Gemini AI platform caught scanning Google Drive files without user consent
Google’s AI Gemini has been caught scanning PDF files hosted on Google Drive without active consent or initiative, reigniting the debate over AI security and privacy concerns.
Senior Advisor for AI Governance Kevin Bankson went to X to express their concerns about an automatically generated AI statement in a private and confidential tax return.
In Bankston’s thread, he described his experiences with Gemini AI reading private documents without permission, and the subsequent difficulties in disabling the functionality on the cloud storage platform.
Gemini may be buried deeper in Google Drive than you thought
Despite attempts to disable the feature, Bankston discovered that Gemini was still working in Google Drive, raising questions about how Google handles user data and privacy settings.
After failing to find the right controls to disable Gemini’s integration, Advisor asked Google’s ChatGPT-rival AI chatbot twice to figure out the settings. A second, more detailed response still brought no joy: “Gemini is *not* in Apps & Services on my dashboard (1st option), and I didn’t have a profile picture in the top right corner of the Gemini page (2nd).”
With the help of another X user, Bankston found the control, which was already disabled. He indicated that it was either a control that was malfunctioning or that there were more settings hidden elsewhere.
Ny Breaking has asked Google for comment on the matter, but the company has not yet responded.
Google’s previous documentation has indicated that the company will not use Google Workspace data to train or improve its generative AI services or to serve targeted ads. However, concerns remain about data hygiene and cross-contamination, which could ultimately lead to data breaches.
Bankston suspects his previous participation in Google Workspace Labs may have influenced Gemini’s behavior, but his concerns about hidden settings and controls are still an issue Google needs to address to appease concerned users.
Moreover, the case underscores the importance of consent and user control as AI continues to evolve and become more deeply embedded in the products we use every day.