Meta to use European users’ public posts for AI training

Facebook and Instagram’s parent company, Meta, will soon use public posts from European users to develop and improve AI, starting June 26. European users of Instagram and Facebook mobile apps were recently informed that their public posts, photos, captions, and messages sent to AI could be used for this purpose, excluding private messages.

Meta stated it will notify UK and EU users about using their public information to enhance AI within the scope of privacy laws. Under the General Data Protection Regulation (GDPR), data processing can occur under “legitimate interest,” which Meta cites as the legal basis for its actions.

Users can opt out by filing a form with Facebook or Instagram, though the Facebook form is currently unavailable. Meta will review opt-out requests according to data protection laws, but data could still be used if users appear in public images or posts by others.

The Irish Data Protection Commission (DPC) noted that Meta delayed the launch following inquiries but has since addressed the concerns. Meta informed the DPC that only public posts by EU users, excluding those under 18, will be used for AI training.

Meta is heavily investing in AI research and development, as highlighted by its recent press release. The company’s large language model, Llama, powers its Meta AI assistant, which is not yet available in Europe. While public Instagram and Facebook posts have been used to train Meta AI, this data did not contribute to Llama 2.

Additionally, Meta is enhancing its AI infrastructure, introducing custom-made chips to support AI workloads.