Meta told to stop mining Brazilians’ data to train AI software

Photo: Frank333/Shutterstock

Brazil’s Data Protection Authority (ANPD) on Tuesday ordered Meta to suspend its policy of using Instagram and Facebook posts to train artificial intelligence software. The company will be fined BRL 50,000 (USD 8,800) for each day it fails to comply with the order.

Two weeks ago, the ANPD had pledged to act on Meta’s new data mining practice.

Back on May 22, Meta said on its Brazilian blog it “may use information that people have shared publicly about Meta products and services for some of our generative AI features,” including “public posts or photos and their captions.”

Users could opt out of this data mining, a complicated process that takes several clicks. However, the ANPD said in a statement that “although [Brazilian] users could object to the processing of personal data, there were excessive and unjustified obstacles to accessing information and exercising this right”.

Furthermore, the agency concluded that the information available on Meta’s platforms is, in general, shared by users for relationships with friends, close communities, or companies they are interested in.

“In a preliminary analysis, there would not necessarily be an expectation that all this information — including that shared many years ago — would be used to train AI systems, which were not even implemented when the information was [first] shared,” the ANPD said.

This is significant, the agency added, because Facebook alone has 102 million active users in Brazil — about half of the country’s total population.

Meta told The Brazilian Report in a statement that it is “disappointed” with the ANPD’s decision. “AI training is not unique to our services, and we are more transparent than many players in this industry who have used public content to train their models and products,” it claimed. The tech giant added it will “continue to work with the ANPD to address its concerns.”

The company did not respond on whether it will comply with the order.

In June, Meta announced a halt in its plan to use data from users in the European Union and the United Kingdom to train its AI systems. The decision came after pushback from the Irish Data Protection Commission (DPC).

European data privacy activists successfully argued that users should be asked for explicit consent before their personal data is processed, rather than merely being given the option to refuse.