On a Tuesday afternoon in June, millions of Google users received an email that most will probably delete without reading. The message announced a new setting called "Search Services History" — a change that, on the surface, sounds administrative, but which quietly expands the range of personal data Google is entitled to keep. For the first time, the images you feed to Google Lens, the voice recordings captured by its real-time Search Live tool, audio spoken into Translate, and uploaded video files will all be saved under a single, distinct category. Previously, these interactions were folded into Google's broader "Web & App Activity" controls, where they were easier to overlook.

The practical consequences are significant. Google has stated that it intends to use Search Services History to "provide, develop, and improve its services" — language that explicitly includes training its AI models. If users also activate the optional "Personalized Recommendations" setting, that same pool of multimodal data can be used to shape the advertisements and suggestions they see. Critics of the tech industry have long argued that AI development has been subsidised, without meaningful consent, by the everyday behaviour of ordinary users.

This restructuring makes that relationship considerably more visible, though it does not fundamentally alter the underlying bargain.