Google recently pushed back against a viral claim that it was secretly using Gmail data to train its AI. It denied the allegation. Now, a similar claim has surfaced — this time involving Google Photos and its 1.5 billion users.
The claim comes from Proton, a privacy-focused company that competes with Google Drive. Proton suggested on social media that Google’s powerful new AI image tools may be trained using photos stored in Google Photos. Proton did not provide evidence to support the claim.
Google has denied this, saying it does not use personal Google Photos data to train AI image models. The company said photos are only processed according to existing policies, such as when users share images with other services.
Google Photos is not end-to-end encrypted, and Google does scan images to detect child sexual abuse material. However, Google says this scanning is not used to train AI image generators.
The situation highlights a larger issue: many cloud platforms rely more heavily on AI, while privacy policies become harder for users to understand. Most people accept the terms without reading them closely.
As a result, claims like these continue to raise questions about how personal data is handled — and how much users should trust cloud storage without full encryption.
