Deadline for objection expiresMeta wants to use your data for AI - how to protect yourself
SDA
24.5.2025 - 09:41
Meta wants to train its AI software Meta AI using user data from Instagram and Facebook. Consumer advocates see a breach of data protection law and have now gone to court. (symbolic image)
Karl-Josef Hildenbrand/dpa
Meta wants to use personal data from Facebook and Instagram users for AI training. Anyone who does not want this must object by Monday.
Keystone-SDA
24.05.2025, 09:41
24.05.2025, 10:48
SDA
No time? blue News summarizes for you
Meta wants to use public data from Facebook and Instagram users to train its artificial intelligence from May 27, 2025.
Users must actively object if they do not want their data to be used.
Data protectionists criticize the practice as problematic and doubt that it complies with EU law.
Meta wants to use public data from adult users on Facebook and Instagram to train its artificial intelligence from May 27, 2025. A court has permitted this as long as those affected do not actively object.
Which data is affected, how the AI training works and how you can protect yourself:
What does Meta want to use the data for?
Meta wants to train its AI models with as much data as possible. For example, the Meta AI chatbot, which is integrated into the Whatsapp messenger, is to be improved. Meta is also working on other models, such as Llama, a so-called large language model that can be accessed via Microsoft's AI platform Azure AI.
What data does Meta want to use?
Meta wants to use all publicly available information. This includes the information that Facebook and Instagram users provide about themselves on their profiles, as well as posts, photos, videos, comments under other posts and reactions such as likes. In this case, "public" means that the visibility of the content is not already restricted in the profile settings. Meta also states that it only wants to use information from adults.
Is Meta allowed to do this?
The Austrian data protection organization None of Your Business (Noyb), for example, assumes that Meta is violating EU law by not explicitly obtaining users' consent to the use of their data. Instead, they only have the right to object, so they have to take action themselves.
The Consumer Association of North Rhine-Westphalia filed an urgent application with the Cologne Higher Regional Court to prevent the AI training, which was launched on Monday. However, the court rejected this and referred to a corresponding assessment by the Irish data protection authority responsible for Meta. There was no violation of EU law.
Why is the use of the data controversial?
Little is known about the exact use of the data. It is also not clear whether and, if so, how Meta intends to protect personal data or is even able to do so once it has been incorporated into the AI. In addition, the Llama language model, for example, is an open source model, i.e. freely accessible. "This means that Meta can hardly recall or update a model once it has been published," explains Noyb.
Can I object to the use later?
According to Meta, it is possible to object to the use of data at any point in the future, although all data stored by Facebook or Instagram will have already been removed by then. According to the current state of technology, experts do not believe it is possible to remove data from an AI at a later date. Data protection activist Marcel Schrems from Noyb therefore also sees a violation of the right to be forgotten guaranteed under EU law.
The form can also be found in the respective apps, but is well hidden. In the Instagram app, it can be found in the "Privacy Center" sub-item of the profile settings. In the Facebook app, there is a "Privacy Policy" item in the settings menu. If the same email address is used for both platforms, it is sufficient to fill out the form for one of them.