If you use Google, you’re training its AI. Here’s how to opt out.
Key Takeaways
- TechCrunch reports that a recent, quietly rolled-out change to Google’s Search services privacy settings now allows the company to retain more user-generated media, including images, files, and audio and video recordings, to help train its AI models.
- According to TechCrunch, the update was communicated to customers via an email in June and introduced two new controls, Search Services History and Personalized Recommendations, which govern how long activity is stored and how it personalizes the Google experience.
- TechCrunch notes users can limit this by unchecking a “Save Media” option separately from Search Services History, and by setting automatic deletion windows of three, 18, or 36 months.
What Changed, According to TechCrunch
TechCrunch reports that Google has adjusted the privacy settings tied to its Search services in a way that expands the kinds of personal data eligible for use in training its artificial intelligence systems. The report states that this now extends to media people upload, including images, files, and audio and video recordings, not just typed queries or browsing history.
Per TechCrunch, the shift was rolled out under the framing of giving users more control over saved history and personalized recommendations, but the practical effect was that people were opted into the broader AI training use by default. The change reportedly arrived through a customer email in June, rather than a prominent public announcement, which is part of why TechCrunch frames this as a belated public service message for readers who may not have noticed.
The new settings described in the report, Search Services History and Personalized Recommendations, let users adjust how their activity shapes their Google experience and how long that activity, including web and app data, is retained. TechCrunch specifies that this update is not confined to the core Google Search product. It also touches other search-adjacent services such as Maps, Shopping, Flights, Hotels, Translate, and News.
TechCrunch gives concrete examples of how this plays out in everyday use. A photo captured through Google Lens for a visual search may now be saved for AI training purposes. Voice input submitted through the Search Live feature in the Google app, or through other Google voice search functions, can likewise be retained. Even audio recorded while practicing pronunciation in Google Translate is covered, according to the report.
Google’s own language, as quoted by TechCrunch, confirms the intent directly: the company told customers that saved media, like Search Services History, is used to develop and improve Google services and technologies, including AI models and safety measures. Google’s help documentation, also cited in the report, states that history is used to provide, develop, and improve its services, including training generative AI models, and to protect Google, its users, and the public with the assistance of human reviewers.
Why This Matters for the AI Economy
TechCrunch situates this move within a broader industry pattern in which large technology companies are shifting away from relying solely on data scraped from the open web to train AI systems. Instead, firms are increasingly drawing on content that users themselves upload or generate while using everyday consumer products. TechCrunch cites Meta as another example of this approach, noting that the company trains its AI on users’ images and other media, as well as on content captured through its AI glasses.
For readers who track the economics of the AI sector, this pattern is worth watching because proprietary, user-generated data of this kind is difficult for competitors to replicate and is not available through public web scraping. Companies that can lawfully and continuously harvest first-party media, voice recordings, and behavioral signals from billions of daily interactions gain a distinct input advantage in model development, independent of how much computing power or capital they can deploy elsewhere. That dynamic can influence how investors and markets assess which AI developers hold durable data moats, even though TechCrunch does not report any specific market, stock, or valuation figures tied to this particular policy change.
There is also a governance dimension. TechCrunch’s account shows that Google separated what used to be a single Web & App Activity control into two distinct settings, Web & App Activity data and a new Search data setting that is on by default. As a result, TechCrunch explains, someone who previously adjusted their Web & App Activity retention preferences to limit data collection would find that this earlier configuration no longer restricts the newer Search data setting, since it now operates independently. That kind of default-on structure and setting fragmentation is a recurring theme in debates over digital consent, and it matters to anyone weighing how much personal content, financial details included, they route through cloud-connected AI tools.
What Users Can Actually Do
TechCrunch outlines a practical path for those who want to limit this data use. Users can visit the Search Services History and Search Services Personalization pages in their Google account settings. On the history page, it is possible to uncheck the “Save Media” box independently of the broader Search Services History box, or to disable both together. Users can also choose an automatic deletion schedule of three months, 18 months, or 36 months for stored data, according to the report. TechCrunch further points readers toward additional privacy controls covering Web & App Activity, Timeline, and YouTube History for a more complete adjustment of what Google retains, including search history, location, and other browsing-derived signals used to personalize ads and recommendations.
Hype Check
Claim: Google is now converting everyday uploads, photos, voice searches, and translated speech, into fuel for its AI models by default. Reality: TechCrunch confirms this is accurate based on Google’s own customer email and help documentation, and the report also confirms that Google provides opt-out controls, including a separate “Save Media” toggle and configurable deletion windows, meaning the situation is neither fully hidden nor fully reversible without user action. Verdict: Substance. This is not financial advice.
Source
Researched with AI assistance, fact-checked and edited by a human. Not financial advice.