Ten Norwegian media groups have asked the Oslo District Court to block Meta's new Commercial Terms for business accounts, saying they would let Meta use journalism, including paywalled content, to build AI products. Euronews reported the filing on October 27, 2026. The terms were due to start on October 30, but Meta has already postponed that to November 20. Nothing has been ruled on, and Meta's reading of its own terms differs sharply from the plaintiffs'. This post sets out who is involved, what each side says, and what is still unverified.
TL;DR
| Question | Answer |
|---|---|
| Who is suing? | 10 Norwegian media groups incl. NRK, Schibsted, Amedia, Polaris, Aller, NTB; MBL involved |
| Where? | Oslo District Court, provisional injunction request |
| Against what? | Meta's new Commercial Terms for business accounts |
| Core claim | Terms let Meta collect articles, photos, videos (incl. paywalled) for AI products and more |
| Meta's timeline | Announced Oct 8; start moved from Oct 30 to Nov 20 |
| Status | Injunction pending; no ruling reported |
What the plaintiffs say
According to Euronews, the plaintiffs argue that the new Commercial Terms apply to all professional account holders, including companies and retailers, and that the only way out is to delete the account. They call this an all-or-nothing choice. Lene Eltvik Vindfeld, head of MBL, is quoted as saying: "The fear is that the companies will use this power to secure rights over journalistic content." She also said Meta's use of media investments to build competing AI services is unacceptable and distorts competition.
The complaint is therefore two-sided: a rights concern (content collected for AI) and a competition concern (a platform that publishers depend on for reach building services that compete with them).
What Meta says
Meta announced the changes on October 8 and set an October 30 start. After pushback from European publishers and antitrust authorities, it moved the date to November 20, saying it needed "more time for us to explain to our partners and stakeholders what's changing and what's not," per Euronews.
Our earlier reporting adds Meta's stated position. The Verge reported that spokesperson Daniel Roberts said the terms were consolidated, that content owners can still use robots.txt to signal preferences, and that the updates do not give Meta new rights that supersede those preferences. We have the full chain in Meta Commercial Terms and the October 30 crawl: claimed vs verified, including Meta's removal of a sentence that mentioned training and legal experts' view that the remaining wording may still be broad enough.
A timeline
| Date | Event |
|---|---|
| October 8, 2026 | Meta announces changes to its Commercial Terms |
| October 14 | MBL and NRK threaten to sue Meta over scam ads that impersonated them (a separate dispute) |
| Mid-October | Meta edits the terms, removing a sentence that referenced training |
| October 22-23 | Legal experts say wording may still allow training; Meta delays the start to November 20 |
| October 27 | Norwegian media groups ask Oslo court for an injunction |
| November 20 | New effective date for the terms |
Rows about the edit and the delay come from our earlier posts and The Verge's reporting; the October 14 and October 27 rows come from Euronews.
Why this matters beyond Norway
Three threads meet in this dispute.
- Terms of service as an AI data pipeline. The fight is not about a crawler. It is about contract language that applies to anyone who uses Meta's business tools, including retailers and publishers' commercial accounts. If broad contractual consent can authorise AI use of content, then the data-access fight moves from copyright law to contract law.
- Europe's regulators. Euronews notes pressure from European publishers and antitrust authorities. In June, France's competition authority warned Meta that its approach to copyright talks could "seriously harm" the press, per the same report. A Nordic court case adds a judicial front to the regulatory one.
- A pattern of opt-outs. Euronews notes that since May 2026, more than 340 news sites, including the Guardian, the New York Times and USA Today, have blocked internet archive crawlers to stop AI training data collection. Publishers are looking for any lever that works, whether technical blocks, lawsuits or licensing deals.
Related legal action is moving in several places. Our AI lawsuit tracker counts hundreds of cases, and a recent example is USA Today suing OpenAI for $250 million over training data.
What is confirmed and what is not
| Item | Status |
|---|---|
| A coalition filed for a provisional injunction in Oslo | Reported by Euronews |
| Meta delayed the start to November 20 | Reported by Euronews and The Verge; Meta statement quoted |
| The terms let Meta use paywalled journalism for AI | Plaintiffs' claim; Meta disputes that it gets new rights |
| Court has ruled | No ruling reported |
| The exact current wording of the terms | Check Meta's page directly; text has been edited since October 8 |
We could not find a second outlet repeating the Euronews account at the time of writing, so details such as the list of plaintiffs rest on that single report. We also could not read the court filing itself.
What a provisional injunction means
A provisional injunction is an interim measure. The plaintiffs ask the court to stop something from happening, here the terms taking effect, until the full case has been heard. Courts generally weigh whether the claimant has a credible case and whether waiting would cause harm that cannot easily be repaired later. We are describing the general idea, not predicting how an Oslo judge will apply Norwegian procedure; we have not seen the filing.
The timing is notable. With Meta's start date moved to November 20, a ruling on interim relief has a natural deadline. If the court declines, the terms would start on schedule unless Meta changes them again. If it grants the request, the terms would be frozen for Norwegian parties while the main case continues, though how far that would reach beyond Norway is unclear.
Three scenarios to watch
| Scenario | What it would look like | Effect |
|---|---|---|
| Meta clarifies | Revised text that states AI training is excluded, or an opt-out that does not require deleting the account | Likely reduces the case's urgency |
| Court grants interim relief | Terms blocked for the plaintiffs pending a full hearing | Sets a European reference point for contract-based AI data access |
| Terms start November 20 unchanged | Publishers rely on robots.txt and Meta's assurances | Pressure shifts to regulators and further suits |
The questions a judge, regulator or reader will ask
- What does "associated with your products or services" cover? The breadth of this phrase was the focus of the legal experts quoted in our earlier coverage. A newsroom with a Meta business account for advertising might find its whole site within scope.
- Does robots.txt control the outcome? Meta says it does. Critics say contract language could be argued to override it. A court is a place where that argument could finally be tested.
- Is deleting the account a real choice? For a publisher that depends on Meta for distribution and ad measurement, losing the account may not be practical. That is the core of the "all-or-nothing" complaint.
- Who bears the burden when AI products make errors? Our companion post on Meta's terms for AI features covers a related clause: customers being solely responsible for mistakes. The same terms package raises liability questions, not only content rights.
Why journalism is the test case
News organisations are an unusually clear test of consent for AI data use. They produce expensive, time-sensitive content, often behind paywalls, and AI answer products can substitute for visiting their sites. If a platform can acquire usage rights through account terms, publishers lose a negotiating position they have been using in licensing talks with AI companies. That is why a coalition that includes a public broadcaster, large commercial groups and a news agency chose to sue together.
What to do if you run a site or business account
- Read the live terms. The text has been edited more than once. Save a dated copy.
- List associated domains. Check which websites are connected to your Meta business accounts or ad accounts.
- Set robots.txt deliberately. Meta says it honours robots.txt preferences; make yours explicit and keep logs.
- Review licensing options. If your content has commercial value, document which uses you do and do not permit.
- Watch November 20. If the injunction is granted or Meta changes the text, the picture changes.
For an angle on how AI crawler traffic strains smaller sites, see our look at AI bot traffic and robots.txt rate limits.
Bottom line
This is a notable escalation: publishers moving from public criticism to a court filing against a contract clause. The outcome depends on how an Oslo judge reads terms that Meta says change little and that publishers say change everything. Until a ruling or a clarified text appears, treat both sides' readings as claims.
Related reading
- Meta Commercial Terms and the October 30 crawl: claimed vs verified
- Meta Commercial Terms: AI features and customer responsibility
- AI lawsuit tracker: 248 cases
- USA Today sues OpenAI for $250 million
- Rightsholders sue Anthropic over 274 songs
- AI bot traffic flood and robots.txt
Details are accurate as of October 27, 2026 and may change as the court acts.