Google now trains AI on your uploaded media. Here's how to protect your privacy settings.

2 Sources

Share

Google quietly updated its privacy settings in June, allowing the company to store and use uploaded images, audio, and video recordings from Search services to train AI models. Users are automatically opted in, but can disable the feature through Search Services History and Personalized Recommendations settings. The change affects Google Lens, Voice Search, Translate, and other search tools.

Google Uses Your Uploaded Search Media to Train AI Models

Google introduced a significant change to its privacy settings in June through a customer email that flew under the radar for many users. The company now collects and stores media that people upload to its Search services, including images, files, and audio and video recordings, to improve its AI models

1

. Users are automatically opted into this expanded Google AI training program, marking a shift in how the tech giant handles user data across its ecosystem.

Source: Engadget

Source: Engadget

The update affects multiple search-related products beyond Google Search itself, including Maps, Shopping, Flights, Hotels, Google Translate, and News

1

. When you use Google Lens to search visually by snapping a photo, that image may now be saved for AI training purposes. Similarly, audio recordings from Google Voice Search and the newer Search Live feature in the Google app could be stored and used to develop AI models

2

. Even practicing pronunciation through Google Translate means that audio becomes part of the training data pool.

Understanding the Google Privacy Settings Update

The Google privacy settings update introduced two new configurations: Search Services History and Personalized Recommendations. These settings allow users to control how their activity personalizes their Google experience and how long their web and app activity is saved

1

. Google essentially opted people into this expanded data collection under the guise of providing more control over saved history and personalized recommendations.

Google confirms the use of user-uploaded media for AI models directly in its customer communications, stating that "saved media is also used to develop and improve Google services and technologies, including AI models and safety measures"

1

. The company's help documentation notes that it uses this history with the assistance of human reviewers to provide, develop, and improve its services, including training generative AI models. While some data storage is temporary and necessary for product functionality, saved media can also be retained specifically to train AI systems.

How to Opt Out of Google AI Training

Users who want to prevent their user-generated data from being used for AI training have options to adjust their privacy settings. To opt out, visit the Search Services History page and uncheck the "Save Media" box, which can be toggled separately from the "Search Services History" box

1

. You can also configure automatic deletion intervals for saved data, choosing between 3 months, 18 months, or 36 months

2

.

Additionally, users should check the Search Services Personalization page to ensure it's not saving anything beyond their preferences

2

. Before this update, Google allowed configuration of historical search data through "Web & App Activity" settings, but that has now been separated into two distinct settings, with the new Search data setting turned on by default

1

.

The Broader Industry Shift in Data Collection

This change reflects a broader industry trend toward gathering data by any means necessary to fuel AI development. Instead of relying solely on information scraped from the web, Google and other tech companies are increasingly collecting data that people upload or create when using their services

1

. Meta represents another example of a consumer-facing tech company implementing similar practices at scale, training its AI on users' images and media, as well as content recorded by its AI glasses

1

.

The appetite for training data appears insatiable, as the "gaping maw of generative AI needs data to feed on and it's running out"

2

. This suggests that users should expect more companies to follow similar paths, making proactive data storage management increasingly important for those concerned about how their personal content is being used to develop AI technologies.

Today's Top Stories

© 2026 TheOutpost.AI All rights reserved