Google revised its privacy policy, effective October 16, to permit the use of content from Gmail, Drive, and Google Photos for personalized advertising under specific conditions. The update introduces a conditional clause stating that Google does not show personalized ads based on this content "unless you choose to make data from these services available to other Google services that support personalized ads." This change is linked to Google's "Personal Intelligence" initiative, which allows AI features like AI Mode in Search and Gemini to draw on connected user apps for more tailored responses.
The company emphasizes that this data usage is off by default. Advertisers do not gain access to the underlying content of users' emails, files, or photos, and Google states it will not share personally identifying details without permission. However, users who opt into these AI conveniences may inadvertently allow their connected data to inform ad targeting. To manage this, users can opt out by disconnecting the relevant applications within the "Connected Content apps" section of their Google Account’s Data & privacy settings or by disabling personalized recommendations entirely.
This adjustment follows a pattern of recent privacy-related changes as Google integrates artificial intelligence across its products. In May, reports indicated that Google’s Chrome browser began silently installing a multi-gigabyte on-device AI model, and the company removed a privacy claim associated with its Gemini Nano model. These moves have drawn scrutiny regarding how Google handles user data while expanding its AI capabilities.
Google’s modification of its privacy policy represents a strategic recalibration of the trade-off between utility and privacy in the AI era. By tying ad personalization to the activation of "Personal Intelligence" features, Google creates a pathway where enhanced AI functionality comes with an implicit cost to user privacy, even if that cost is opt-in rather than automatic. This approach allows the company to maintain its long-standing public commitment against scanning private communications for ads while simultaneously leveraging the same data streams for commercial purposes when users engage with new AI tools.
The operational risk here lies in user comprehension and consent granularity. While the setting is technically off by default, the complexity of Google’s ecosystem means that many users may not fully understand that enabling AI-driven search or assistant features could expose their metadata or content signals to ad targeting algorithms. Furthermore, the note that previously collected data may linger in search history after disconnection suggests that revoking consent does not immediately erase all traces of past interactions, potentially complicating compliance with data minimization principles and eroding trust among privacy-conscious consumers.
