Meta to Use Facebook and Instagram Data to Train AI: How to Opt-Out
hassan benyoub 
In a significant shift in its European policy, Meta has announced that it will begin using public data from European users of its platforms to train its generative AI models. The change, which is set to take effect on May 27, will allow Meta to use public posts, comments, and other information from Facebook and Instagram to improve its AI's understanding of European communities.
What Data Will Be Used?
The data used to train Meta's AI will include:**
- Public posts and comments on Facebook and Instagram
- Photos, stories, and reels
- Information publicly available on profiles, such as name, username, and profile picture
- Activity in groups, pages, and public channels
- Reviews and ratings
Meta's AI will also use data from Messenger conversations
The data will be used to improve Meta's AI's understanding of European languages, dialects, and cultural nuances. Meta believes that this will enable it to develop AI that is "not only available to Europeans but also designed for them."
Users Can Opt-Out Using a New Form
Meta will provide a notification to European users in the coming days, along with a link to a form where they can opt-out of having their data used to train the AI. The form is already available online. Users who opt-out will receive an email confirmation, and their data will no longer be used to train Meta's AI.
Exceptions to the Opt-Out
While users can opt-out of having their data used to train Meta's AI, there may be exceptions in certain situations. For example, if a user's information is publicly shared by another user, or if their information is mentioned in a public post or comment, Meta may still process their data.
Meta's New Policy is Based on EU Data Protection Rules
Meta's new policy is based on an opinion from the European Data Protection Board (EDPB) published in December. The opinion states that AI developers can use the concept of "legitimate interest" to use personal data for the development and deployment of AI models in the EU, as long as they respect the anonymity of the users whose data is used.
