Meta has announced plans to begin training its generative AI models using public content and user interactions from individuals based in the European Union. The move marks a significant step in the company’s effort to tailor its AI systems to better understand the diverse cultures, languages, and nuances across Europe.

Starting this week, users across Meta’s platforms—Facebook, Instagram, WhatsApp, and Messenger—will receive in-app notifications and emails outlining how their publicly shared posts, comments, and interactions with Meta AI will be used to improve the company's large language models. Importantly, Meta is offering EU-based users the ability to opt out of this data usage through a dedicated objection form, which the company says is simple to access and submit.

According to Meta, the data will only include public content shared by adults and will exclude private messages as well as any information from users under the age of 18. The company emphasized that no private conversations between friends and family will be used for training purposes.

Meta AI, which launched in the EU last month, is now being integrated into messaging platforms to provide users with AI-powered support directly within apps they already use. This latest update is intended to enhance the model's performance by exposing it to real-world examples of European languages, slang, humor, and cultural references.

“We believe we have a responsibility to build AI that’s not just available to Europeans, but is actually built for them,” the company stated in its announcement. “That means training AI to understand local dialects, regional context, and even the different ways humor is expressed across European communities.”

The move comes after a year-long delay, during which Meta paused its original AI training plans to await regulatory guidance from European authorities. In December, the European Data Protection Board (EDPB) affirmed that Meta’s approach met legal requirements. Since then, Meta has been in ongoing dialogue with the Irish Data Protection Commission (IDPC), its lead data authority in the EU.

Meta’s latest effort to train AI using public European content mirrors similar strategies already employed by Google and OpenAI. However, the company claims its rollout is more transparent and user-focused, particularly with the inclusion of the opt-out mechanism.