Courts Block Meta's Plan to Mine Publisher Content
Meta's proposed terms would let it crawl business websites for AI development, a plan still pending a 20 November start date.
Updated on Nov. 5, 2026 in Artificial Intelligence
Live Poll
Should tech companies have to pay news publishers when they use their articles?
Meta has proposed new Commercial Terms, announced 8 October 2026, that would let it crawl and collect content from the websites of any business using Facebook, Instagram, WhatsApp or Messenger for business purposes. The development is a proposed policy, not a shipped product: Meta has delayed the terms until 20 November and removed an explicit reference to AI training, though its revised wording still permits use of content to develop business products, including generative AI services.
Why it matters
The dispute centres on whether publishers' journalism can be used to train generative AI systems without payment or authorisation, a question now being tested in courts and competition inquiries across several countries. Courts in Norway and Brazil have already moved to block the terms, while publishers in South Africa and elsewhere have asked competition authorities to act.
The reported evidence consists of court interim orders and the wording of Meta's proposed terms as described by publishers, rather than technical benchmarks. The terms reportedly permit Meta to use content for developing business products, including generative AI services, while the commitment to respect robots.txt, a web standard that lets sites signal crawling limits, is not written into the revised text.
The players
Meta
Meta operates Facebook, Instagram, WhatsApp and Messenger, and is building generative AI services that rely on large-scale data collection.
SANEF
The South African National Editors' Forum is a national body of South African newsroom editors that opposed Meta's proposed terms on 5 November 2026.
The details
Crawling refers to automated software that visits web pages and copies their content for indexing or analysis. Under the proposed terms, that collection would extend to websites of any business using Meta's messaging and social platforms for business purposes. Generative AI services are systems that produce new text or images from patterns learned in training data, which is why the use of journalistic content for training is contested.
Timeline
8 October 2026: Meta announced the proposed Commercial Terms.
28 October 2026: The Oslo District Court issued an interim order preventing Meta from implementing the terms.
5 November 2026: SANEF issued a statement opposing the terms.
20 November 2026: The delayed terms are scheduled to take effect.
The Tech Race
Publishers have asked competition authorities in the United Kingdom, the European Union, Denmark and Sweden to act against the proposed terms. The dispute places Meta's AI data-collection plans against a growing set of legal challenges from news organisations seeking payment or consent for use of their journalism.
If the terms take effect on 20 November, content from business websites using Meta's platforms could be collected for AI development. The Oslo and São Paulo court orders currently block implementation in Norway and restrict collection of Brazilian journalism in Brazil.
The takeaway
Watch whether Meta's terms take effect on 20 November 2026 and whether the Oslo and São Paulo court orders are upheld on the merits. Those outcomes will indicate how far publishers' claims over AI training data can restrain platform collection.
What happens next
Meta's delayed Commercial Terms are scheduled to take effect on 20 November 2026, a date on which the dispute over the terms will reach a decision point.
Further reading
For broader context on how these data-use disputes fit into the wider field, see Artificial Intelligence.
Source note: This article includes information reported by SANEF | Protecting Media Freedom.
Live Poll
Should tech companies have to pay news publishers when they use their articles?
