SAO PAULO (Reuters) – Social media giant Meta Platforms will inform Brazilians how it intends to use their personal data to train generative artificial intelligence (AI), it said on Tuesday, following a demand from the country’s data protection watchdog.

Meta’s users in Brazil will receive the warnings starting on Tuesday by email and notifications on Facebook and Instagram, and will be able to reject the use of their data by the firm to train generative AI, it said in a statement.

Brazil’s National Data Protection Authority (ANPD) in July suspended Meta’s new privacy policy over the use of personal data to train AI, but overturned this last Friday saying Meta had agreed to issue the disclosures.

In July, Meta also actively suspended the use of generative AI tools in Brazil, which included tools used to create popular AI-made stickers for messaging platform WhatsApp. Brazil is WhatsApp’s second-largest customer base.

At the time, Meta said it had decided to suspend the tools while it was in talks with ANPD to address the authority’s doubts over its use of generative AI.

Asked by Reuters on Tuesday if it would resume the tools after ANPD lifted the suspension, Meta only reaffirmed that the suspension had been issued while it was in talks with the ANPD.

(Reporting by Paula Arend Laier and Andre Romani in Sao Paulo; Editing by Stephen Coates)