Meta takes action: How to protect your data from AI use!

Meta takes action: How to protect your data from AI use!

April 28, 2025 is an important date for users of meta services such as Facebook, Instagram and WhatsApp. Starting May 27, 2025, Meta plans to use publicly available information and user content to train its AI technology, known as “Meta AI.” There are new regulations in this regard that users should urgently observe.

An objection to the use of personal data is possible and does not require giving reasons. However, users must withdraw their consent by May 26, 2025 at the latest to prevent their already published data from being used in the future. After this date, the data can no longer be retrieved or deleted. Meta intends to use content from users over the age of 18, which includes, for example, messages, photos and posts.

Appeal procedure

To object to the use of data, users must use the online forms on Facebook and Instagram in the logged in area. On Facebook, this takes place in several steps: First, the user must log into their account and go to the help page. There he selects the option “Objection to data use for AI training” and enters his registered email address to submit the form. On Instagram, you follow similar steps, starting with logging in and accessing the privacy settings.

• Anmeldeschritte für Facebook:

1. Anmeldung beim Facebook-Konto.
2. Besuch der Hilfeseite unter facebook.com/help/contact/[formular-ID].
3. Wahl der Option „Widerspruch gegen Datenverwendung für KI-Training“.
4. Eingabe der registrierten E-Mail-Adresse.
5. Absenden des Formulars.

• Anmeldeschritte für Instagram:

1. Anmeldung beim Instagram-Konto.
2. Öffnen der Einstellungen über das Profilbild.
3. Scrollen zu „Datenschutz und Sicherheit“.
4. Auswahl von „Datennutzung und Berechtigungen“.
5. Klick auf „Widerspruch gegen KI-Training einlegen“.
6. Bestätigung mit der registrierten E-Mail-Adresse.

Data protection and possible risks

Although private conversations on WhatsApp are protected by end-to-end encryption, they lose this status when an interaction with “Meta AI” takes place. It is recommended not to reveal sensitive information such as names or health details, especially in chats with the AI, as requests and messages can also be used for AI training purposes. Users can only ignore the functions of “Meta AI” as they cannot be deactivated or hidden.

In Austria, the use of publicly available user data for AI training is controversial under data protection law, and complaints and investigations against Meta's projects are ongoing with the responsible data protection authorities. A survey has shown that around 62% of Austrians oppose the use of their social media data for AI purposes. In light of these concerns, Meta previously temporarily postponed its plans, but is now starting a new attempt with the announced measures.

In summary, it is essential for users of meta platforms to be aware of the upcoming changes and the possibility of objecting to the use of their data. Failure to respond by May 26, 2025 could be irreversible with respect to future use of their content. For detailed information on the objection procedures, Kosmo and Ververbraucherzentrale offer helpful explanations Consumer Service Bavaria provides contextual information on the legal framework.