Brazil’s National Data Protection Authority (ANPD) has temporarily banned Meta processing of personal data of users to train the company’s artificial intelligence (AI) algorithms.
ANPD said it found “evidence of personal data processing on the basis of an inadequate legal hypothesis, lack of transparency, limitation of the rights of data subjects and risks for children and adolescents”.
The decision was made by the social media giant update its terms which allow the use of publicly available content from Facebook, Messenger and Instagram for AI training purposes.
A recent report published This was revealed by Human Rights Watch LAION-5Bone of the largest sets of images and text used to train artificial intelligence models contained links to identifiable photos of Brazilian children, putting them at risk of malicious deepfakes that could put them at even greater risk of exploitation and harm.
Brazil has about 102 million active users, making it one of the largest markets. The ANPD noted that Meta’s update violates the General Data Protection Regulation (LGBD) and has an “unavoidable risk of causing serious and irreparable or hard-to-repair damage to the fundamental rights of affected data subjects.”
Meta has five business days to comply with the order or face daily fines of 50,000 reais (about $8,808).
U statement As shared with the Associated Press, the company said the policy “conforms to Brazilian privacy laws and regulations” and that the ruling is “a step back for innovation, competition in the development of artificial intelligence and further delays that allow people in Brazil to enjoy artificial intelligence. .”
The firm in social networks received similar repulsion in the European Union (EU), forcing it to suspend plans to train its artificial intelligence models using user data in the region without obtaining explicit user consent.
Last week, Meta’s president of global affairs, Nick Clegg said that the EU is “no longer a breeding ground for innovation and world-class companies”, adding that “the era of generative artificial intelligence offers the opportunity to change history”.
The development comes after Cloudflare released a new one-click tool that prevents AI bots from sampling content on customers’ websites to train large language models (LLMs).
“This feature will be automatically updated over time as we see new fingerprints of offending bots that we identify are making extensive use of the network to train models,” the web infrastructure company said. said.
