TikTok and YouTube are still feeding harmful content to 73% of UK teens

TikTok and Alphabet's YouTube have failed to set out meaningful steps to protect British children from harmful online content, media regulator Ofcom said on Thursday, citing data showing widespread exposure on their platforms.

Ofcom said neither company had made significant new commitments to make recommendation feeds safer, despite evidence that these feeds are the main route through which children encounter harm.

Governments are stepping up efforts to improve child safety online, with Prime Minister Keir Starmer urging social media firms to take greater responsibility.

Britain is consulting on tighter curbs, including a possible ban on under-16s using social media, modelled on Australia's landmark move, to tackle what it calls addictive design features.

73% of UK teens exposed to harmful content via recommendation feeds

Ofcom research found 73% of 11- to 17-year-olds were exposed to harmful content over four weeks, mainly via personalised feeds. TikTok was cited most often, followed by YouTube, Meta's Instagram and Snap's Snapchat.

For instance, Cybernews has previously reported that nearly half of porn users in the UK have admitted to watching content on sites that evade age verification checks, leaving them vulnerable to viewing illegal material.
Today Ofcom said TikTok and YouTube maintained their existing systems were sufficient, but the regulator said its evidence suggested their feeds "are still not safe enough".

A YouTube spokesperson said: "YouTube provides industry-leading, age-appropriate, high quality experiences for young viewers, working with child safety experts to deliver protections that support millions of families across the UK.

"We welcome today's news that others across the industry will soon adopt similar features."

A TikTok spokesperson said it was "very disappointing that Ofcom has failed to acknowledge both our longstanding and newer safety features."

"We will continue to make ongoing investments in safety measures for our users," the person added.

Snap, Meta and Roblox have all agreed to introduce stronger protections against online grooming following Ofcom's demands last month.

Under those commitments, Snap will block adult strangers from contacting children by default and expand age checks in Britain. Meta plans new controls on teen accounts and AI tools to detect suspicious conversations. Roblox will allow parents to disable direct messaging for under-16s.

84% of 8-12 year olds bypass age restrictions

Nearly a year after new child safety duties under the Online Safety Act took effect, Ofcom said there had been little overall improvement in children's exposure to harmful content.

The regulator said major platforms dominate children's online activity, with YouTube used by 67% of children and TikTok by 60%, and 95% using at least one social media or video-sharing service.

It also flagged weak enforcement of minimum age rules, noting 84% of children aged eight to 12 use services requiring users to be at least 13.

Cybernews has previously reported that British children are finding creative ways to bypass age verification systems despite the UK Online Safety Act. A study by Internet Matters found that nearly a half of children believe online age checks are easy to circumvent, with a third admitting they've done so. Methods include using VPNs, fabricating data, borrowing parents' IDs, and even drawing fake mustaches with eyebrow pencils to trick facial recognition systems.

Ofcom said current legislation does not clearly require companies to keep underage users off their platforms and urged the government to strengthen the law.

Curious what others think about this story? Contribute your thoughts to the debate below.

Separately, Elon Musk's X has agreed to step up enforcement against illegal hate speech and terrorist content, including reviewing such material within 24 hours on average and sharing quarterly data with the regulator.

Ofcom is still examining X's systems and its Grok chatbot after a Reuters report earlier this year found the tool could generate sexualised images in many cases despite user warnings.

Unlock more exclusive Cybernews content on YouTube.