Brazil has ordered Meta to halt the use of data from its Instagram and Facebook platforms for training AI models, citing privacy risks and the potential exploitation of user data.
Short Summary:
- Brazil blocks Meta’s use of Brazilian user data for AI training.
- Meta given five days to comply with the ruling, or face daily fines.
- Decision mirrors similar actions in Europe against Meta’s AI practices.
Full Story:
In a significant development, Brazil’s national data protection authority (ANPD) has mandated an immediate halt in Meta’s use of Brazilian user data from Instagram and Facebook for training its AI models. This decision follows the agency’s assessment that such practices pose “imminent risk of serious and irreparable damage, or difficulty repairing fundamental rights” of the affected individuals.
Meta, the parent company of Instagram and Facebook, updated its privacy policy to allow public user posts to be utilized for AI training, including generative models such as their chatbots. However, the ANPD’s ruling now blocks this in Brazil. The ruling highlights ongoing global scrutiny and backlash against Meta’s data practices.
A spokesperson for Meta responded to the decision, stating, “We are disappointed by the decision, as we believe our approach complies with local privacy laws. This is a step backwards for innovation and competition in AI development, further delaying the benefits of AI for people in Brazil.”
“This is a step backwards for innovation, competition in AI development and further delays bringing the benefits of AI to people in Brazil,” Meta said.
Brazil represents a crucial market for Meta, with 102 million Facebook users and over 113 million Instagram users. The ANPD’s decision reflects considerable concern over potential harm and privacy violations that could arise from using personal data to train AI systems. Under the ruling, Meta has five working days to revise its privacy policy to exclude the use of public posts for AI training. Non-compliance could result in a daily fine of R$50,000 (approximately £6,935 or $8,820).
“The imminent risk of serious and irreparable or difficult-to-repair damage to the fundamental rights of the affected data subjects necessitated this action,” the ANPD stated.
Meta’s policy to use public posts for AI training not only faces resistance in Brazil but has also been scrutinized in Europe. The company had planned to start using public posts from users over 18 in the UK and EU for AI development under a policy change set to take effect on June 26. Posts, images, comments, and Stories shared publicly on Facebook and Instagram were to be included in this data collection, excluding private messages. However, due to a request from the Irish Data Protection Commission (DPC) on behalf of other European stakeholders, Meta delayed this plan.
Pedro Martins from Data Privacy Brasil flagged the differences in data protection measures between Brazil and Europe, stressing that Brazilian users, including minors, could be vulnerable due to less stringent regulations and complex opt-out mechanisms. In Europe, users under 18 are not subject to the AI training data collection, showcasing a more protective approach.
“In Europe, the steps users can take to prevent Meta from using personal information are more straightforward than in Brazil, where it can take as many as eight steps for users to block the company from using their posts,” Pedro Martins explained.
The ANPD’s ruling aligns with a global trend of increased scrutiny on how tech companies like Meta utilize user data. Similar concerns have been raised in the United States, where Human Rights Watch reported the use of personal photos of Brazilian children from large databases for AI image-generation tools without consent. This practice underscores the risks of exploiting personal data for AI development, particularly for vulnerable populations.
“Meta was severely punished for being the only one among the Big Tech companies to clearly and in advance notify in its privacy policy that it would use data from its platforms to train artificial intelligence,” said Ronaldo Lemos of the Institute of Technology and Society of Rio de Janeiro.
Despite Meta’s assertions that its policy complies with Brazilian privacy laws, the ANPD criticized the lack of transparency and accessibility regarding users’ options to opt out. They pointed out “excessive and unjustified obstacles” for users wanting to prevent their data from being harnessed for AI training, exacerbating potential privacy risks.
Hye Jung Han from Human Rights Watch highlighted the ruling’s protective impact on children, noting that it would help shield them from unforeseen harm resulting from the misuse of their personal data shared on social media platforms.
“This action helps to protect children from worrying that their personal data, shared with friends and family on Meta’s platforms, might be used to inflict harm back on them in ways that are impossible to anticipate or guard against,” Hye Jung Han stated.
Brazil’s action against Meta reflects a broader push among nations to safeguard user privacy and ensure that technological advancements such as AI development do not come at the expense of individuals’ fundamental rights. This decision signals a decisive stance that user data should not be exploited without proper consent and protective measures.
Meanwhile, Meta is likely to face ongoing challenges as data protection authorities around the world closely monitor and regulate how tech giants manage user data for AI training and other purposes. This case sets a precedent for other countries to potentially follow suit in curbing the use of personal data for AI without stringent safeguards.
The ANPD has given Meta a tight deadline to demonstrate compliance with the ruling, emphasizing that data protection and user rights must be prioritized in any technological innovation process. As the global conversation about ethical AI and data use continues to evolve, Brazil’s decisive action underscores the paramount importance of safeguarding user privacy in an increasingly digital world.