Google expands data collection for AI training, users can opt out

Google has quietly expanded the data it collects through Search services to include media such as images, audio, and video for training its AI models. The change, rolled out gradually, automatically opts in users but offers a way to disable it.

What changed

Google's updated privacy settings for Search services (including Search, Maps, Shopping, Flights, Hotels, Translate, and News) now save media uploaded via these tools. For example, photos taken with Google Lens, voice recordings from Google voice searches, and audio from Google Translate can be retained and used to train AI Source: TechCrunch, euronewsen. Google Photos is explicitly excluded Source: euronewsen. The company states it uses saved media to "develop and improve Google services, including AI models and safety measures" Source: TechCrunch.

How to opt out

Users can disable “Save Media” on the Search Services History page or configure automatic deletion after 3, 18, or 36 months Source: engadget, euronewsen. Additionally, personalized recommendations can be turned off via Search Services Personalization page Source: engadget. These steps prevent Google from using uploaded media for AI training.

Industry context

This shift reflects a broader trend where tech companies move from web scraping to direct user data collection. Meta also trains AI on European users' public posts, and OpenAI defaults to data sharing for consumer accounts Source: euronewsen. Anthropic offers an opt-in policy Source: euronewsen.

Key points:

• Google's updated policy applies to all Search-related products, including Maps and Translate.
• Users are automatically opted in, but can opt out by unchecking “Save Media”.
• ReCAPTCHA previously used user input for digitization, predating the AI boom Source: euronewsen.