Google is rolling out a major update for its Gemini AI on Android, enabling it to access WhatsApp, Messages, Phone, and Utilities apps—even if users have disabled “Gemini Apps Activity” in settings. Starting July 7, Gemini can automate tasks like sending texts or initiating calls, but Google insists it will not read or summarize message contents, images, audio, or GIFs.

However, privacy experts remain uneasy. Information collected via notifications may still allow Gemini to preview messages, and Google retains related data for up to 72 hours, even when activity tracking is off. The Economic Times reports this raises alarms about how much control users truly have.

To disable the feature, users must navigate to the Gemini app → Profile → Apps, and manually toggle off access per app. Alternatively, turning off Gemini Apps Activity via Profile → Gemini Apps Activity helps—but may not be sufficient. For stronger protection, experts recommend fully disabling or uninstalling the Gemini app.

This move echoes broader concerns over tech giants quietly enabling AI integrations by default. Earlier integrations from Microsoft and Meta drew similar objections for embedding AI tools without clear user consent .

Users should review Gemini permissions immediately, decide whether convenience outweighs privacy, and stay alert for future Android updates that may re-enable AI integrations.