Brazil's data protection watchdog has slammed the brakes on Meta's plans to use Brazilian user data for AI training. The National Data Protection Authority (ANPD) ordered the social media giant to immediately stop harvesting data from its platforms like Facebook and Instagram for AI development, citing risks to users' rights.
At the heart of the issue is Meta's recent privacy policy update, which permits the use of public posts, photos, and captions from its platforms – including Facebook, Instagram, and Messenger – to train and refine generative AI systems. This has raised alarm bells among privacy advocates and regulators, particularly given the vast trove of personal data at Meta's disposal. (Btw, here's how to opt out of your data being used to train Meta AI models) Now, Meta faces daily fines of 50,000 reais ($8,820) if it doesn't comply within five days.
Why did Brazil take action? The ANPD highlights several key concerns:
- Inadequate legal basis: The authority questions Meta's reliance on "legitimate interest" as justification for data processing, especially when dealing with sensitive personal information.
- Lack of transparency: Regulators argue that Meta failed to provide clear, precise, and easily accessible information about the privacy policy changes and their implications.
- Excessive limitations on user rights: The ANPD found that Meta placed undue obstacles in the way of users seeking to exercise their data protection rights.
- Insufficient safeguards for minors: The authority expressed particular concern over the potential use of children's and adolescents' data without adequate protections.
This hits Meta hard. Brazil is a massive market, with Facebook alone boasting 102 million active users in the country of 203 million people.
Meta pushed back, claiming its approach "complies with privacy laws and regulations in Brazil." A spokesperson called the decision "a step backwards for innovation" that will delay AI benefits for Brazilians.
This isn't Meta's only AI data headache. The company recently paused similar plans in Europe after facing regulatory heat there. Meanwhile, in the US, where online privacy laws are weaker, Meta's data collection for AI continues.
The Brazilian decision follows a Human Rights Watch report exposing how a popular AI training dataset (LAION-5B) contains identifiable photos of Brazilian children, raising serious ethical concerns.
As tech giants race to develop AI, they're running into a wall of privacy concerns. Brazil's move highlights the growing tension between rapid AI advancement and protecting people's personal information.
For Meta, threading this needle is crucial. They need to push AI forward while navigating a patchwork of international privacy rules. For users, Brazil's stance is a win in the fight to control personal data in the AI era.