Power & systems

Norwegian media outlets take Meta to court

Yet Meta has wiped the word "train" from its clause.

6 min read
Norwegian media outlets take Meta to court

Ten names before the Oslo court

On Tuesday, October 27, the Norwegian Media Businesses' Association, MBL, announced an unusual move. NRK, TV 2, Schibsted, Amedia, Polaris, Aller, DN, Mentor Medier, the NTB news agency and the daily Hordaland Folkeblad are asking the Oslo District Court for a temporary injunction. The goal: block Meta's new commercial terms until a judge decides whether they are lawful.

A public broadcaster, big private groups, a wire service. MBL describes an industry standing behind a single filing.

The terms apply to anyone who runs a business page on Facebook or Instagram. They include the clause we picked apart on October 17: Meta will be able to send its crawlers to the websites linked to those accounts, and come back regularly for updates.

That day, we showed that overnight the US English version had lost the word "train", while the French page still said "entraîner". So we went back to read it again.

The word is gone everywhere

On Wednesday, October 28, at 11:18 a.m., from a connection in France, we opened the page with the upcoming terms in five versions: US English, UK English, French, German and Norwegian. None of them contains the verb "train" any more, or an equivalent. All now say the same thing: the crawlers are used to provide, develop, improve, research and maintain "advertising and business products", and to protect the integrity and security of Meta's products.

The edits rolled out unevenly. Wayback Machine snapshots show the UK version changed between October 17 and 20. For French, all we know is that the word was there on the afternoon of October 17 and isn't any more. We found no announcement from Meta about these changes.

So the ten publishers are challenging a crawler clause that no longer mentions training. Pulling a word out of a contract is like taking the sign off a door: you no longer know what the room is for, but the door is still open.

What the clause still allows

The start of the sentence hasn't moved in any language. Crawlers, regular visits, and "test credentials", meaning the means to sign up on a site the way a reader would. According to MBL, every article, photo and video from Norwegian media could be scraped that way, including behind the paywall. There is no box to refuse: the only exit, the association says, is to delete your account.

There is also a gray area the text doesn't settle. Could "developing" and "researching" advertising products cover AI tools that generate ads? None of the five versions says, in either direction.

Meta's answer is a file. In an email to NRK, a spokesperson said content owners will still be able to use robots.txt, the "keep out" notice a website shows to crawlers. On October 17, Meta had already told Nieman Lab, Harvard's journalism lab, that it rewrote the text because it wasn't clear enough that these preferences are honored. Beyond that, Meta declined to comment to NRK.

A promise, not a clause

That is where MBL pushes. The term robots.txt appears in none of the five versions we read. "If Meta believes reservations are being respected, it should write that in the terms," says MBL chair Lene Eltvik Vindfeld. For the association, a spokesperson's assurance isn't enough.

MBL makes two other demands. Any use of journalism to train AI, or for standalone AI products, should go through a separate, voluntary agreement with each outlet. And opting out shouldn't cost audience: no reduced reach for publishers who block the crawlers.

Speaking to NRK, Vindfeld calls it an "abuse of a dominant position". The filing itself hasn't been made public.

Leave, or stay and sue

Why not simply close the pages? "Unfortunately, it isn't that simple," Olav Nyhus, NRK's chief legal officer, told NRK. NRK's mission is to reach everyone, and he says these platforms have become unavoidable for reaching the public. He also notes that NRK publishes content on its pages whose rights belong to others.

Think of a shopkeeper whose store is on the only busy street in town: she contests the new lease rather than moving.

Others do the math the other way. The local paper Fjell-Ljom, in Røros, will delete its Facebook page for good on January 1, 2027, NRK reports. A small title can afford what a public broadcaster can't.

Déclic is in scope too: our articles are shared on a Facebook page and an Instagram account, and these terms apply to us as they do to any outlet that posts its links there.

A delay with convenient timing

The terms were due to take effect on October 30. Meta's page now says November 20, in all five versions. That delay isn't a response to the Norwegian filing: it came first. On Friday, October 23, Meta told Nieman Lab it wanted more time to explain to its partners what is changing and what isn't, robots.txt included.

Thomas Höppner, a lawyer and competition law professor in Germany, reads it differently. He told Nieman Lab he fears the delay is mainly meant to lower the risk of requests for interim measures. A judge asked to act urgently wants to see urgency. The more time plaintiffs let pass after learning of the changes, the easier it is to conclude there is none.

Pushing back the deadline is like moving a fire away from the house: it isn't out, but the call to the fire brigade looks less pressing. The Norwegians announced their filing four days after the delay. MBL and NRK believe they are, to their knowledge, the first in the world to ask a judge to stop these terms.

Elsewhere, people are watching

The European Commission welcomes the delay and says it will keep watch over compliance with EU law, including the rules on abuse of a dominant position and the Digital Markets Act, according to a statement passed to Nieman Lab. In the UK, the competition authority says it is still reviewing the terms. Also per the American lab, two German publishers' associations have asked Brussels to bar the terms from taking effect before a full review.

No hearing date has been made public. To matter, the Oslo decision will have to come before November 20. Until then, Meta can still tweak its page, as it has done language by language since October 16. What no version has written so far is the commitment the Norwegian press is asking for: to honor a website's refusal by contract.

Was this article useful?

Topics covered:

RegulationMetaNews

Frequently asked questions

Who is suing Meta in Norway?
Ten media companies: NRK, TV 2, Schibsted, Amedia, Polaris, Aller, DN, Mentor Medier, the NTB news agency and Hordaland Folkeblad. Their trade body, MBL, announced on October 27 that they are asking the Oslo District Court for a temporary injunction against Meta's new commercial terms until a judge rules on whether they are lawful.
What do they object to in Meta's commercial terms?
One clause lets Meta send its crawlers to the websites linked to business accounts on Facebook and Instagram, come back regularly, and use test credentials. According to MBL, that could pull in every article, photo and video from Norwegian media, paywalled content included. There is no box to tick to say no: the only way out, the association says, is to delete your account.
Has Meta taken AI training out of its terms?
The word is gone, but the question isn't settled. On October 28 at 11:18 a.m., none of the five versions of the page (US English, UK English, French, German, Norwegian) still contained the verb train or its equivalent. But could "developing" and "researching" advertising products cover AI tools that generate ads? None of the five versions says, one way or the other.
What is the Norwegian press asking for?
Three things, according to MBL: that respect for robots.txt be written into the terms rather than merely promised by a spokesperson; that any use of journalism to train AI, or for standalone AI products, go through a separate, voluntary agreement with each outlet; and that saying no carry no penalty in reach for publishers who block the crawlers.
When will the Oslo court rule?
No hearing date has been made public. The terms were due to take effect on October 30. Meta pushed the date back to November 20, and explained why to Nieman Lab on October 23, four days before the Norwegian filing was announced. To matter, a ruling would have to land before that date. MBL and NRK believe they are, to their knowledge, the first in the world to ask a judge to stop these terms.
Katja Liersch

Katja Liersch

Co-founder & Journalist

Katja is a journalist and TV producer. With decades of experience in mainstream media, she brings to Declic Media the perspective of those discovering AI: curious, demanding and pragmatic. She ensures every piece of content truly speaks to everyone.

All articles by Katja →
The free AI newsletter