Meta to resume AI training on content shared by Europeans

Meta announced today that it will soon start training its artificial intelligence models using content shared by European adult users on its Facebook and Instagram social media platforms.

The content used for AI training includes posts and comments from adult users, as well as questions and queries made when interacting with the company’s Meta AI assistant. However, the company says it won’t use “people’s private messages with friends and family” or the “public data” of Europeans under 18 to train its AI models.

People in the EU who use the social media giant’s platforms will start receiving in-app and email notifications this week explaining the kind of data Meta will begin using.

These alerts will also include a link to an online form where they can object to their data being used this way. Meta says the objection form is easy to find and use and will honor all submitted forms.

“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,” it said in a Monday press release. “People based in the EU who use our platforms can choose to object to their public data being used for training purposes.”

Rollout resumes after June delay

Today’s announcement comes after Meta delayed training its AI using public content shared on Facebook and Instagram at the request of the Irish Data Protection Commission in June 2024.

The 2024 delay was the result of NOYB, an Austrian nonprofit organization working to enforce data protection laws, filing complaints in 11 European countries, including Ireland, where the DPC initially approved the introduction of Meta AI in the EU/EEA but reversed its decision following pushback from other EU data protection authorities.

However, on Monday, Meta said the European Data Protection Board (EDPB), the EU’s data protection regulator, approved this new rollout because the approach meets the European Union’s stringent data privacy laws.

“We welcome the opinion provided by the EDPB in December, which affirmed that our original approach met our legal obligations,” it said. “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.”

Last month, the social media giant also started rolling out its Meta AI assistant in WhatsApp across Europe. Meta says that personal chats will not be used to train Meta AI, but conversations with Meta AI could be used to train future versions.