Meta Begins Training AI Models on EU Public Content

On Monday, Meta announced a significant shift in its AI strategy within the European Union, revealing plans to train its AI models using public posts and comments from Facebook and Instagram. This decision comes after a temporary halt due to regulatory concerns about data privacy. Starting this week, Meta will begin this training process, incorporating not just public content but also users' interactions with Meta AI to enhance its models.

The move follows a limited rollout of Meta AI in the EU last month, which was significantly delayed compared to its launch in the U.S. and other regions. While Meta has been using user-generated content to train its AI in the U.S. for years, the EU's stringent privacy laws, notably the General Data Protection Regulation (GDPR), have posed challenges. GDPR mandates a clear legal basis for processing personal data, which has been a sticking point for Meta's AI training efforts.

Back in June 2024, Meta had to pause its plans to train AI systems with user data in the EU and U.K. after facing pushback from the Irish Data Protection Commission (DPC), which oversees Meta's operations in the EU. By September 2024, Meta resumed these efforts in the U.K., and now, it's extending this to the EU as well.

“Last year, we delayed training our large language models using public content while regulators clarified legal requirements,” Meta explained in a recent blog post. “We welcome the opinion provided by the EDPB in December, which affirmed that our original approach met our legal obligations. Since then, we have engaged constructively with the IDPC and look forward to continuing to bring the full benefits of generative AI to people in Europe.”

Starting this week, EU users will receive notifications via in-app messages and emails, informing them that Meta will use their public data and interactions with Meta AI for training purposes. These notifications will include a link to a form allowing users to opt out of this data usage. Meta has committed to respecting all existing and new objection forms.

It's worth noting that Meta does not use private messages or public data from users under 18 in the EU for training its models. “We believe we have a responsibility to build AI that’s not just available to Europeans, but is actually built for them,” Meta stated. “That’s why it’s so important for our generative AI models to be trained on a variety of data so they can understand the incredible and diverse nuances and complexities that make up European communities. That means everything from dialects and colloquialisms, to hyper-local knowledge and the distinct ways different countries use humor and sarcasm on our products.”

Meta is following in the footsteps of companies like Google and OpenAI, which have already utilized data from European users to train their AI models. However, the DPC continues to keep a close eye on how AI services are trained, as evidenced by their recent investigation into xAI’s training of Grok.