Meta decides not to launch its AI in Brazil due to ‘regulatory uncertainties’ and is criticized by consumer protection entity

A Meta announced yesterday (23) that its generative AI, integrated into Facebook, Messenger, Instagram and WhatsApp, in addition to having a web version, has arrived in Latin American countries. But one detail did not go unnoticed: the absence of Brazil among the countries that received the resource.

Powered by the Llama 3.1 405B language modelgenerative AI has image editing capabilities, stopping and removing people and objects, and can even create texts.

But why did Meta decide to leave Brazil out of this launch? The answer has to do with the conflict between the company and ANPD (National Data Protection Authority). The entity decided that Meta cannot use data from Brazilians to train its artificial intelligence models, and established a daily fine of R$50,000 in case of non-compliance.

Meta talks about ‘regulatory uncertainties’

Meta even stated that the decision was a setback. And, citing the issue again, the company now justifies its decision by saying that Brazil offers “regulatory uncertainties”.

“Due to local regulatory uncertainties, Meta’s AI will not yet be available in Brazil. We will continue to work closely with the relevant local authorities to ensure that Brazil has access to – and is properly served by – the same level of AI innovation that we are bringing to other countries, including in Latin America,” Meta said in a statement.

Argument rejected by Idec

Meta’s position was harshly criticized by the Brazilian Institute for Consumer Protection (Idec)classifying as a “fallacy” what the company claimed as justification:

“It is a fallacy that Meta could not implement this AI in Brazil. However, it would have to do so while respecting the laws – and the fact that they are not bringing this technology to the country may show that they are assuming that they do not want to comply with them,” said Idec in a note.

The Institute also added that “Meta’s statement can be interpreted as a kind of ‘collective punishment’ for users who depend so much on these platforms: if someone claims their rights, all users will be left without access to Meta’s new products.”

ANPD cites irregularity in relation to the General Data Protection Law (LGPD)

The ANPD’s ban on Meta using Brazilian data to train its AI is based on the General Data Protection Law (LGPD). The entity argues that Meta’s new privacy policy is irregular from the LGPD’s point of view and that this training could represent “risks of serious and difficult-to-repair damage to users”.

The ANPD listed the following points in the decision:

• Use of hypotheses inadequate legal framework for the processing of personal data;
• Failure to provide clear, accurate and easily accessible information about the privacy policy change and about the treatment carried out;
• Excessive limitations on the exercise of the rights of the holders;
• Processing of personal data of children e adolescents without proper safeguards.