Google just flipped another switch on your data, and this time it’s not just your search queries at stake. The company has begun rolling out a restructured privacy setting called Search Services History that, by default, saves user-uploaded media, including images, audio recordings, and files, for the explicit purpose of training its generative AI models.
If you had Web & App Activity turned on before (most people do), the new “Save Media” toggle is already enabled for you. In English: Google is now collecting the photos you upload to Lens, the voice clips you record in Translate, and the files you share through Search, and feeding them into its AI training pipeline unless you manually opt out.
What Google is actually collecting
The scope here goes well beyond your typical “we save your search history” disclosure. Search Services History spans Google Search, Lens, Translate, Maps, and Shopping. The media it captures includes Google Lens images, Search Live voice recordings, Translate speaking practice audio, and uploaded files you interact with through any of these services.
Google says the data used for AI model training is anonymized and decoupled from individual user accounts. The retention period stretches up to four years. Human review of your media requires separate consent, which is a small but meaningful guardrail.
