.
Q: What happens if TikTok is found to have broken the rules? Ofcom has not yet reached any conclusions. If the investigation finds TikTok failed to comply, it could face a substantial fine and be required to change its age-checking methods. In the most serious cases, Ofcom could seek a court order to restrict or block access to the platform in the UK.
What happens next? Ofcom’s investigation is ongoing, with no set deadline for a final decision. Meanwhile, the UK government is preparing to implement a social media ban for under-16s early next year, which will increase pressure on all platforms to get age verification right. TikTok has said it will work with Ofcom to demonstrate compliance, but the regulator has urged any platform using age inference to “switch to other methods listed in our guidance as highly effective without delay.”