Know what’s on your site before a plaintiff’s firm does
A free privacy audit shows which cookies, pixels and trackers are running — and which ones commonly appear in privacy claims.
Three of the largest services regulated under the UK’s Online Safety Act sat in front of peers on November 17, 2026, and spent nearly an hour and a half explaining what they have changed. The House of Lords Communications and Digital Committee had brought in Google, Meta and Roblox as part of its inquiry into whether the Act is making people in the UK safer online. The backdrop was uncomfortable for the companies. Ofcom’s own statutory report, published a month earlier, found no significant overall change in the share of children encountering harmful content in the first nine months of the children’s safety duties. And the regime’s biggest remaining change, an age-based restriction on social media for under-16s, is still being drafted.
TL;DR
- The Lords Communications and Digital Committee took evidence from Google, Meta and Roblox on November 17, 2026, as part of its inquiry into the Online Safety Act’s implementation and impact.
- Ofcom reported on October 22, 2026, that there was no significant overall change in the proportion of children encountering harmful content in the first nine months after the children’s safety duties took effect in July 2025.
- The under-16 social media restriction is not yet law in force: the power to impose it was inserted into the Act in April 2026, and the government says the first regulations will be laid by the end of 2026 for implementation in spring 2027.

Who gave evidence, and what peers wanted to know
The committee’s session page lists three witnesses: Katie O’Donovan, Director for Government Affairs and Public Policy at Google UK; Tim Scott, Senior Director for Public Policy, EMEA, at Roblox; and Rebecca Stimson, Director of Public Policy, UK, at Meta. The committee said the witnesses would be asked what they have changed because of the Act and why those changes “do not appear to have so far reduced users’ exposure to harmful content.” The published areas for discussion were how far the platforms prioritise innovation in safety, how effective their age verification is, what they do to find and remove harmful content proactively, and how they engage with Ofcom.
The hearing followed earlier sessions in the same inquiry with campaigners, academics, lawyers and former ministers who helped shape the Act. Written evidence closed on September 21, 2026. The official transcript of the November 17 session had not been posted on the committee’s site when this article was written, so the account below relies on the parliamentary broadcast record and published reporting of the session, and should be read as a summary rather than a verbatim record.
What the platforms told the committee
The companies largely argued that the Act has changed how they work, even if outcome data has not yet moved. Google’s representative described the Act as one of the most complex pieces of safety regulation in the world and pointed to the risk assessment process, which she said had required Google to produce roughly twenty assessments. Meta’s representative said the company’s proactive removal rate for harms such as terrorism and child sexual exploitation is in “the high ninety percentiles,” meaning most material is caught before anyone reports it. She also described the Act as a floor rather than a ceiling for what Meta does.
Age assurance took up much of the session. Roblox said it had rolled out facial age estimation to gate chat as well as some content and features, and described separate experiences for younger users. Google said SafeSearch is locked for users it believes are under 18. Meta argued that age checks should also happen at the level of app stores and operating systems, suggesting that app stores may hold signals such as bill-payer information that help establish who a parent is. Peers, for their part, pressed the companies on grooming risks, suicide and self-harm content in search, illegal gambling results and why so much detection still depends on users and parents reporting problems.
Proactive detection was a recurring theme. Both Google and Roblox described automated, AI-based moderation as the biggest change in how violative content is found. That aligns with the direction regulators are pushing: Ofcom’s new measures on proactive detection of child sexual abuse material would extend hash matching and automated tools to more services from early 2027, subject to parliamentary approval.
The evidence problem behind the questions
Ofcom’s report on content harmful to children, its first under section 63 of the Act, is the reason the committee framed its questions the way it did. Most children in the UK told Ofcom they had encountered at least one piece of harmful content online in a four-week period. At a total level, there was no significant change in that proportion over the first nine months after the children’s safety duties came into force in July 2025. Ofcom called the finding concerning. It also recommended adding body stigma content and depression content to the list of priority content harmful to children, which would require the government to act.
A flat headline number does not mean nothing changed. It does mean that platform claims about removal rates and new safety features are now being measured against population-level outcomes, and that is the test peers applied on November 17.
What is still to come under the Online Safety Act
The summary circulating about the hearing described the under-16 social media ban as an amendment and as the last remaining piece of the Act’s full implementation, slated for spring 2027. Both points need correcting. The power exists, but the ban itself does not yet. Section 214A, inserted into the Online Safety Act on April 29, 2026, by section 70 of the Children’s Wellbeing and Schools Act 2026, lets the Secretary of State make regulations requiring providers to prevent or restrict access by children of a specified age to specified services or features. The text of section 214A ties that power to the outcome of the “Growing up in the online world” consultation launched on March 2, 2026. The government’s fact sheet says the first set of regulations will be laid before the end of 2026 and implemented in spring 2027. Reporting in mid-November said ministers expected to lay them in December, and that a second tranche covering more features could follow in the spring.
The ban is also not the only outstanding item. Ofcom’s implementation roadmap lists several pieces still in progress:
- Final policy statements on the extra duties for categorised services, covering terms of service, user empowerment, identity verification, fraudulent advertising and protections for news and journalistic content, by mid-2027 at the latest.
- First transparency reports from all categorised services, due in 2027.
- A statutory report on children’s use of app stores by January 2027, which will inform whether app stores are brought into the Act’s scope.
- New code measures on proactive detection of child sexual abuse material, expected from early 2027 subject to parliamentary process.
- Minimum accuracy standards and accreditation for technology that Ofcom can require through technology notices.
Spring 2027 is therefore a milestone, not the end of implementation.
What services in scope should do now
For compliance and risk teams at services used by children, the hearing is a preview of the questions regulators and Parliament will keep asking. A practical sequence:
- Refresh your children’s risk assessment against outcomes, not just controls. Ofcom’s report shows the regulator is looking at exposure rates. Document what your own data says about whether children’s exposure is falling, and why.
- Map your features against section 214A. The power covers specified functionalities, including contact from unknown users, live audio or video with strangers and location sharing, as well as whole services. Identify which of your features could be named in regulations.
- Plan for age assurance at 16, not only 18. The government expects Ofcom to set out options for establishing whether someone is over 16. If you rely on self-declaration anywhere, expect to replace it.
- Treat proactive detection as a data protection project. Automated scanning and facial age estimation both process personal data at scale, and in the case of face scans, biometric data. A data protection impact assessment, retention limits and a clear lawful basis under UK GDPR should be in place before deployment.
- Prepare for 16- and 17-year-old defaults. The government has said livestreaming and contact from strangers will be off by default for this age group, with overnight curfews and personalised feeds and autoplay switched off by default on affected social media platforms.
- Watch the December regulations closely. Scope definitions built around recommender systems and user-generated content could reach services that assumed they were exempt.
Age verification has already been tested in other jurisdictions. In the United States, the Supreme Court’s age-gating ruling upheld a state law requiring age checks for adult sites, and European governments are drawing their own lines, as Italy’s under-13 social media proposal shows. For companies operating across several of these regimes, the youth digital safety compliance window is narrowing at the same time in several places.
Frequently asked questions
Is the UK under-16 social media ban law yet?
No. The legal power to impose it is in force as section 214A of the Online Safety Act, but the regulations that would name the services and features covered had not been laid before Parliament as of November 20, 2026. The government says it will lay them by the end of the year for implementation in spring 2027.
Did the committee find that the platforms broke the law?
No. This was an evidence session in a Lords inquiry, not an enforcement action. Ofcom enforces the Act. The committee will publish its own conclusions and recommendations in a report later.
Why does a privacy team need to care about an online safety hearing?
Most of the safety tools discussed, including facial age estimation, device-level age signals and automated content scanning, involve processing personal data, often of children. Those choices fall under UK GDPR and the ICO’s Children’s Code as well as the Online Safety Act, and the two regimes need to be designed together.
Will researchers get access to platform data?
Peers raised independent researcher access during the session. Ofcom has said the Data (Use and Access) Act 2025 allows the government to create a framework for it, but no such scheme is in place yet. Some companies have moved on their own; for example, one AI developer has opened usage data to independent researchers.
The floor is about to rise
Meta told peers the Online Safety Act is a floor. That floor is about to rise again: a ban for under-16s, new defaults for 16- and 17-year-olds, more proactive scanning and a full set of duties for the biggest services, all in 2027. The companies came to Westminster to show their work. The committee’s message, and Ofcom’s data, is that showing activity is no longer enough. Services now have to show results.
Put this into practice
Turn privacy guidance into working controls.
Captain Compliance helps teams discover tracking technologies, enforce consent, manage privacy requests and document the evidence behind every decision.