(Credit: Cheng Xin/Getty Images)
UPDATE (3/18): Google is expanding its Personal Intelligence feature to all US users. The feature taps Google apps like Gmail and Photos to provide tailored responses to your queries in the Gemini app, AI Mode in Search, and the Gemini side panel in Chrome.
Since launching in January, the feature has been limited to Google AI Pro and AI Ultra subscribers. It now supports free users, but is disabled by default. To connect apps via the Gemini app, go to Settings > Personal Intelligence > Connected Apps. In Search, click on your profile photo, go to Search personalization > Connected Content Apps.
Some of the benefits of the crossover include suggestions that match your preferred style and brand when shopping, solutions to your tech queries without needing to mention your model number, and custom itineraries that match your travel interests.
Original Story (1/15): Google is testing a new Gemini experience that allows the chatbot to personalize its responses using data from connected apps such as Gmail, YouTube, Search, and Photos. The feature, called Personal Intelligence, requires users to opt in and choose the apps Gemini can access.
In a blog post, Google’s VP for Gemini, Josh Woodward, shared a couple of examples to show how he uses Personal Intelligence. When Woodward needed to buy new tires for his car, he asked Gemini for advice, but didn’t mention his car model or his typical use cases. The chatbot, however, tapped into Woodward’s emails and photos to find the required details, such as car model, tire size, and trips taken, to suggest the most suitable tires and their pricing.
In another example, Gemini analyzed the family’s interests and past trips from Gmail and Photos to prepare an ideal plan for their upcoming spring break.
Regarding privacy, Google says that since all of the data “already lives at Google securely,” it doesn’t have to send the data elsewhere to begin personalizing.
Additionally, the company claims it does not train Gemini “directly” on your photos or emails. Instead, it may use specific information from prompts and responses to improve Gemini over time. For example, if the car’s license plate was referenced to generate a response, Google says it will use the prompt and response for AI training only after it “obfuscate[s] personal data from the conversation.”
As with most of its AI features, Google warns that Personal Intelligence may not always get it right, adding that it still struggles with timing and nuances.
Personal Intelligence is currently available in beta for Google AI Pro and AI Ultra subscribers in the US. You can enable it from Gemini’s Settings > Personal Intelligence > Connected Apps. The graphics in Google's blog post also indicate that the feature may soon support Google Calendar, Drive, and other Workspace apps.
Apple’s Siri was supposed to receive a similar ability to draw from context across apps in 2024, but that never happened. The company replaced its AI boss, now faces multiple lawsuits, and is aiming for a much-delayed release date: spring 2026. Google will be the one to help push that along. It just announced a partnership with Apple to power Siri using Gemini. When Siri’s context awareness finally arrives, will it be Gemini’s Personal Intelligence in disguise?