Key Takeaways
- Meta will begin training its AI models using content from adult users in the EU, aiming to make its tools more culturally relevant.
- The Facebook parent claimed all users will be able to opt out of the scheme.
- Meta’s announcement follows initial backlash against its plans in 2024.
Meta has announced that it will begin using public posts, comments, and other interactions on its platforms to train its artificial intelligence models within the European Union.
This development follows similar moves by other major AI players, including OpenAI and Google, who have begun using user data in the EU to improve their AI systems.
Meta’s AI Training Strategy
On Monday, April 14, Meta stated it would use publicly shared content from adult users to train its AI tools.
Facebook and Instagram’s parent company said the initiative aims to better support millions of people and businesses across the EU.
In a blog post , Meta emphasized that training on public content is essential for teaching its AI systems to “better understand and reflect [users’] cultures, languages, and history.”
Meta has ensured that users would have the ability to opt out before the data collection begins.
According to the tech giant, EU users will receive notifications explaining the new policy and a link to a form where they can object to the use of their data.
“We’ve made the objection form easy to locate, read, and complete, and we’ll honor all previously submitted forms as well as new ones,” the company stated.
Only data from users aged 18 and older will be used. Private messages are explicitly excluded from the training process.
Regulatory Hurdles
Meta AI was first introduced in the U.S. in 2023, but its rollout in Europe has slowed due to stringent EU privacy laws.
Last year, Meta postponed its AI training plans after the Irish Data Protection Commission (IDPC) requested a pause on using social media content for training purposes.
The privacy advocacy group None of Your Business (NOYB) subsequently filed more than ten complaints against Meta, arguing that the approach violated the General Data Protection Regulation (GDPR).
NOYB founder Max Schrems told Forbes : “Meta is basically saying it can use any data from any source for any purpose and make it available to anyone, as long as it’s through ‘AI technology.’ This is clearly not compliant with the GDPR.”
Schrems also highlighted the vagueness of Meta’s terminology: “‘AI technology’ is extremely broad—like saying ‘using your data in databases’—and carries no clear legal boundaries.”
He added that Meta failed to clearly explain how the data will be used, raising concerns about its potential applications—from targeted advertising to more controversial uses.
AI Nuance
Meta maintains that training its AI on public EU data is necessary to ensure the models are helpful and relevant to European users.
“We believe we have a responsibility to build AI that isn’t just available to Europeans, but is actually built for them,” the company stated.
The company said that training on diverse and localized data will help Meta’s AI better understand regional dialects, cultural nuances and humor.
“This is particularly important as AI models become more advanced with multi-modal functionality, which spans text, voice, video, and imagery,” it added.
Meta did not respond to CCN’s request for comment at the time of reporting.
Meta Antitrust Trial Starts: Timeline, Key Witnesses, and What’s at Stake
