Brazil’s data protection authority, Autoridade Nacional de Proteção de Dados (ANPD), has temporarily banned Meta from processing users’ personal data for training the company’s artificial intelligence (AI) algorithms.
The ANPD stated that it found “evidence of processing of personal data based on inadequate legal grounds, lack of transparency, limitation of the rights of data subjects, and risks to children and adolescents“.
This decision comes after the social media giant updated its terms to permit the use of public content from Facebook, Messenger, and Instagram for AI training.
A recent report by Human Rights Watch revealed that LAION-5B, one of the largest image-text datasets for AI model training, contained links to identifiable photos of Brazilian children, exposing them to potential malicious deepfakes and further exploitation.
With approximately 102 million active users, Brazil is one of Meta’s largest markets. The ANPD noted that Meta’s update violates the General Personal Data Protection Law (LGPD) and poses “imminent risk of serious and irreparable or difficult-to-repair damage to the fundamental rights of the affected data subjects”.
Meta has five working days to comply with the order or face daily fines of 50,000 reais (around $8,808).
In a statement to the Associated Press, the company claimed its policy “complies with privacy laws and regulations in Brazil,” and described the ruling as “a step backwards for innovation, competition in AI development, and further delays in bringing the benefits of AI to people in Brazil”.
Meta has faced similar opposition in the European Union (E.U.), forcing it to halt plans to train its AI models using user data from the region without explicit consent.
Last week, Meta’s president of global affairs, Nick Clegg, remarked that the E.U. is “no longer a fertile ground for innovation and world-class companies,” suggesting that the “era of generative AI presents an opportunity to change the narrative”.
Meanwhile, Cloudflare has introduced a new “one-click” tool to prevent AI bots from scraping content from its customers’ websites for training large language models (LLMs).
“This feature will be updated automatically as we identify new fingerprints of bots widely scraping the web for model training,” stated the web infrastructure company”.
