Recently, Brazil’s national data protection authority ruled that Meta, the parent company of Instagram and Facebook, is not allowed to utilize data from Brazil to train its artificial intelligence.
Despite Meta’s updated privacy policy allowing the use of public posts for AI training, this practice will not be allowed in Brazil.
The agency stated in the official gazette that the decision was made due to the high likelihood of significant and irreparable harm to the fundamental rights of the individuals affected.
Brazil is a major market for Meta, with Facebook alone having approximately 102 million active users in the country, as reported by the agency. According to the country’s 2022 census, Brazil has a population of 203 million.
In a statement, a representative for Meta expressed disappointment and maintained that the company’s approach is in line with privacy laws and regulations in Brazil.
The spokesperson also stated that the decision is a setback for innovation, competition in AI development, and will further postpone the delivery of AI benefits to the people in Brazil.
Additionally, the social media company has faced opposition to its privacy policy revision in Europe, leading to the postponement of plans to use public posts for training AI systems, which was originally scheduled to commence last week.
In the United States, online privacy protection is not mandated by national law, but training in this area is already underway.
Meta stated in May on its Brazilian blog that it may utilize publicly shared information about its products and services for certain AI features, such as public posts, photos, and their captions.
Although Meta stated that opting out is an option, the agency expressed concerns about the difficulty and lack of justification in accessing information and exercising the right to opt out.
Meta did not provide enough information for people to understand the potential risks of using their personal data for developing generative AI.
This is not an isolated issue, as other companies have also used data from Brazilians to train their AI systems.
A recent report by Human Rights Watch revealed that personal photos of identifiable Brazilian children were taken from various online sources and used to create AI image-generator tools without the knowledge of their families. In some instances, these tools were used to generate n*de imagery.
On Tuesday, Hye Jung Han, a researcher based in Brazil for a rights group, stated in an email that the regulator’s action is a step towards protecting children from the potential harm of their personal data being used in unforeseeable ways on Meta’s platforms.
However, Ronaldo Lemos from the Institute of Technology and Society of Rio de Janeiro, expressed concern that this decision may lead other companies to be less transparent about their data usage in the future.
“Meta faced severe consequences for being the sole Big Tech company to explicitly disclose in its privacy policy its intention to utilize data from its platforms for artificial intelligence training,” he stated.
The company is required to show compliance within five business days of receiving the decision, and the agency imposed a daily fine of 50,000 reais ($8,820) for non-compliance.
Stories You May Like