A London court heard a three-day challenge from Meta, TikTok and X to Ofcom's information notices, which opened on 5 October 2026. The notices, issued in February under the Online Safety Act 2023, demand granular moderation data: how many posts were removed or had their visibility restricted, and how many users saw harmful content. According to The Star's report of the hearing, X's witness statement called it "the most burdensome information request X has received from any regulator in any jurisdiction."
This is among the first court tests of the Act's enforcement machinery. The judgment will matter well beyond these three companies.
The strongest case for Ofcom
Ofcom's position deserves a fair hearing. A regime that has never been measured cannot be judged. Ofcom says Parliament charged it with regulating an industry that was unregulated and unaccountable for more than 20 years, and a regulator cannot assess whether safety systems work without data on what they remove and what users still see. Ofcom says the data is "genuinely needed" to evaluate the new regime, and that it narrowed the notices before the hearing.
There is also a transparency argument. Platform-published figures are self-selected and not comparable across services. If every company defines "harmful content" and "views" differently, neither Parliament nor the public can compare them. Ofcom said in July 2024 that it would issue transparency notices to categorised services, requiring disclosure of how prevalent illegal content is and how many users come across it. These are exactly the kinds of figures now in dispute.
Where the companies have a point
The companies' objections are not all special pleading, though some surely are. Three distinct arguments surfaced, as reported by The Next Web:
- Purpose. Meta says Ofcom wants "wide-ranging and granular information" about seven of its services without a clearly defined regulatory purpose.
- Burden. X says the demand is out of line with anything it faces elsewhere.
- Process. TikTok argues Ofcom bypassed an alternative monitoring regime that carries its own safeguards.
The process point is the most interesting in legal terms. The Act sets up a dedicated transparency-reporting route, in which Ofcom issues notices to categorised services and specifies what must be reported. If the same metrics can be extracted through general information-gathering powers, the structured route and its protections become optional. That is a reasonable administrative-law worry whichever way the court rules on these specific notices.
Why proportionality is not a loophole
Ofcom's powers are wide. Ofcom's own explanation of the regime says they include audits, remote inspection of algorithms, and entry to UK premises, and that non-compliance can draw fines of up to £18 million or 10% of worldwide revenue, whichever is higher. Powers of that scale should come with a clear statement of why each demand is needed and what will be done with the answer.
The case for insisting on this is practical as well as legal. Granular moderation metrics are easy to misread. A higher removal count can mean better enforcement or worse content. A lower exposure figure can reflect better systems or simply a different measurement method. If Ofcom collects data without a published methodology and a stated analytical use, it risks the worst of both worlds: heavy compliance costs for firms, and numbers that tell the public little.
The costs are also uneven. Meta, TikTok and X have large compliance teams. A mid-sized forum or a startup that receives a comparable notice will not. Precedent set in a fight between giants becomes the template applied to everyone else, so the court's reasoning on purpose and proportionality matters for smaller services that are not in the room.
There is a speech angle too. Moderation metrics put pressure on platforms to show high removal numbers. Regulators that reward removal volume invite over-removal of lawful speech, which is the opposite of what a proportionate regime should produce. Data requests should be designed to measure accuracy and error rates, including wrongful takedowns, not only the quantity of content taken down.
What a good outcome looks like
The court need not choose between a regulator that cannot see and one that sees everything. A sensible resolution would have three features:
- Ofcom states, for each category of data, the supervisory question it answers.
- Data that fits the structured transparency-reporting route goes through that route, with its safeguards.
- Requests include error and appeal-overturn rates, so the public sees over-removal as well as under-removal.
A separate Meta challenge to how Ofcom calculates Online Safety Act fees is due in court the week after this hearing, according to The Next Web. Taken together, the cases will shape how much discipline the UK regulator must apply to its own powers. Ofcom is right that the regime needs evidence. The companies are right that evidence should be gathered for a stated purpose, through the route Parliament designed. Both can be true, and a judgment that says so would strengthen the Act rather than weaken it.
No judgment had been reported at the time of writing, so the above is analysis of the arguments, not of a ruling.