Overview

• Meta began training its AI models on May 27 using publicly available Facebook and Instagram posts from users in Germany and Europe under the GDPR’s ‘legitimate interest’ provision.
• Those who did not file an opt-out by May 26 cannot exclude their earlier public content, though they can object via Meta’s online forms to prevent future posts from being used.
• Private messages and protected group content remain excluded from the AI datasets, but interactions with Meta’s AI chatbot in WhatsApp can still be collected for training.
• Verbraucherzentrale Nordrhein-Westfalen and other consumer advocates have ongoing legal challenges against Meta’s use of public data, though a court rejected calls to halt the training.
• Privacy experts advise reviewing account settings and avoiding sensitive information in public posts or AI chatbot interactions to limit data exposure.