Meta halts AI-linked commercial terms after injunctions in Brazil and Norway

Meta has suspended planned changes to its commercial terms that would have obliged business account holders, including news publishers, to permit automated programmes to access, copy, store and analyse content on external sites linked from Facebook, and to allow that material to be used in training artificial intelligence systems. A company spokesperson said it was pausing the update "as we continue to engage with stakeholders and take our time to get this right", adding that the existing terms remain in effect. The retreat matters to AI policy because legal experts cited in reporting on the terms warned that the vague language could let Meta disregard scraping preferences set out in robots.txt files and exhaustively crawl off-Facebook content for AI training, potentially avoiding licensing payments to publishers.

The changes were initially due to take effect on 30 October and had already been pushed back once, to 20 November, before Friday's indefinite suspension. Under the proposed rules, continued use of the platforms would have counted as acceptance of the new terms, and publishers would have been required to create test credentials so that Meta's automated systems could reliably bypass paywalls. Meta has not said when, or in what form, revised terms might return.

In Brazil, the 44th Civil Court of the Central Forum of the Capital in São Paulo barred Meta on Tuesday from using robots to collect reports, photographs, videos and other material from outlets affiliated with three industry bodies — the National Association of Newspapers (ANJ), the Brazilian Association of Radio and Television Broadcasters (Abert) and the National Association of Magazine Publishers (Aner) — which had brought a joint action. Judge Guilherme Madeira Dezem also prohibited the company from using the material to develop or train AI systems such as Meta AI, and set a fine of 1 million reais per day, capped at 100 million reais. He characterised the requirement as contractual abuse in an adhesion contract, writing that imposing on news organisations the free and universal surrender of their copyright assets as a toll for remaining connected to the channels where contemporary audiences consume news amounted to manifest abuse of right and unjust enrichment. Meta said it would not comment.

In Europe, the pressure has been regulatory as well as judicial. Courts in Norway issued preliminary injunctions blocking the terms after organisations there filed lawsuits, and the UK Competition and Markets Authority opened a formal investigation into whether Meta was exploiting a dominant position to seize valuable AI training data. Theo Bamber, chief executive of the UK's News Media Association, described what he called "the continued arrogance of this company to think everyone will just be OK with its actions" as "staggering". Before the first postponement, the Movement for an Open Web, whose members include digital advertisers and news businesses, said it had lodged a complaint with the European Commission and the UK competition authority and sought interim measures.

The dossier now rests on two tracks: a suspended corporate decision, with the current terms left in force and no new implementation date announced, and litigation and regulatory scrutiny that remain live, including the São Paulo injunction and the London investigation.