NextFin News - Google’s latest privacy change is a reminder that the boundary between everyday search use and AI training has become thinner than most users realize. A recent update to Google’s Search Services controls introduced separate settings for Search Services History and Personalized Recommendations, while also making saved media part of the data that can be used to improve Google’s services. The practical effect is simple: unless users review the new controls carefully, more of what they upload or save around Google’s search ecosystem can be used to develop the company’s products, including AI systems.
The shift matters because it changes how users think about consent. Before the update, one history setting covered more of the activity stack. Now Google has split the controls, so managing how long web and app activity is stored is no longer the same as deciding whether media can be saved for service improvement. Google says the Search Services pages let users change preferences separately, including the option to uncheck “Save Media,” and to choose automatic deletion after 3, 18, or 36 months. That is a more granular system, but it is also more complex, and complexity often works in favor of the platform rather than the user.
Google’s own wording makes clear that activity can be used well beyond simple personalization. Its Gemini Apps Privacy Hub says, “Google uses your activity to provide, develop, and improve its services (including training generative AI models).” The same hub adds that activity can be reviewed by human reviewers and that some settings allow audio, Gemini Live videos, and screenshares to be used to improve and develop Google’s services. That is an explicit statement that user activity can become training material for generative AI models when the relevant setting is turned on.
The Search Services update extends that logic into another corner of Google’s product family. The company says saved media can be used to “develop and improve Google services and technologies, including AI models and safety measures.” That language is broader than a simple storage policy. It links ordinary user uploads to model development and safety improvements, which is exactly why privacy advocates and power users have been treating the update as a material change rather than a cosmetic redesign.
Google’s search, maps, video, email, photos, and cloud products all sit inside the same account ecosystem, so any adjustment to data controls can ripple across services. Search queries show intent. Images can reveal objects, places, and context. Files may contain receipts, travel plans, documents, or work notes. If those inputs are stored and reused under new controls, the data pool available for personalization and AI improvement becomes richer and more sensitive.
That is the tension at the heart of the change. Google is not hiding the setting, but it is expanding the number of places users must check to understand how their data is handled. A person who believes turning off one history toggle is enough may still be leaving other activity types active. In that sense, the update is less about a single opt-out and more about a new map of permissions, one that requires deliberate auditing if users want to limit data reuse.
Google’s privacy hub also shows how the company is segmenting the data it can use. It says your activity is auto-deleted after 18 months by default, though users can change that to 3 or 36 months or choose not to auto-delete. It also says some audio, video, and screenshare data are not used to improve Google services by default unless a specific setting is turned on. That separation matters, because it means Google is attempting to distinguish between categories of content rather than applying one blanket policy to everything. Still, each category is another possible source of training data.
Google uses your activity to provide, develop, and improve its services (including training generative AI models).
That line is the clearest summary of the company’s approach. Google is not merely retaining data longer; it is building a framework in which product improvement, personalization, and model training all sit close together. For the company, that makes technical and commercial sense. For users, it means the privacy cost of convenience may be more distributed than they think.
One reason the update drew attention is that it affects more than classic Search. Google says the change applies to other search services as well, including Maps, Shopping, Flights, Hotels, Translate, and News. That widens the scope from web search alone to a broader set of daily actions. A route search, a shopping query, or a travel plan can now sit in the same policy universe as a typed search, which gives Google a more complete picture of user behavior and likely improves the company’s ability to personalize results and train systems that depend on behavioral patterns.
The company’s controls still give users options, but the burden is now on them to use them carefully. Google says users can review and delete activity in Gemini Apps Activity, turn settings on or off, and manage auto-delete preferences. The Search Services pages similarly let users separate media storage from history storage. Yet the broader lesson is not that privacy is gone; it is that privacy management has become fragmented. The more granular the controls, the more likely it is that many users will leave the defaults untouched.
That default behavior is where the policy shift becomes most consequential. If a platform can widen the range of data that is eligible for improvement and AI training while presenting the change as a set of user controls, it can expand its data advantage without needing a dramatic product launch or a loud public announcement. In the AI era, that is often how meaningful changes arrive: not as a single headline, but as a settings update.
For users who want to reduce how much data Google can use, the practical response is to review the Search Services History page, the Personalized Recommendations page, and the broader Gemini and account activity controls. That does not eliminate the tradeoff, but it makes it visible. And visibility is the first requirement for any meaningful opt-out.
The broader market implication is straightforward: Big Tech’s AI systems are increasingly being fed by ordinary consumer settings rather than just by obvious AI prompts. As companies connect more services to model improvement, the real contest is not only over which AI product is best. It is also over which platform can collect the richest data with the least friction.
That is why this change matters beyond Google itself. It shows how the AI era is turning privacy settings into infrastructure. The question for users is no longer just what stays on their account. It is what becomes training data.
Explore more exclusive insights at nextfin.ai.
