Brazil’s data protection authority, Autoridade Nacional de Proteção de Dados (ANPD), has issued a temporary ban against Meta from processing users’ personal data for training its artificial intelligence (AI) algorithms.
The ANPD found evidence of improper processing of personal data by Meta, which included inadequate legal basis, lack of transparency, infringement on data subjects’ rights, and risks to children and adolescents.
This decision follows Meta’s recent update to its terms, allowing it to use public content from Facebook, Messenger, and Instagram for AI training.
A report by Human Rights Watch revealed that LAION-5B, a major image-text dataset used for AI training, contained identifiable photos of Brazilian children, exposing them to potential harm from malicious deepfakes.
The ANPD noted that Meta’s update violates the General Personal Data Protection Law (LGBD) and poses a significant risk to data subjects’ fundamental rights.
Meta has five days to comply with the order or face daily fines of 50,000 reais.
In a statement, Meta defended its policy as compliant with privacy laws in Brazil and criticized the ruling as hindering innovation and AI development in the country.
Similar pushback in the European Union has caused Meta to pause plans to use EU user data for AI training without explicit consent.
Last week, Meta’s president of global affairs expressed concerns about innovation in the EU and highlighted the opportunities presented by generative AI.
Cloudflare has launched a new tool to prevent AI bots from scraping websites for training large language models, aiming to protect its customers’ content from unauthorized use.
The tool will be updated to combat new bot fingerprints that attempt to scrape the web for model training.
