Meta resumes AI training using EU user data

After nearly a year’s pause due to regulatory concerns, Meta has begun harvesting public content from its European users to train its AI models, just as EU officials prepare to issue the first-ever fines under the bloc’s Digital Markets Act (DMA).

Meta announced Monday it will start using public posts, comments, and AI interactions from adult users across Facebook, Instagram, and WhatsApp in the EU to improve its generative AI systems. The European Data Protection Board (EDPB) approved the rollout.

“This training will better support millions of people and businesses in Europe, by teaching our generative AI models to better understand and reflect their cultures, languages, and history,” Meta said in its official announcement.

Meta was previously barred from using EU data, stating in 2024 “Without EU user data Meta says “we’d only be able to offer people a second-rate experience. This means we aren’t able to launch Meta AI in Europe at the moment.”

Meta is offering opt-out options 

European users will begin receiving notifications this week, both in their apps and via email, explaining exactly what data will be collected and how it will be used. These notifications will include a link to an objection form where users can opt out.

“We have made this objection form easy to find, read, and use, and we’ll honor all objection forms we have already received, as well as newly submitted ones,” Meta’s press release stated.

The company emphasized that “we do not use people’s private messages with friends and family to train our generative AI models.” 

Additionally, “public data from the accounts of people in the EU under the age of 18 is not being used for training purposes.”

Meta also pointed out it’s “following the example set by others including Google and OpenAI,” noting both companies have already used data from European users to train their AI models.

DMA fines on the horizon

The timing of Meta’s announcement is noteworthy, coming just as the European Commission prepares to issue what are expected to be substantial fines against both Meta and Apple for alleged violations of the new Digital Markets Act.

Competition Commissioner Teresa Ribera reinforced the Commission’s enforcement intentions Tuesday, telling the European Parliament: “If we do not see willingness to cooperate we will not shy away from imposing the fines identified by the law.”

Companies found in breach of the DMA can be fined up to 10% of their total worldwide turnover, increasing to 20% for repeated infractions.

While the EU wants to enforce fines against Big Tech, preventing model training on citizens’ data has proved not to be possible. 

For European users who don’t want their data harvested by Meta, keep an eye out for Meta’s notifications. 

The post Meta resumes AI training using EU user data appeared first on DailyAI.