Meta has paused generative AI features in Brazil after the country’s National Data Protection Authority (ANPD) barred the company from training AI models on personal data from Brazilians. The decision interrupts Meta’s effort to expand its AI products in a market with more than 200 million people.
What Brazil’s data authority decided
The ANPD banned Meta from using personal data from Brazilians to train its AI models. In response, Meta suspended the use of its AI assistant in Brazil.
The Brazilian authority described the action as a preventive measure. As cited in the official document, it said the decision was tied to “the imminent risk of serious harm and irreparable or difficult-to-repair damage to the fundamental rights of guardians.”
The order also carries a financial consequence. The ANPD set a daily fine of 50,000 reais if Meta does not comply.
For Meta, the immediate result is that genAI features that had already been available in Brazil are no longer active while the company engages with the regulator.
Meta says it is engaging with ANPD
A Meta spokesperson confirmed the suspension in a statement to TechCrunch. The company said, “We decided to suspend genAI features that were previously live in Brazil while we engage with the ANPD to address their questions around genAI.”
That statement frames the suspension as a step taken during regulatory discussions rather than a full withdrawal from the Brazilian market. Still, the pause affects a central part of Meta’s AI product rollout: the ability to keep live generative AI features available while using local personal data for model training.
The source article does not describe which individual genAI features were affected beyond Meta’s AI assistant and the broader genAI tools that had been live in Brazil. It also does not provide a timeline for when, or whether, those features could return.
Why the training issue matters
The core issue is not simply whether Meta can offer an AI assistant in Brazil. The ANPD action focuses on whether Meta can train AI models on personal data from Brazilians.
Training is a separate question from user access. A company may want to offer AI tools in a market, but regulators can still challenge how the data behind those tools is collected, used, or incorporated into model development.
In this case, the ANPD’s concern centers on personal data and fundamental rights. The source article identifies the measure as preventive, meaning the authority acted before the disputed training practice could continue under the challenged terms.
The decision also shows how AI product launches can be constrained by data protection rules. For companies building generative AI products across multiple markets, the same product strategy may face different regulatory responses depending on the jurisdiction.
A wider pattern of regulatory pushback
Brazil is not the only market where Meta’s AI training plans have faced resistance. Meta has already been training its AI using user-generated content in the U.S. and other markets for several years.
But in May, Meta had to pause plans to train its AI models in Europe and the U.K. after the Irish Data Protection Commission pushed back against those plans.
The Brazil suspension therefore sits alongside a broader challenge for Meta: expanding generative AI products while answering questions from data protection authorities about the use of personal data and user-generated content.
For users in Brazil, the visible impact is the suspension of Meta’s AI assistant and previously live genAI features. For Meta, the strategic impact is larger. Brazil is described as a market with more than 200 million people, and a pause there affects the company’s ability to build out AI products in a major market.
What remains unresolved
The source article leaves several questions open. It does not say how long Meta’s suspension will last, what changes the ANPD may require, or whether Meta will alter its approach to AI training in Brazil.
What is clear is that the company is now in a regulatory process with ANPD. Until those questions are addressed, Meta’s genAI features in Brazil remain suspended under the conditions described by the company and the authority.
The case highlights a practical reality for generative AI deployment: product availability, model training, and personal data governance are now tightly connected. In Brazil, that connection has already forced Meta to stop features that were live while it responds to the regulator’s concerns.