A Silent Data Deal with Major Consequences
Whether it's holiday photos, birthday greetings, or dance videos – what we've shared on Facebook and Instagram over the years is now being used for a new purpose: training Meta's Artificial Intelligence. Without active objection from users, publicly posted content will flow into the database for AI development starting May 27, 2025. Data protection advocates are strongly opposing this, but many users don't even know they are affected.
1. What exactly is Meta planning?
Meta (formerly Facebook) is preparing its AI models for new capabilities – they are intended to generate images, write texts, and hold conversations. The basis for this? Public content on Facebook and Instagram. Anyone over 18 who has posted publicly is potentially providing material for AI training. This includes:
- Photos and videos (including captions)
- Comments under posts
- Profile pictures and usernames
- Content in public groups
Private messages (e.g., on WhatsApp) are excluded due to end-to-end encryption – with one exception: conversations with the AI chatbot "Meta AI" can indeed be analyzed and stored.
2. Why is this problematic?
The model Meta is relying on is called "opt-out": those who do not object automatically agree. This practice contradicts European data protection law, which requires users to actively consent. Consumer advocates and lawyers speak of a massive breach of trust and have already sent Meta a warning.
The main criticisms are:
- Users are not directly notified.
- Content from third parties can also be affected (e.g., party photos in which you appear).
- Content that has been learned is difficult to delete.
3. How can you object to the use of your data?
Fortunately, there is a way to protect yourself:
- Log in to Facebook or Instagram
- Here is the link to the Instagram form and here is the link to the Facebook form
- Access Meta Accounts Center: https://www.facebook.com/privacy/center/meta-ai
- Fill out the objection form
- Wait for email confirmation
Important: You must object separately for each social network (Facebook AND Instagram), unless you use the Accounts Center with linked accounts.
4. What happens if you do nothing?
If you do nothing, your data will automatically be considered released. Starting May 27, 2025, Meta can include your public content in the training of its AI.
Even if you are not on Facebook or Instagram yourself, but appear in a public photo of another person, you could be affected.
5. Can you request subsequent deletion?
Only in rare cases. For example, if Meta AI provides information upon request that should not have been used, you can report this case. For this, you need:
- The exact prompt that led to the illegitimate answer
- A screenshot of the AI response
- A link to the deletion request form
However, it is hardly realistic that data once learned can be effectively removed from the system.
6. What does this mean for the future?
This development is just the beginning. Platforms like Meta will continue to try to use their data treasures for AI purposes in the future. It is all the more important for users to take active steps and understand their rights.
Conclusion: Act now before it's too late
Anyone who wants to protect their privacy should file an objection by May 26, 2025 at the latest. Because once something has entered Meta's AI system, it is almost impossible to retrieve.
More information and direct links to the forms can be found via your Meta account or on the website of the Verbraucherzentrale NRW.
