July 4, 2024 at 03:11AM
Brazil’s data protection authority, ANPD, has temporarily banned Meta from using users’ personal data to train AI algorithms due to inadequate legal basis, lack of transparency, and risks to children. Meta must comply within five days or face fines. This mirrors pushback in the EU over AI training data. The company insists it complies with privacy laws.
Key Takeaways from the Meeting Notes on Artificial Intelligence and Data Privacy:
1. Brazil’s data protection authority, ANPD, has temporarily banned Meta from processing users’ personal data for training AI algorithms due to inadequate legal basis, lack of transparency, and risks to children and adolescents.
2. The ban was a result of Meta’s update to its terms allowing it to use public content from Facebook, Messenger, and Instagram for AI training, which violated Brazil’s General Personal Data Protection Law.
3. Human Rights Watch found that an AI training dataset, LAION-5B, contained identifiable photos of Brazilian children, raising concerns about potential exploitation through deepfakes.
4. Meta has five working days to comply with the ban or face substantial daily fines.
5. Meta has received pushback in the E.U. for similar AI training plans, leading to a pause in its efforts due to a lack of explicit user consent.
6. Meta expressed its disagreement with the ban, stating that it complies with privacy laws in Brazil and that the ruling hinders innovation and AI benefits for people in Brazil and the E.U.
Feel free to reach out for further information or action items following these meeting notes.