Google trains AI on your media by default – here's how to stop it

Google has quietly expanded its privacy settings to automatically store user media – including images, audio, and video recordings – for training its AI models. This under-the-radar update, announced via email in June, applies to all Search services: Maps, Shopping, Flights, Hotels, Translate, News, and especially features like Google Lens (visual search) and Live Search (voice input). The company states that saved media is used to “develop and improve Google services and technologies, including AI models and safety measures,” and may involve human reviewers. This reflects an industry shift from scraping public web data to collecting user-uploaded content, similar to Meta's approach with its AI glasses and social media content.

To regain control, users must manually adjust two new settings introduced in the update: Search Services History and Search Services Personalization. On the Search Services History page, you can uncheck the “Save Media” box separately from the general history box, or disable both. You can also set auto-deletion of saved data after 3, 18, or 36 months. Google also split the previous Web & App Activity setting into two separate options, meaning changes to Web & App Activity no longer affect Search-specific data storage. For those concerned about privacy, the fix is straightforward but requires action – and the default remains set to opt-in for AI training.