Meta will train models in the EU using public posts and AI interactions.

On Monday, Meta Platforms (META.O), which launches a new tab, announced that it will train its AI models in the EU using user interactions with the system as well as adult users’ comments and public posts on its platforms.

The Facebook parent company’s action follows Meta’s AI technology roll-out in Europe this month, which was first announced in June 2024 but postponed due to regulatory worries over privacy and data protection.

Although Meta AI was introduced in the United States in 2023, the EU’s strict privacy and transparency regulations presented a number of challenges for its rollout in Europe.

On Monday, Meta said that users of its Facebook and Instagram platforms in the EU would begin to get alerts outlining the types of data the firm will collect. A link to a form where users can protest the usage of their data for training will also be sent to them.

Private conversations and public data from users under the age of 18 will not be used in the training, even if the corporation will use data like user enquiries and Meta AI questions.

A request for comment on Meta’s action was not answered by the European Commission.

Following a directive from Ireland’s Data Protection Commission (DPC) to postpone its plan to incorporate data from social media posts, Meta made the decision to halt the rollout of its AI models across Europe last June. The advocacy group NOYB also criticised it, calling for national privacy watchdogs to prevent this kind of social network content exploitation.

Alphabet’s (GOOGL.O) and Elon Musk’s X open a new tab The Irish privacy regulator is also investigating Google.

While the DPC launched an investigation into Google in September to see if the business sufficiently protected user data before using it to aid in the development of its AI model, X is currently under investigation for exploiting the personal data of EU users to train its AI system, Grok.