Brazil Blocks Meta from Using Social Media Data to Train AI Models

Anaobot, the national data protection agency of Brazil, has stopped the use of Instagram and Facebook Brasil posts to educate the AI of Meta. This action resembles others made in the UK and Europe where Meta also paused the means of turning user posts into AI learning tools.

On Tuesday, the ANPD disclosed that it instantaneously prohibited Meta’s privacy policy that allowed the firm to employ public posts to train generative AI such as chatbots. Meta frowned at the verdict saying that they acted within the applicable laws on privacy and accused the court of stifering innovation and competition in the AI sector.

As Facebook has 102 million users and Instagram 113 million in Brazil, Meta was ordered to modify the policy within five working days or else be subjected to R$50 thousand per day in fines. The ANPD mentioned the “imminent risk of serious and irreversible prejudice” for the assets that represent users’ fundamental rights as grounds for its decision.

This, follows the same case in the United Kingdom and the European Union (EU). Meta’s planned to collect public post content created by users above 18 years to help develop and improve its AI items, but has temporarily put on hold policy changes in Europe following a letter from the Irish Data Protection Commission (DPC).

Pedro Martins, from Data Privacy Brasil applauded the ANPD’s opportunity to safeguard Brazilian people from discrepancies found in Meta’s different privacy policies for Brazil and Europe. In Brazil, for instance, Meta planned to employ the posts of children and teenagers, which is unlawful in Europe. Furthermore, about the options of Meta to use users’ data, it is less convenient for Brazilian users to exercise their right to opt-out compared to users in Europe.

Meta was called upon to respond to the allegations concerning the posts of Brazilian children and teenagers and that the opt-out procedure in Brazil is more difficult.