Meta has hit pause on the rewrite of its Meta business terms that publishers feared most, but the fight over it has only spread. In the space of three days, a court in Oslo blocked the change and Britain’s competition regulator opened a formal case. Meta, for its part, says the old contract stays in force while it consults. The stakes, as the Competition and Markets Authority put it on October 30, are terms that “could allow it to collect content from the websites of businesses and publishers that use its services” and use it for its own products, including AI features, “without those businesses getting a choice.”
Most coverage has treated this as a publishers-versus-Meta story. However, there is a second, more practical question for anyone who runs a website and a Facebook or Instagram page: what can you actually control today, while lawyers and regulators argue over the Meta business terms? To answer it, we lined up the legal challenges side by side, then read Meta’s own crawler documentation bot by bot.
Four venues, four very different powers
The challenges are not equal. A court injunction binds Meta now. A regulator’s investigation can take months and may end with nothing. A complaint is only a request. Here is where each challenge to the Meta business terms stands as of November 1.
| Venue | Who brought it | Status on Nov 1 | What it can do |
|---|---|---|---|
| Oslo District Court, Norway | Ten Norwegian media groups, backed by the MBL trade body | Temporary injunction granted Oct 28; Meta ordered to pay NOK 1.2 million in costs | Bars Meta Platforms Ireland from bringing in the change until a final ruling |
| UK Competition and Markets Authority | Concerns from the News Media Association and Movement for an Open Web | Formal investigation opened Oct 30 | Interim measures, commitments, or a fine after an infringement finding |
| European Commission | European Publishers Council and Movement for an Open Web | Complaints filed; no public action yet | Could open a case under EU competition rules or the Digital Markets Act |
| Meta itself | Response to the pressure | Rollout paused Oct 28 pending consultation; old terms still apply | Can restart the rollout with a new date at any time |
The Norwegian order is the only one with immediate legal force. According to the Norwegian Media Businesses’ Association, NRK, TV 2, Schibsted, Amedia, Polaris, Aller, DN, Mentor Medier, NTB and Hordaland Folkeblad filed the request, and the Oslo District Court ruled for them on October 28. Schibsted says a total of 180 companies stand behind the action. The ruling came without Meta being heard, so Meta can still ask for a new hearing.
Why the UK case reaches the current Meta business terms
The CMA case is slower but broader. It uses Chapter II of the Competition Act 1998, which means the regulator must first decide whether Meta holds a dominant position in a UK market. Only then does it ask whether the terms are an abuse. Notably, the CMA’s case page covers Meta’s “existing commercial terms, or revisions to those terms,” as PPC Land pointed out. In other words, pausing the rewrite does not end the inquiry.
The regulator also has a faster tool it has not used yet. The Act allows interim measures “where temporary action is needed to prevent significant damage.” Meanwhile, the theoretical ceiling for a finding of abuse is a fine of up to 10% of worldwide turnover. Nobody should read that as a forecast, though. The CMA says no decision has been made, and its first round of evidence gathering is scheduled to run until January 2027.
Meta rejects the premise. A spokesperson told Reuters there were “no grounds” for an investigation and added, “We’ve already committed not to change our terms without consulting publishers and other stakeholders.”
Five crawlers, and robots.txt only governs some of them
Here is the part most site owners have not seen. Meta has said publishers can still use robots.txt to signal their preferences. Yet the UK magazine body PPA notes that this assurance “does not appear in the terms.” So the file is a technical request, not a contract right. Moreover, Meta runs more than one bot, and its web crawler documentation treats them differently.
| Crawler (user agent) | Purpose, per Meta | Robots.txt, per Meta | Cost of blocking it |
|---|---|---|---|
| facebookexternalhit | Builds the title, description and thumbnail preview when your link is shared | Might bypass it during security or integrity checks | Shared links can lose their preview cards |
| meta-externalads | Improving advertising and other business-related products | Respects a disallow rule | None stated by Meta |
| meta-externalagent | Training foundation AI models or indexing content directly | Respects a disallow rule | None stated by Meta |
| meta-webindexer | Improving Meta AI search results | Allowing it helps Meta cite and link your pages | Fewer citations in Meta AI answers |
| meta-externalfetcher | Fetching single links a user asks for, including for AI agents | May bypass it, because a user requested the fetch | Blocking may not work |
Two details stand out. First, two of the five crawlers come with a written caveat that they may bypass robots.txt in some situations. Second, the crawler whose stated purpose sits closest to the disputed wording is Meta-ExternalAds. The PPA says the proposed Meta business terms would let Meta use crawled content to develop its “advertising and business products.” That is our reading of two documents side by side, not a statement from Meta, but it suggests that blocking only the AI training bot may not cover the use publishers are most worried about.
What a small publisher can set up before the Meta business terms return
For a small news site or creator blog, the safest move is to separate the bots you need from the ones you do not. Link previews are the main reason to keep facebookexternalhit, because without it a shared article can lose its title, description and thumbnail card. By contrast, nothing in Meta’s documentation says blocking the training or ads crawlers changes how your links display.
Based on Meta’s documented user agents, a robots.txt section like this one blocks the two crawlers tied to AI training and ad products while leaving previews alone:
User-agent: meta-externalagent
Disallow: /
User-agent: meta-externalads
Disallow: /
User-agent: facebookexternalhit
Allow: /There are trade-offs. If you block Meta-WebIndexer as well, Meta says it is less likely to cite and link your pages in Meta AI answers. Also, Meta warns that crawlers can cache robots.txt for up to 24 hours, so changes are not instant. Finally, check your server logs after a week, because a rule only proves itself when the visits stop.
This is the same logic we used when looking at how Discord’s Google search opt-out works: the switch exists, but it only covers what the platform says it covers.
The Google precedent publishers want copied
Publishers already have a template for what they want. In June, the CMA used its separate digital markets powers to require Google to let publishers opt out of having their content used to train and ground AI features such as AI Overviews, without being punished in normal search. Those main obligations begin on December 3.
The CMA cannot simply copy that order for Meta. Google has “strategic market status” under the 2024 digital markets law, while Meta does not. As a result, the regulator has to prove dominance and abuse the slow way. The next set of status proposals goes to the CMA Board in spring 2027, so a Google-style rulebook for Meta is at least months away.
That gap is why the PPA wants any new version of the Meta business terms to separate access to Meta’s services from consent to crawling. Put simply, a publisher should be able to run a Facebook page without signing away its archive. The same pressure is visible elsewhere: Ofcom recently won its court fight over platform data demands, a sign that UK regulators are winning early tests against the largest platforms.
Dates to watch on the Meta business terms
November 20 was Meta’s second target date, and it no longer holds; Meta says the consultation could run past it. The more useful markers are a possible Meta request for a new hearing in Oslo, any CMA move on interim measures, and the end of the CMA’s first phase in January 2027.
Separately, the CMA is asking Meta about Meta One, its new business subscription, after small firms raised concerns about limits on posting external links without paying. That is outside the formal case. Still, it touches the same dependency from the other side: one fight is about what Meta can take from your site, and the other is about how easily you can send readers there. Platforms are tightening access on several fronts, much as Reddit’s login wall for logged-out readers showed in late October.
Until one of those dates arrives, the Meta business terms in force are the old ones. Site owners who care about AI reuse have a window to set their robots.txt rules deliberately, rather than finding out later what the default allowed.
