TikTok and YouTube have failed to commit to making significant changes to their recommender systems to stop kids from seeing harmful content on their platforms, according to the UK’s online safety regulator, Ofcom, which expressed concern about a lack of industry action on Thursday.

The regulator made the remarks in a press release alongside the publication of evidence showing that nearly three-quarters of children aged between 11 and 17 saw harmful content online, with little change since the UK’s Online Safety Act (OSA) came into effect.

The OSA, the country’s equivalent of the EU’s Digital Services Act (DSA), has been implemented in phases. But safety mechanisms to protect children from a range of harms, including those posed by algorithms that promote harmful content, came into force last summer.

Ofcom said TikTok and YouTube have claimed that their content feeds are already safe for children and do not require any changes. Yet the regulator found that personalised feeds are the main reason for children’s exposure to harmful content. It also said that 35% of children reported seeing harmful content while scrolling.

Back in February, the European Commission also found fault with TikTok’s addictive design features – including its highly personalised recommender system – although the DSA breach finding was preliminary and enforcement remains ongoing in the EU.

Independent audit

Ofcom previously sent legally binding requests for information to Meta, which owns Facebook and Instagram, as well as TikTok and YouTube, asking about the platforms’ content feeds.

On Thursday, it said it is continuing to investigate its responses and is setting out a new action plan to dial up its engagement with platforms.

Notably, the regulator said it’s “exploring” its inspection powers under the OSA, which could mean platforms will face an independent audit. Ofcom also said it could issue “remote inspection notices” that will allow it to observe how content moderation, detection and age checks work in practice.

Ofcom has warned that if platforms are found to be failing to comply with the rules, the next step would be enforcement action. The OSA allows for penalties of up to 10% of a sanctioned company’s global turnover for confirmed breaches.

In another finding from its research, Ofcom said none of the platforms that set a minimum age of 13-years-old for using their services was effectively enforcing it. It found that 84% of kids aged 8 to 12 were still using YouTube, Facebook, TikTok, Instagram and Snapchat.

In a similar finding in April, the EU Commission said Meta had not taken sufficiently strong steps to prevent under-13s from accessing Facebook and Instagram.

Safety measures

Separately, Ofcom announced that Snap, Meta and video-gaming platform Roblox have committed to rolling out anti-grooming safety measures aimed at protecting kids from stranger danger, including by tightening default contact settings, applying AI detection tools and direct chat controls.

Meta has also pledged to expand its use of 13+ “movie-style” content ratings from Instagram to Facebook in the UK, although Ofcom said it’s too early to determine whether the move would reduce harm in practice.

(nl, cm)