Meta announced on Monday that it’s going to train its AI models on public content, such as posts and comments on Facebook and Instagram, in the EU after previously pausing its plans to do so in response to regulatory pressure due to to data privacy concerns. The company will start training its AI on users’ content in the EU this week, it said. Users’ interactions with Meta AI will also be used to train its models.
The announcement comes after a limited version of Meta AI launched in the EU last month, well after its debut in the U.S. and other global markets.
While Meta has been training its AI on user-generated content in the U.S. for years, it has faced resistance in the EU due to the bloc’s strict privacy laws, particularly 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 start training its AI systems using user data in the EU and U.K. following pushback from the Irish Data Protection Commission (DPC). The DPC 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 its AI systems using public posts from its U.K. user base.
Fast forward to today, Meta has announced that it will do so with public posts from its EU userbase as well.
“Last year, we delayed training our large language models using public content while regulators clarified legal requirements,” Meta said in its blog post. “We welcome the opinion provided by the EDPB in December, which affirmed that our original approach met our legal obligations. Since then, we have engaged constructively with the IDPC and look forward to continuing to bring the full benefits of generative AI to people in Europe.”
Starting this week, users in the EU will start receiving in-app and email notifications to explain that Meta will start using public data and interactions with Meta ai to train its models. These notifications will include a link to a form that will allow users to opt out of their data being used. Meta says it will honor all objection forms it has already received, as well as newly submitted ones.
Meta notes that it doesn’t not use private messages, nor public data from users under the age of 18 in the EU, to train its models.
“We believe we have a responsibility to build AI that’s not just available to Europeans, but is actually built for them,” Meta says. “That’s why it’s so important for our generative AI models to be trained on a variety of data so they can understand the incredible and diverse nuances and complexities that make up European communities. That means everything from dialects and colloquialisms, to hyper-local knowledge and the distinct ways different countries use humor and sarcasm on our products.”
Meta says it’s following the example set by companies like Google and OpenAI, both of which have already used data from European users to train their AI models.
Meanwhile the DPC is not moving on entirely from scrutinizing how Large Language Model creators are training their AI services. Last week, the regulator announced it was investigating xAI’s training of Grok.