Google’s Gemini AI now integrates Gmail, Drive, and Chat data into research reports, raising privacy and security concerns.

Google has improved the Deep Research feature of its Gemini AI model to extract data straight from users' Google Chat, Gmail, and Drive accounts. The latest update was released today, enables the tool to add web-sourced data along with personal emails, files, spreadsheets, slides, PDFs, and chat conversations into thorough research reports.

The new feature enables easier interaction among groups and experts. By sharing their Drive creative documents, relevant email threads, and project chats with Gemini, users may immediately start a market analysis. This provides an extensive report that links external data with internal strategies. In the same way, creating a competitor analysis may include submitting comparison spreadsheets while Gemini searches the public internet for information on similar products.

Users mistakenly expose huge amounts of private information, such as proprietary strategies, client interactions, or proprietary data, to Google's processing infrastructure by giving AI access to sensitive repositories like Gmail and Drive.

Cybersecurity experts advise about threats like prompt injection attacks, in which malicious inputs could deceive the AI into ignoring or collecting personal data. In order to reduce data exposure, organizations now need to apply blind trust principles and conduct thorough audits of AI permissions. The need for clear data handling processes is shown by Google's own experience, including previous Gmail scanning concerns. Establishing a balance between increasing efficiency and ensuring security is essential as AI systems like Gemini progress. Although this update is creative, it serves as a reminder that data control cannot be sacrificed for convenience.

Since Google focuses on high value on user controls, such as selecting specific sources before beginning research, the default accessibility may result in unintentional data breaches. Google states this as a highly requested feature, and it is currently available to all Gemini users on desktop through the Tools menu. A mobile release is planned for a future date.

Keywords: