Meta's new business terms ask businesses that keep using its business tools, including news publishers, to accept sweeping crawl-and-reuse rights or stop using those tools. The dispute shows how far the UK's competition regime has come, and where it still falls short.
What Meta is asking for
According to Press Gazette, the updated terms let Meta, or anyone acting for it, use automated tools to access, copy and crawl sites linked to its products. Meta may use that content to "provide, develop, improve, research, and maintain" its products. The original wording referred to training AI. Meta has since removed the word "train", and a US media lawyer quoted by Press Gazette argues that "develop" is still broad enough to cover it.
The News Media Association's chief executive, Theo Bamber, called it a "shameless landgrab attempt". Movement for an Open Web has asked the European Commission and the CMA for interim measures, and the European Publishers Council is writing to Commission executive vice-presidents Teresa Ribera and Henna Virkkunen.
Meta's reply is that the crawling terms it consolidated "only applied to our ads and business products". It also says publishers can signal preferences through robots.txt, and that the update does not override them.
The original 30 October deadline has moved. Press Gazette's 23 October update reports that Meta has delayed the change to 20 November. That is a short reprieve, not a resolution.
The strongest case for intervention
Regulators have a serious argument. Publishers cannot realistically leave Facebook and Instagram, which are major distribution channels. If the price of staying is a licence over content published elsewhere, the choice is a coerced one. A take-it-or-leave-it term, imposed by a firm with leverage and with no opt-out, is the kind of conduct competition law exists to scrutinise. Rosewell's warning that harm is "imminent and irreversible once content has been ingested into models" is also fair. Ingested training data cannot easily be withdrawn.
That case deserves a proper hearing. It does not follow that the right response is a ban on crawling, or an assumption that Meta is acting unlawfully before the facts are established.
Where the CMA's powers stop
Sarah Cardell told the Lords Communications and Digital Committee on 20 October that the CMA shares the concerns "at face value" but needs to "understand it better", according to Press Gazette's report. She also acknowledged: "We obviously don't have a live designation in relation to Meta."
That matters because the Digital Markets, Competition and Consumers Act 2024 works through strategic market status. A firm is designated first. Only then can the CMA impose tailored conduct requirements. Without a designation the CMA is left with ordinary competition enforcement, which is slower and case-by-case. Cardell named it as an option, alongside informal and formal engagement.
The NMA has urged the CMA to begin an SMS investigation if Meta does not publicly commit to dropping the contested elements. That is a request from the trade body, not a CMA commitment.
The Google precedent shows what proportionate looks like
The CMA has already shown what a targeted remedy looks like. On 3 June 2026 it imposed a publisher conduct requirement on Google's search services. Publishers can opt out of having their content used to power AI features such as AI Overviews, and out of having it used to fine-tune Google's AI models. Cardell called it a "world-first requirement". Google has nine months to implement all the changes.
The requirement itself lists five obligations:
- effective publisher controls over generative AI use of search content;
- clear information on how that content is used;
- engagement metrics for publishers;
- accurate, clear attribution in general search;
- published information on Google's approach to attribution.
Notice what this does well. It starts from a designation process and an evidence record. It targets a specific harm, the lack of control over AI reuse. It leaves Google free to build AI products, and gives publishers a choice without forcing either side out of the market. That is the pro-innovation way to correct a bargaining imbalance.
What a proportionate response looks like for Meta
The facts are still contested. Meta says the terms cover ads and business products only, and that robots.txt preferences are honoured. Legal experts disagree about what "develop" covers. A regulator should resolve that ambiguity before acting. The CMA's urgent information request and its meeting with Meta are the right first steps.
If the facts bear out publishers' fears, the remedy should mirror the Google one. It should be a clear, granular, genuinely usable opt-out from AI reuse, separate from the decision to hold an account on Facebook or Instagram. Publishers should not have to choose between reach and rights over their archive. Meta could also fix much of this itself by putting an unambiguous opt-out and a plain statement of scope into the contract. A voluntary commitment, made public, would be faster than any designation process.
What the UK should avoid is two opposite mistakes:
- Reflexive blanket restrictions on crawling. These would damage search, research and archiving, which all rely on the open web.
- Waiting for designation as an excuse for inaction. Cardell herself flagged competition enforcement as available, and publishers' content is being exposed now.
The wider lesson
The episode shows both the strength and the limit of the UK's ex ante model. The Google case shows it can deliver precise, workable remedies. The Meta case shows the model reaches only firms the CMA has already investigated and designated. Whether Meta should be examined next is a question for evidence, not for a deadline set by a platform's own terms of service.
The 20 November date gives the CMA, the Commission and Meta a few weeks to establish what the terms actually do. Publishers deserve a clear answer. The open internet is best protected by regulators who test claims, impose narrow remedies where harm is shown, and leave room for innovation.