AI platform Gemini caught scanning Google Drive files without user permission

Google’s Gemini AI has been caught scanning PDF files hosted on Google Drive without permission or active initiative, sparking a new discussion around AI security and privacy concerns.

Senior Advisor, AI Governance Kevin Bankson approached X with concerns about an AI-generated summary in a private and confidential tax return.

Bankston’s thread details his experience with the Gemini AI reading private documents without consent and subsequent issues with disabling the feature on the cloud storage platform.

Gemini may be deeper in Google Drive than you thought

Despite attempts to disable the feature, Bankston found that Gemini continued to work in Google Drive, raising questions about Google’s handling of user data and privacy settings.

After failing to find the proper commands to disable Gemini integration, the advisor twice asked Google’s AI chatbot, a competitor to ChatGPT, to figure out the settings. A second, more detailed response still didn’t satisfy: “Gemini is *not* in my dashboard apps and services (option 1), and I didn’t have a profile picture in the top right corner of the Gemini page (option 2).”

With the help of another X user, Bankston found the control, which was already disabled, highlighting either a faulty control or indicating that other settings are hidden elsewhere.

TechRadar Pro asked Google for comment on the matter, but the company did not immediately respond.

Google’s previous filings have confirmed that the company will not use Google Workspace data to train or improve its generative AI services or to power targeted ads. However, there are still concerns about data hygiene and cross-contamination that could ultimately lead to data leaks.

Bankston speculates that his prior involvement with Google Workspace Labs may have influenced Gemini’s behavior, but his concern about hidden settings and controls remains an issue that Google needs to address in order to satisfy concerned users.

Additionally, the case reaffirms the importance of user consent and control as AI continues to advance and become increasingly integrated into the products we use every day.