The ANPD banned Meta from using Brazilian data to train its AI models due to the risks of serious damage and difficulty to users1. This decision came after an update to Meta's privacy policy allowed the company to use public data from Brazil for AI training, raising concerns about potential harm to fundamental rights of Brazilian users.
The ANPD cited the risks of serious damage and difficulty to users, as well as imminent risk of serious and irreparable or difficult-to-repair damage to the fundamental rights of Brazilian users. The decision was based on concerns about the use of personal data, including photos of Brazilian children, for training AI models, which could lead to deepfakes and other forms of exploitation.
Meta's updated privacy policy allowed the company to use public Facebook, Messenger, and Instagram data from Brazil, including posts, images, and captions, for AI training purposes2. The policy aimed to provide more clarity around the types of information the company collects, uses, and shares, in line with evolving privacy regulations.