Meta is moving ahead with a major AI data change in the EU: public content from Facebook and Instagram users will be used to train its AI models. The company says the rollout begins this week and will also include users’ interactions with Meta AI.
What Meta is changing in the EU
Meta announced on Monday that it will train its AI models on public content in the EU. The material includes public posts and comments on Facebook and Instagram, along with interactions people have with Meta AI.
The company had previously paused these plans after regulatory pressure tied to data privacy concerns. That pause made the EU a different case from markets such as the U.S., where Meta has been training AI on user-generated content for years.
The change comes after a limited version of Meta AI launched in the EU last month. That EU launch arrived well after Meta AI had already appeared in the U.S. and other global markets.
Why the rollout was delayed
The EU has been a difficult environment for companies that want to use personal data to train AI systems. The source article points specifically to the General Data Protection Regulation (GDPR), which requires a clear legal basis for processing personal data to train AI models.
Meta said back in June 2024 that it would pause plans to begin training AI systems with user data in the EU and U.K. That move followed pushback from the Irish Data Protection Commission (DPC), which regulates Meta in the EU and was acting on behalf of several data protection authorities across the bloc.
In September 2024, Meta said it was restarting efforts to train AI systems with public posts from its U.K. user base. The new announcement extends that direction to public posts from users in the EU.
Meta has linked the restart to regulatory clarification. The company said it delayed training large language models using public content while regulators clarified legal requirements, and it welcomed an EDPB opinion in December that Meta said affirmed its original approach met its legal obligations.
How users will be notified
Starting this week, Meta says EU users will begin receiving in-app and email notifications. These notices are meant to explain that the company will use public data and interactions with Meta AI to train its models.
The notices will include a link to a form that lets users opt out. Meta says it will honor objection forms it has already received, as well as newly submitted ones.
For users, the key practical details are:
- Public Facebook and Instagram posts and comments can be used for AI training.
- Interactions with Meta AI can also be used.
- Users in the EU will receive in-app and email notices.
- The notices will link to an opt-out form.
- Meta says existing and new objection forms will be honored.
Meta also says there are limits to what it will use. It does not use private messages to train its models, and it does not use public data from users under the age of 18 in the EU.
Meta’s argument for local training data
Meta’s case is that European data can help make its generative AI more useful for people in Europe. The company says AI models need a variety of data to understand the nuances and complexities of European communities.
According to Meta, that means learning from things such as dialects, colloquialisms, hyper-local knowledge, and the distinct ways different countries use humor and sarcasm on its products. In plain terms, Meta is arguing that AI systems trained without local public content may be less able to reflect how people in Europe actually communicate online.
Meta also says it is following the example of companies like Google and OpenAI, which it says have already used data from European users to train AI models.
Scrutiny is not over
The announcement does not mean EU regulators are stepping away from AI training questions. The DPC is still examining how large language model creators train their AI services.
Last week, the regulator announced it was investigating xAI’s training of Grok. That separate investigation shows the broader issue remains active: how AI companies gather, process, and justify the data used to train their systems.
For Meta, the immediate change is clear. Public EU Facebook and Instagram content is being brought into its AI training pipeline, with notices and opt-out forms offered to users. The larger question is how regulators, companies, and users continue to define the limits of AI training on public social content.