Google rolls out Gemini Personal Intelligence

Google began rolling out a beta feature called Personal Intelligence within the Gemini app, letting the assistant draw on a user’s Gmail, Google Photos, YouTube history and Search history to answer questions with personal context rather than generic replies. Google’s example was a travel query: Gemini could pull an itinerary from Gmail, recall a destination documentary watched on YouTube, and check saved places from Search history to assemble a route and packing suggestions in one response.

The feature was opt-in and off by default. Users chose which of the four apps to connect, in a single tap each, and could disconnect them at any time. Google said Gemini would not train directly on the contents of a user’s Gmail inbox or Photos library, though it would use limited information from prompts and responses for model improvement generally.

Access began with subscribers to Google’s paid Google AI Pro and AI Ultra tiers in the United States, rolling out over the following week, with the company saying it planned to expand to more countries and to free tiers, and to bring the same capability to Google’s AI Mode search feature. The launch positioned Gemini more directly against Apple’s efforts to fold personal context into its own assistant, and extended a broader industry move — also visible in ChatGPT’s memory features — toward assistants that persist context across a user’s own data rather than starting each conversation from nothing.