Facebook parent Meta is to begin training its artificial intelligence (AI) offerings on public posts from EU users this week, after delaying doing so last year over regulatory concerns.

The company said public posts from Facebook and Instagram would be used to train its AI models, as well as users’ interactions with its Meta AI chatbot.

Private messages and material from users under the age of 18 would not be used, Meta said.

Delayed launch

The company launched a limited version of Meta AI in the EU in March, after postponing the launch last June over questions around how the use of users’ data would comply with the EU’s GDPR data protection requirements.

Meta AI launched in the US last year, but was delayed in the UK and the EU as well as some other markets. The tool launched in the UK last October.

Last June Meta faced scrutiny from Ireland’s Data Protection Regulator (DPC), its main EU data regulator, over compliance requirements for the AI offering.

Meta said an opinion provided by European Data Protection Board last December affirmed that its original approach met its legal obligations.

The opinion, which was requested by the Ireland DPC, discussed legitimate data collection for AI models under the GDPR.

Beginning this week users in the EU will see in-app and email notifications explaining how Meta will use their data to train its AI models, and providing a link to a form enabling them to opt out.

Opt-out model

Privacy group Noyb and other EU civil liberties groups had objected to Meta’s plans to begin using users’ data without asking for their consent, which they said is required under the GDPR.

“It seems that Meta is doing everything it can to never get opt-in consent for any processing,” Noyb said at the time.

Meta’s current plan does not involve asking users to opt-in for data usage.

Ireland’s DPC last week opened an investigation into social media platform X over its use of users’ posts for training its Grok AI model without informing them until afterward.

Last September the DPC also began a probe into Google’s AI training and whether it properly safeguarded users’ data in the process.