Meta will pause its plans to train its AI technology by using public content shared on Facebook and Instagram across the EU/EEA, the Irish Data Protection Commission (DPC) announced on Friday.

The decision follows privacy complaints in 11 European countries brought by the European Center for Digital Rights (NOYB), which also urged the authorities to take action on the matter.

Using content since 2007

Meta recently informed millions of European users that it was changing its privacy policy to be able to use user information to train its AI technology. The changes, due to come into force on June 26, would have given Meta the right to use personal posts, private images or online tracking data that were published across both social media platforms since 2007.

“Meta’s new privacy policy basically says that the company wants to take all public and non-public user data that it has collected since 2007 and use it for any undefined type of current and future ‘artificial intelligence technology’,” NYOB said.

The company would also have been able to ingest personal data from any source, both public and non-public, and potentially share this with third parties.

“This is a step backwards for European innovation, competition in AI development and further delays bringing the benefits of AI to people in Europe.”

Meta

The only way to opt out of this processing was to fill out an objection form before June 26, 2024, and explain why – which had to be done for each account. NOYB also said that after that date, users appeared to have no option of being able to remove the data.

The only information that would not be used in the training were messages and pictures that are posted in a private profile, and data from accounts of users under the age of 18.  

By introducing such AI models, NOYB alleges that Meta would violate EU GDPR Articles 5(1), 5(2), 6(1), 9(1), 12(1), 12(2), 13(1), 13(2), 17(1)(c), 18(1)(d), 19, 21(1), and 25.

Irish DPC and UK ICO

In a statement, Meta expressed disappointment. “This is a step backwards for European innovation, competition in AI development and further delays bringing the benefits of AI to people in Europe,” the company said.

“If we don’t train our models on the public content that Europeans share on our services and others, such as public posts or comments, then models and the AI features they power won’t accurately understand important regional languages, cultures or trending topics on social media.”

Meta also said it is consulting with the Irish DPC, and has incorporated feedback to ensure that the AI training is complying with EU privacy laws. During the pause, Meta will also address specific requests from the UK’s Information Commissioner’s Office (ICO).

The DPC also said that it will co-operate with its fellow EU data protection authorities to continue engage with Meta on this issue.