Meta indefinitely pauses AI data-scraping terms after injunctions in Brazil and Norway

Meta has indefinitely shelved revisions to three sets of business terms that were scheduled to take effect November 20. The changes would have allowed automated scraping of websites linked to Meta Business accounts, with the content used to train generative models. The existing rules remain in place.

The pause followed legal pressure. Courts in Brazil and Norway granted preliminary injunctions against the terms. News publishers also organized opposition, arguing that businesses were being forced to hand over training data in exchange for advertising access.

This is a notable retreat in the fight over training data. Meta needs data to feed its Muse models, and advertiser websites would have been a large, low-friction source. The pause keeps Meta out of further litigation for now, but leaves that data gap open.

The skeptical take is in the fine print. Legal observers say the revised language Meta kept still allows content to be used for 'any business product,' so the core threat may be unchanged. Watch whether Meta returns with narrower terms, and whether other regulators follow the Brazilian and Norwegian courts.