Meta Platforms has opted to suspend its generative AI tools in Brazil in response to objections from the nation’s government concerning new privacy policies. According to Reuters, this pause comes amid regulatory scrutiny from Brazil’s National Data Protection Authority (ANPD).
Importance of Brazil for Meta
Brazil, home to over 200 million people, is an important market for Meta. It is the second-largest user base for WhatsApp, following India. This makes Brazil a strategic priority for Meta, particularly for its AI initiatives.
In June, Meta unveiled its first AI-powered ad targeting program tailored for businesses using WhatsApp at an event in Sao Paulo. The project aimed to use AI to improve ad targeting on the popular messaging app, aligning with Meta’s broader AI integration strategy.
ANPD’s Regulatory Intervention
Earlier this month, the ANPD halted Meta’s new privacy policy that permitted the use of personal data for training AI systems. The regulatory body instructed Meta to revise its policy to exclude personal data processing for AI training.
In reaction, Meta decided to pause its AI tools and entered into discussions with the ANPD. The company expressed its commitment to resolving the issues and aligning with Brazil’s data protection laws.
The ANPD’s decision was driven by worries about the potential risk to fundamental rights of individuals. The authority imposed a daily fine of 50,000 reais (around $8,836) for non-compliance, indicating the seriousness of the matter.
Meta’s Response and Impact on Development
Meta described the ANPD’s ruling as a disappointment and a hindrance to innovation. Nevertheless, the company is focusing on addressing regulatory concerns while continuing to explore responsible AI integration.
The suspension of AI tools in Brazil underscores the increased regulatory scrutiny tech companies face over data privacy and AI use. As Meta works to meet these requirements, the outcome could affect how AI technologies are implemented in other regions with stringent data laws.
Meta has also chosen to pause the release of its future multimodal Llama 3 AI in Europe. The imminent release of multimodal AI models, capable of processing a confluence of data types including video, audio, images, and text, promises integration with various devices. However, a regulatory hurdle has emerged, as European businesses will be excluded from initial access to these advancements, despite the models being offered under an open-source license.
The focus of Meta’s challenge lies in ensuring compliance with the General Data Protection Regulation (GDPR) when utilizing data from European individuals for model training purposes. Earlier this year, Meta announced plans to leverage publicly accessible posts from Facebook and Instagram for this endeavor, notifying over two billion EU users of their right to opt-out.