Google Gemini Introduces Personal Intelligence Beta to Pull Data from Gmail, Photos and More
Key Points
- Google launches Personal Intelligence beta for Gemini AI assistant.
- Feature accesses Gmail, Photos, Search and YouTube history to provide proactive answers.
- Opt‑in design; the feature is off by default and users choose which apps to connect.
- Gemini can combine text, images and video to answer queries, e.g., tire size from photos.
- Assistant can offer personalized travel planning, media recommendations and more.
- Personal data is referenced only for responses, not used to train the model.
- Guardrails prevent unsolicited use of sensitive information.
- Initial rollout to Google AI Pro and AI Ultra subscribers in the United States.
- Google plans to expand to additional regions and the free tier.
Google has launched a beta feature for its Gemini AI assistant called Personal Intelligence. The feature lets Gemini access a user’s Gmail, Photos, Search and YouTube history to deliver proactive, context‑aware answers. It is off by default, giving users the choice to connect their apps. Gemini can combine text, images and video to answer queries such as locating a tire size from a photo or planning a trip based on past emails and travel photos. The company says the model only references personal data without using it for training and includes safeguards for sensitive topics. The rollout begins with Google AI Pro and AI Ultra subscribers in the United States, with plans to expand to more regions and to the free tier.
New Personal Intelligence Feature in Gemini
Google announced a new beta capability for its Gemini AI assistant, branded as Personal Intelligence. This feature enables Gemini to retrieve and reason across a user’s personal data stored in Google services, including Gmail, Google Photos, Search history and YouTube watch history. While Gemini already could fetch information from these sources, the beta allows it to synthesize that data to provide proactive, tailored responses without the user explicitly directing it to a particular app.
The Personal Intelligence experience is disabled by default. Users must opt in and grant permission for Gemini to access their selected Google apps. Once enabled, Gemini will decide when to draw on personal data, aiming to intervene only when it believes the information will improve the answer.
Google describes two core strengths of the feature: the ability to reason across complex sources and the capacity to retrieve specific details from items such as an email or a photo. In practice, Gemini can pull together text, images and video to answer a single query. For example, a user asking about a car’s tire size could receive a suggestion that references family road‑trip photos stored in Google Photos. Similarly, the assistant can locate a license‑plate number from a picture in the user’s photo library.
Beyond simple facts, Gemini can help with planning and recommendations. By analyzing past travel emails, photos of previous trips and YouTube videos watched, the assistant can suggest itineraries that avoid typical tourist traps, propose alternative transportation options, and even recommend board games for the journey. The system also offers personalized suggestions for books, shows, clothing and meals based on a user’s documented interests.
Google emphasizes that Gemini does not train directly on a user’s Gmail inbox or Photo library. Instead, the model uses the personal data only as a reference point when generating a response to a specific prompt. The underlying AI model continues to be trained on its standard datasets, separate from the individual’s private content.
To address privacy concerns, the company has built guardrails that limit proactive use of sensitive data. Gemini will avoid making unsolicited assumptions about health‑related information or other delicate topics, though it will discuss such data if the user explicitly asks for it.
The beta rollout is initially available to subscribers of Google AI Pro and AI Ultra in the United States. Google plans to extend the feature to additional countries and to users on Gemini’s free tier in the future. The company also released a list of example prompts that illustrate the range of tasks Personal Intelligence can handle, such as planning a weekend based on personal preferences, recommending documentaries, or identifying YouTube channels that match a user’s cooking style.