Brazil's national data protection agency (ANPD) has suspended Meta's latest privacy policy, which would have allowed the company to train generative AI models using posts from its users on Facebook and Instagram. The decision follows concerns about the potential risks to users' fundamental rights and the difficulty in repairing any damage. Meta has been given five working days to amend its privacy policy to exclude the use of personal information from public posts for AI training, or face a daily fine of R$50,000. The company's updated policy has also faced scrutiny in the UK and the EU, where it would have allowed the use of users' information to develop and improve its AI products. Brazil's ANPD's decision comes after a report by Human Rights Watch revealed that a dataset used to train AI models contained identifiable photos of Brazilian children, raising concerns about potential exploitation. Meta expressed disappointment with the decision, stating that it complies with privacy laws and regulations in Brazil and that the ruling is a setback for innovation and competition in AI development. The company also highlighted that AI training is not unique to its services and that it is more transparent than many industry counterparts in using public content for training. The decision by Brazil's ANPD reflects the country's significant market for Meta, with over 102 million Facebook users and more than 113 million Instagram users. The authority's move underscores the growing global scrutiny of tech companies' use of personal data for AI training and the need to balance innovation with privacy and fundamental rights.
