Today’s Focus

UK Prime Minister Keir Starmer told Apple and Google on Monday that they have three months to build device-level safeguards preventing children from sending or receiving nude images, according to the BBC.

The order, announced as part of a wider online safety package, would require operating system makers to detect and block intimate images on devices used by under-18s. The BBC reported that ministers want the controls active by early September.

The Wall Street Journal said the policy puts the burden squarely on device manufacturers rather than individual apps, marking a shift from the approach taken by the Online Safety Act, which focused on platforms. Starmer framed the move as a response to a surge in self-generated child sexual abuse material flagged by the Internet Watch Foundation.

Bloomberg reported that the announcement comes alongside a planned ban on social media use for children under 16, which Starmer is expected to formally unveil within days. Downing Street has not confirmed the exact legal mechanism.

The Guardian reported that the White House has privately urged the UK not to proceed with the under-16 ban, arguing it would harm American tech companies and set a precedent other governments could follow. Neither Apple nor Google has publicly committed to meeting the September deadline.

New Scientist noted that both companies already offer on-device nudity detection tools through their parental control systems, though usage rates are low and the features are opt-in rather than default. The UK proposal would make detection mandatory on accounts registered to minors.

The Debate

Supporters argue

Starmer told reporters the technology already exists and that companies have no excuse for delay, according to the BBC. He said the government would legislate if the firms refused to act voluntarily.

Technology Secretary Peter Kyle said the policy targets a “pipeline of harm” that begins with image sharing and ends in exploitation, the WSJ reported. He pointed to Internet Watch Foundation data showing self-generated abuse material rose sharply among children under 13 last year.

The National Society for the Prevention of Cruelty to Children (NSPCC) endorsed the proposal, telling the BBC that device-level controls are more effective than app-by-app rules because children migrate between platforms quickly. The charity said voluntary measures have failed to stem the flow of harmful content.

Conservative shadow ministers offered qualified support, with shadow home secretary Chris Philp telling the Guardian the deadline should be enforced with criminal penalties for non-compliance. Australian officials, whose under-16 social media ban took effect last year, told Bloomberg they had shared implementation lessons with London.

Critics argue

The White House warned that the broader package, especially the under-16 social media ban, would damage US tech firms and amount to an extraterritorial restriction on speech, the Guardian reported. American trade officials have raised the issue in ongoing UK talks.

Digital rights group Open Rights Group told the BBC that scanning images on a user’s device, even for a narrow purpose, opens the door to broader surveillance and could be replicated by authoritarian governments. The group said the policy revives concerns raised when Apple shelved a similar client-side scanning plan in 2022.

New Scientist quoted security researchers who said current on-device nudity classifiers produce false positives and could flag innocent photos sent between family members. They warned that a forced rollout in 12 weeks risks shipping immature technology to millions of children.

Some Labour backbenchers told the Guardian they worry the under-16 social media ban will push children toward unmoderated corners of the internet rather than reduce harm, citing mixed early evidence from Australia.

What the experts say

The Internet Watch Foundation, an independent UK charity that catalogs child sexual abuse material, reported a 22% rise in self-generated imagery involving children aged 7 to 10 in 2024, according to figures cited by the BBC. Most of the content was produced on smartphones.

Researchers at the Oxford Internet Institute have found that device-level interventions tend to be more durable than platform-level ones because they survive app switching, but they also raise harder questions about parental consent and false detection rates.

A 2023 report from the European Union Agency for Cybersecurity (ENISA) concluded that on-device scanning can be implemented without sending images off the phone, but warned that the threat model expands every time new categories of content are added.

Sonia Livingstone, a London School of Economics professor who studies children and digital media, told New Scientist that age-verification and device-level controls work best when paired with education, and that bans alone have a poor track record of changing behavior among teenagers.

By the Numbers

3 months: the deadline Starmer set for Apple and Google to implement the new controls, according to the BBC.

16: the proposed minimum age for social media use under the parallel ban Starmer is expected to announce within days, per Bloomberg.

22%: the rise in self-generated child sexual abuse material involving children aged 7 to 10 in 2024, per Internet Watch Foundation figures cited by the BBC.

2022: the year Apple abandoned its earlier plan to scan iCloud photos for child sexual abuse material after a backlash from privacy advocates, per New Scientist.

2024: the year Australia’s under-16 social media ban took effect, the model London is studying, according to Bloomberg.

2 platforms: Apple’s iOS and Google’s Android together account for nearly all smartphones used by UK children, per Ofcom data referenced by the WSJ.

Sources

Get the briefing in your inbox every morning.

Subscribe