Google has rolled out a new feature that enables Gemini AI to utilize personal data from applications like Gmail, Photos, and YouTube to respond to inquiries. This feature, called ‘Personal Intelligence’, is currently in beta for Google AI Pro and Ultra subscribers in the US. By activating this option, Gemini can leverage a user’s interconnected data to offer specific responses, like identifying a license plate number from an image or recommending a trip based on previous emails.
According to Google, the feature has been developed with a focus on user control and privacy. It remains deactivated by default, requiring users to manually select which apps to link. The company emphasizes that while data from a user’s inbox or photo library is not used for training its primary AI models, it is temporarily accessed to address a query. Users retain the ability to disconnect apps or deactivate the feature whenever they choose.
Google has acknowledged certain limitations of the beta version, including potential inaccuracies and instances of “over-personalization”, where the AI may misinterpret a user’s preferences based on their data. The company is actively seeking feedback from initial users and plans to extend the feature to more countries and free users in the future. Notably, this functionality is not accessible to business or educational accounts.