Brazil blocks Meta from using social media posts to train AI
Brazil has blocked Meta from using Brazilians' Instagram and Facebook posts to train its artificial intelligence (AI) models. This decision comes after the company had to abandon similar plans involving UK and European users' posts for the same purpose.
On Tuesday, Brazil's national data protection agency (ANPD) announced the immediate suspension of Meta's latest privacy policy. This policy allowed the company to train generative AI models, such as chatbots, based on posts from its users.
A Meta spokesperson expressed disappointment, stating that the company's approach was in compliance with local privacy laws. They added that this move is seen as a setback for innovation, competition in AI development, and the progress of bringing AI benefits to the people in Brazil.
Significant Market Presence
Meta holds a significant market presence in Brazil with 102 million Facebook users and over 113 million Instagram users in the country.
The ANPD justified its actions citing the imminent risk of serious and irreparable damage to the fundamental rights of the affected account holders. Meta has been given a deadline of five working days to revise its privacy policy excluding the use of personal information from public posts to train generative AI. Failure to comply will result in a daily fine of R$50,000 (£6,935).
Policy Scrutiny in the UK and EU
The company's updated policy faced scrutiny in the UK and the European Union as well. The planned privacy policy changes, slated to be effective from 26 June, aimed at using Meta users' information for the development and enhancement of its AI products.
However, following a request from the Irish Data Protection Commission (DPC) and other European stakeholders, Meta decided to delay the training of large language models (LLMs). These models play a crucial role in powering chatbots like OpenAI's ChatGPT and Google's Gemini. The delay was deemed a setback for AI development in Europe.
Despite the postponement in Europe, Meta proceeded with the policy changes in Brazil. Pedro Martins from Data Privacy Brasil praised the ANPD's decision and pointed out the disparities in Meta's data protection measures for Brazilian and European users.
Concerns and Criticisms
Martins highlighted that Meta had intended to utilize posts from Brazilian minors to train its AI models, while in Europe, such practices were limited to users above 18 years old. The data protection regulator in Brazil raised concerns regarding the potential collection and usage of personal data from minors, which could violate the country's data protection laws.
Moreover, Martins emphasized that the process for users to prevent Meta from using their personal information was more complex in Brazil compared to Europe. Users in Brazil reportedly have to navigate up to eight steps to block the company from utilizing their posts.
The BBC has reached out to Meta for comments on the allegations concerning the use of posts from Brazilian minors to train AI models and the differences in opt-out procedures between Brazilian and European users.
