Policy & Regulation4 November 2026, 02:02 UTC2 min read

Can Meta Train AI on Brazilian News?

A São Paulo court issued a provisional order on November 3, 2026, prohibiting Meta from using news content from three Brazilian media groups to train its AI systems. The ruling includes a potential fine of up to $18.2 million if violated. The order applies to journalistic content but does not explicitly cover other types of web data, leaving some ambiguity for Meta’s broader scraping practices.

1 source

Key points

São Paulo court barred Meta from training AI on news content from three Brazilian media groups on November 3, 2026
Fine of up to **$18.2 million** applies if Meta violates the provisional order, per the ruling
Order does not clarify whether Meta’s broader web scraping for non-news content is affected

The decision follows a legal battle over Meta’s new commercial terms, which publishers argue allow unchecked access to copyrighted material. While Norway’s October 2026 ruling against Meta’s terms paused similar changes there, Brazil’s ruling is the first to directly target AI training. Meta has not yet commented on the order’s implications or whether it will appeal. The case highlights growing legal pressure on tech giants over data usage in AI development.

Full story from riotimesonline.com · via Search: MetaOpen source ↗