Britain's communications watchdog Ofcom announced on Monday that it cannot require tech companies to deploy scanning technology on users' personal devices to detect child sexual abuse material, marking a significant moment in the debate over encrypted messaging and child safety. The statement came as the regulator unveiled finalized rules under the Online Safety Act, responding to calls from campaigners who argue that abusers exploit end-to-end encryption to evade detection.

Ofcom stated it "does not have powers under the Act to regulate devices" when responding to those demanding scanning capabilities. The agency's new framework instead focuses on what platforms can do within their own systems rather than on individual hardware.

What the new rules require

Under the finalized guidance, large platforms must deploy tools to identify child sexual abuse material proactively—including previously unknown content without digital fingerprints—but only when users post such material publicly. The rules take effect in February 2027 if Parliament approves them.

For encrypted conversations, Ofcom's updated approach permits platforms to assess entire groups based on profile pictures, group descriptions, or coded language, provided this assessment draws on a British police report. The regulator acknowledged that such enforcement "may interfere significantly with users' privacy."

All platforms hosting user-generated content must permanently remove adults caught sharing abuse material or grooming minors. The rules include carve-outs for victims and potentially for those who report unwanted images to address concerns about harassment. Content depicting terrorism, hate speech, suicide, and self-harm must be excluded from children's recommendation algorithms.

Ofcom decided against imposing restrictions on how other users can comment on, gift to, or record children's livestreams, determining such protections were redundant given the government's proposed ban on social media use by under-16s.

The client-side scanning question

Device-level scanning—running detection tools on users' phones or computers before messages are encrypted—remains one of the most contested issues in child safety policy. Those backing the approach contend it represents the only viable method to uncover abuse material shared through end-to-end encrypted services. Conversely, security experts and privacy advocates warn that on-device scanning undermines the fundamental protections that encryption provides.

Apple tested this method in 2021 with a proposed system to scan on-device content for child abuse imagery. Following backlash, the company abandoned the initiative in 2022.

While Ofcom has not excluded future deployment of client-side scanning, it has deferred the matter to a separate provision in the Online Safety Act. Section 121 grants the regulator authority to issue a "technology notice" compelling companies to employ accredited software to detect abuse material, including in private communications.

Before Ofcom can invoke this power, the technology in question must receive accreditation based on accuracy benchmarks established by government. These standards took effect in September, though Ofcom has not yet confirmed that any tool meets the accreditation threshold.

Limits and exceptions

Ofcom stated its rules would not apply where compliance would "materially compromise the security" of a service. However, the regulator cautioned that substantial implementation costs or necessary design modifications alone would not qualify as valid defenses.

The agency also withdrew earlier proposals, announced alongside the abuse material measures in June 2025, that would have mandated automated fraud detection by platforms. Ofcom indicated it possessed "a high degree of confidence" only in tools designed to identify abuse images and signaled it would seek input on addressing other harms through consultation in the coming year.