Meta blocked from using social media posts for AI training in Brazil 

Meta’s new privacy policy to use public social media posts for training their artificial intelligence (AI) has been blocked once again by Brazil’s data protection regulator. 

Brazil’s national data protection authority announced this week that the policy poses “an imminent risk of serious and irreparable or difficult-to-repair damage to the fundamental rights of the affected data subjects,” as stated in the nation’s official gazette.

Brazil is an important market for Meta, with approximately 102 million active Facebook users, according to the agency’s statement. Meta has faced similar resistance in Europe, where it recently paused its plans to use public posts for training AI systems, a rollout that was initially scheduled for last week.

In Europe, the policy changes would have involved using publicly shared posts, images, image captions, comments, and stories from users over 18 on Facebook and Instagram, excluding private messages. 

The Irish Data Protection Commission (DPC) intervened, requesting Meta to delay training large language models (LLMs) that power chatbots like OpenAI’s ChatGPT with this data, in response to concerns from other European bodies.

Meta argued that without access to local data, it could only offer MetaAI users “a second-rate experience” with AI products that would struggle to accurately understand regional languages, cultures, or trending topics. Consequently, Meta delayed its European rollout. If granted access to this data, the company intended to launch Llama, its AI chatbot, and the MetaAI assistant.

Meta maintained that its European approach “complies with laws and regulations” and is “more transparent than many of [their] industry counterparts.” A Meta spokesperson expressed disappointment to the Associated Press, asserting that the company’s methods comply with Brazilian privacy laws and regulations.

“This decision is a setback for innovation, competition in AI development, and further delays bringing the benefits of AI to people in Brazil,” the spokesperson added.

Meta must demonstrate compliance within five working days of the notification, otherwise, the company could face a daily fine of 50,000 reais. 

Be the first to comment

Leave a Reply

Your email address will not be published.


*


This site uses Akismet to reduce spam. Learn how your comment data is processed.