Conceptual illustration of a smartphone displaying an abstract photo gallery and a closed padlock, with subtle Italian flag colours in the background.

Meta fined €805 million in Italy over Facebook’s access to phone galleries

Last updated: Reading time: 2 min

Italy’s data protection authority, Garante, announced an €805 million fine against Meta Platforms Ireland Limited on 14 November 2026. According to the regulator, Facebook accessed Android users’ photos, videos and metadata in Italy without separate, specific consent or adequate information. The case covers around 20 million people.

The investigation concerned a feature that suggested material from a phone’s gallery as ideas for Facebook posts. Garante’s findings therefore relate to the app’s access to content stored on a device, rather than only photos and videos that users had already chosen to publish on the platform.

According to the Italian regulator’s statement, that access also covered metadata — information describing the files. Garante found that the feature did not give users adequate information or obtain separate, specific consent for this use of their data.

60 days to change how the feature works

Alongside the fine, the regulator ordered changes to the service. Meta has 60 days to ensure that the feature is switched off by default and that activating it requires the user’s separate, specific consent. The distinction matters: the order concerns not just information about access to the gallery, but also the feature’s default setting and the way consent is obtained.

The announcement does not establish that Meta has paid the fine or implemented the required changes. Garante’s statement confirms the penalty and sets out the company’s obligations, but does not confirm that they have been fulfilled.

The investigation did not determine whether content from users’ galleries was used to train artificial intelligence. According to the regulator’s statement, Meta denies using it for that purpose. The findings about access to photos and videos should therefore not be treated as a finding that those files were used to train AI models.

Sources

Found an error or outdated information? Let us know.

Latest posts