New privacy controls let users block their search media from AI training. The setting is on by default. Find out how to manage your data before it’s used for years.
Google has begun rolling out updated privacy settings that automatically allow the company to use users’ search-related media—such as images, audio, and video—for artificial intelligence training. The new feature, called Search Services History, is enabled by default for most accounts and can be managed through the My Activity page.
When the update reaches your account, you’ll find that the option to save uploaded media for AI training is already checked. Unless you previously disabled both Web & App Activity and Search Personalization, your search uploads—including photos for reverse image search and voice recordings from Google Translate—may be retained and used to improve Google’s AI models.
To prevent your media from being used in this way, visit the My Activity page, select the Search Services History tab, and uncheck the box labeled Save media. This section also allows you to turn off the entire setting and delete your activity. Google warns that if your media is used for AI training, it is separated from your account but may be kept for up to four years, even if you delete the original activity.
According to Google spokesperson Davis Thompson, these settings are designed to help users revisit past searches, including visual and voice queries, and can be toggled on or off at any time. However, Thompson did not address why the feature is enabled by default.
The scope of data collection is expanding beyond typed queries. Google’s description clarifies that saved media includes images, files, and audio or video recordings from interactions with Search services, such as Google Lens, Search Live, Translate speaking practice, and voice searches. This broader data pool is intended to help Google’s AI models learn from more diverse inputs, potentially giving the company an edge over competitors.
Thorin Klosowski, a senior security and privacy activist at the Electronic Frontier Foundation, notes that Google’s vast ecosystem of services makes it uniquely positioned to collect large amounts of user data. He argues that requiring users to opt out, rather than opt in, places an unfair burden on individuals who may not realize their data is being used for AI training.
Ben Winters, director of AI and privacy at the Consumer Federation of America, says the constant need to opt out of data collection can leave users feeling powerless. He warns that this approach may contribute to widespread fatigue and resignation about personal data protection.
Google’s email notification about the change, sent to users on June 23, frames the update as giving people “even more control over saved history.” The message highlights benefits like revisiting past visual searches or continuing conversations in Search Live, but offers little detail about how saved media will be used for AI model training.
This shift in privacy policy comes as scrutiny of AI-powered search features intensifies. For example, a recent German court decision held Google liable for misinformation generated by its AI Overviews, underscoring the growing legal and ethical challenges facing tech giants as they expand AI capabilities.
Google, founded in 1998, remains the world’s dominant search engine, processing over 8.5 billion searches per day as of 2026. The company’s AI initiatives span products like Google Lens, Translate, and Gemini, and its vast user base provides a unique advantage in collecting diverse data for model training. Google employs more than 180,000 people worldwide and reported annual revenues exceeding $300 billion in 2025, reflecting its continued influence in both search and artificial intelligence.
