Court Bars Meta From Using News Articles to Train AI in Brazil - 04/11/2026 - Business - Folha de S.Paulo

A São Paulo court on Tuesday (3) prohibited Meta, the owner of Facebook, Instagram, WhatsApp and Threads, from using bots to collect news articles, photographs, videos and other content from media outlets affiliated with three Brazilian press associations.

The ruling also prevents the use of that material to develop or train artificial intelligence systems, such as Meta AI. The lawsuit sought to suspend the company's new commercial terms, which are set to take effect on November 20.

According to the associations, the rules would allow Meta to access, copy, store and analyze content from news websites. The judge suspended the provision requiring companies to authorize, without compensation, the use of "web scrapers" and ruled that the requirement constituted contractual abuse.

He said scraping news websites to feed language models constitutes reproduction of copyrighted works, which requires prior authorization, and that Brazilian law provides no exception for data mining for commercial purposes. According to the judge, the terms create an "insoluble legal trap" because media outlets publish third-party material.

The fine for noncompliance is R$1 million ($182,400) per day, capped at R$100 million ($ 18 million). Meta has five days to challenge the ruling, and the associations have 30 business days to file their main claim. The company was not heard and said it would not comment. Marcelo Rech, executive president of the Brazilian Newspaper Association, called the injunction "historic" and said it shows that there are limits to the abuse of economic power by platforms.

The ruling comes nearly a week after a Norwegian court suspended the same rules.