Osegi
AI·1 min read

Brazil Court Bars Meta From Using Media Content for AI Training

A Brazilian court has ruled that Meta's new terms, which required free use of media content, are unlawful. The company faces fines if it does not comply.

Gavel resting in front of a Meta logo and a Brazilian flag, symbolizing a court ruling on AI training data.
Image: heise online

A court in Brazil has prohibited Meta from using media content to train its AI systems under the company's new terms of use. The ruling declares those terms impermissible and threatens the company with coercive fines if it fails to comply.

At the heart of the dispute is a clause in Meta's updated terms. They provided that content would be handed over to the company free of charge. The court found this arrangement unlawful, so Meta cannot rely on it as a basis for feeding media material into its AI training.

Why it matters

The decision adds to the growing legal pressure on large technology companies over the data used to build AI models. Publishers and other media producers have increasingly questioned whether their work can be used for training without consent or compensation. The Brazilian ruling takes a clear position against a blanket, unpaid transfer of rights through terms of service.

For Meta, the threat of a coercive fine gives the order practical weight: the company would have to adjust its terms or its training practices in the Brazilian market, or risk recurring penalties.

What comes next

The material available does not specify the amount of the possible fine, the deadline for compliance or whether Meta plans to appeal. Those details will determine how far-reaching the consequences are, both for Meta's operations in Brazil and for how other companies draft similar terms elsewhere.

Apps worth a look

SUMO: the number puzzle

Tap the right tiles, add them up, hit the target. Easy to learn, hard to master. Free, for all ages.

Tags:metabrazilai trainingcopyrightmedia content

More from this day

See all →

More in: AI

See all →