Starting September 2nd, Google will begin utilizing user-uploaded content samples including documents, images, videos and screenshots to train its Gemini AI engine. This marks a significant shift in the company's approach to AI development.
The update involves a strategic rebranding of current Gemini activity preferences to "Retention of Activity Logs." Notably, this new preference will be enabled by default unless users manually opt-out, triggering widespread privacy discussions.
Privacy Concerns with Activity Retention Settings
The implementation has immediately raised questions about consent protocols, data transparency and protection measures. Privacy experts warn that sensitive personal information could be inadvertently exposed through AI training without explicit user awareness.
While Google emphasizes responsible data usage limited to sample processing, critics highlight the opacity of the renaming process. The new "Retention of Activity" label lacks the clarity of its predecessor "Gemini Application Activity" designation.
Disabling Gemini Activity on Desktop Devices
Data-conscious users can easily disable the feature via Gemini.Google.com by: logging in > navigating to "Settings & Help" > selecting "Activity" > turning off Gemini activity tracking. Existing data deletion is also possible to minimize exposure.
On mobile devices, users should open the Gemini app, tap the account icon, access Gemini activity settings and disable the function. Historical activity deletion remains available.
Google warns that previously collected data may remain on servers for up to 72 hours post-disablement before permanent removal. Multi-account users must repeat the opt-out process across all active Google profiles.
September Implementation Changes
Currently labeled "Gemini Application Activity," the feature will transition to "Retention of Activity" display across web and mobile platforms starting September. This name change combined with the default-enabled setting increases user awareness regarding potential AI training implications on private communications and files.
The latest Gemini AI enhancement initiative demonstrates a clear trade-off between performance optimization and escalating privacy considerations for users.
