Your Data, Google's AI: The Shocking Truth About Training AI and How to Escape It
Google has updated its privacy settings to store user media and search data for AI model training, a change that impacts various Google services. Users are opted in by default, but options are available to adjust preferences and opt out through new dedicated privacy settings.
Consider this an important public service announcement: Google has recently updated its privacy settings, enabling the company to store a broader range of user data, specifically including media such as images, files, and audio and video recordings. This stored data is then utilized to enhance Google's artificial intelligence models. Essentially, any media uploaded to Google's Search services is now being used for AI training by default, unless users actively choose to opt out.
This significant change was implemented through an under-the-radar update to Google's Search services privacy settings, which was initially announced to customers via email in June. Google framed this expansion of AI training as a means of providing users with more control over their saved history and personalized recommendations, thereby automatically opting users into this data collection for AI purposes.
The update introduced two distinct new settings: 'Search Services History' and 'Personalized Recommendations'. These settings allow users to configure how their activity is leveraged to personalize their Google experience and determine the duration for which their web and app activity is saved. The scope of this update extends beyond just Google Search, encompassing other integrated search services such as Maps, Shopping, Flights, Hotels, Translate, and News.
Practical examples of this data collection include images snapped and uploaded when using Google Lens for visual searches, which may now be saved for AI training. Similarly, audio recordings from using the newer 'Search Live' feature for voice input in the Google app, or any other Google voice search, can be stored. Even audio data from using Google Translate to practice speaking is subject to being saved.
These changes are indicative of a wider industry trend where tech companies are increasingly gathering diverse forms of data to improve their AI services. Instead of solely relying on information scraped from the public web, companies like Google, and Meta (which trains its AI on users’ images, media, and content from its AI glasses), are actively collecting data directly uploaded or created by users within their services.
Google explicitly confirms this media-training usage. In its customer email, the company stated, "Like your Search Services History, your saved media is also used to develop and improve Google services and technologies, including AI models and safety measures." This is further corroborated by its help documentation, which notes that Google "uses your history to provide, develop, and improve its services (such as training generative AI models) and to protect Google, its users, and the public with the help of human reviewers." While some of this data storage is temporary and necessary for product functionality, Google's own language indicates that saved media can be specifically retained for AI training.
Fortunately, users retain a degree of control over these settings. Preferences can be adjusted on the 'Search Services History' and 'Search Services Personalization' pages. Within the 'Search Services History' page, users have the option to uncheck the 'Save Media' box independently from the 'Search Services History' box, or uncheck both. Additionally, users can configure automatic data deletion after specified periods: 3 months, 18 months, or 36 months. From these pages, users can also navigate to other privacy settings, including 'Web & App Activity', 'Timeline', and 'YouTube History'.
It's important to note that beyond saved media, Google also utilizes search history, location data, and other information from websites users visit to personalize their Google experience, including targeted advertisements. Prior to this recent update, Google allowed users to configure historical search data saving through its 'Web & App Activity' settings. However, this has now been bifurcated into two distinct settings: the original 'Web & App Activity' data and a new 'Search data' setting, which is activated by default. This separation means that making changes to the 'Web & App Activity' data retention settings will no longer automatically impact the storage of your Google Search services data, as it is now a separate, default-enabled option.