Key Points:
- Meta will restart its AI training using publicly available content from European users after a pause due to privacy concerns.
- The training data includes public posts, comments, and user interactions with Meta AI — excluding private messages and content from minors.
- The decision follows the launch of Meta AI in Europe and regulatory approval from EU privacy authorities.
- Users in the EU will be notified and given the option to opt out of their data being used
Meta to Resume AI Training With Public Content From European Users After Regulatory Green Light
Meta, the parent company of Facebook and Instagram, announced on Monday that it will resume training its artificial intelligence models using public data from European Union users — a move it had paused last year in response to widespread data privacy concerns.
The tech giant said it will utilize public content such as posts, comments, and user interactions with its Meta AI assistant to enhance its machine learning models. This update comes just weeks after Meta rolled out its AI assistant across platforms in the European Union, including Facebook, WhatsApp, and Instagram.
According to Meta, only information from adult users that is already publicly visible will be used. Private messages, sensitive content, and data from users under 18 will remain off-limits. “Interactions with Meta AI — like questions and prompts — will also be included to refine our models and improve user experiences,” the company said in a blog post.
Meta’s renewed push into AI training in Europe comes after it received confirmation from the European Data Protection Board that its approach aligns with the continent’s strict privacy standards.
The Vienna-based privacy advocacy group NOYB, led by Max Schrems, had previously urged regulators to block Meta’s original plans, citing potential violations of users’ data rights.
To comply with EU regulations, Meta is rolling out a transparent notification system for users. Starting this week, individuals across the EU will begin receiving in-app and email notices explaining what kind of data will be used and how. Each message will include a direct link to an opt-out form, allowing users to decline participation in AI data training.
Meta emphasized that this strategy follows the precedent set by competitors like Google and OpenAI, both of which have incorporated publicly available European data into their AI training pipelines.
“This training will help us build AI tools that better reflect European languages, cultures, and communities,” Meta stated. “We’re committed to honoring any objection forms previously submitted, as well as new ones moving forward.”
The move marks a significant step in Meta’s broader AI development strategy as it looks to compete globally in generative AI while navigating regional regulatory hurdles.
Related blog post: OpenAI Strikes Back in Legal Clash with Elon Musk