Google launches Personal Intelligence beta for Gemini across Google apps

Google launched a beta of Personal Intelligence for its Gemini artificial intelligence on Jan. 14, 2026, debuting an opt-in system that can draw on a user’s Gmail, Photos, Search and YouTube activity to produce more personalized assistance. The initial release is limited to eligible personal accounts in the United States, offered to subscribers of Google’s AI Pro and AI Ultra plans, and will roll out to users over the following week in English only.

Personal Intelligence is described by Google as a secure connector that, with explicit user consent, combines information from multiple Google services to help Gemini answer questions, recommend options and plan tasks. It can access Gmail, including receipts and messages, Google Photos saved content, YouTube watch history, and Search-related data such as search history, Shopping, News, Maps, Google Flights and Hotels. Google Workspace data including Calendar and Drive may also be linked for personal accounts. The feature is available across the web and in the Gemini app on Android and iOS, and Google says it will work with all models currently in the Gemini model picker.

Google emphasized user control in the rollout. Personal Intelligence requires an opt-in and allows users to selectively connect particular apps or services to Gemini. Users may toggle the feature off, view or delete past Gemini chats, and ask the system to remember personal preferences or life goals to improve future interactions. Google said Gemini will attempt to identify the connected sources it used in an answer so users can verify the basis for recommendations. The company also stated it will avoid making proactive assumptions about sensitive data such as health information, though it will discuss such topics if a user raises them.

Despite those safeguards, Google warned users that the beta could produce errors and “over-personalization,” linking unrelated information or failing to account for temporal nuance, for example, not recognizing changes in relationships or the current status of events referenced in a photo. The company advised participants to provide feedback through in-app tools such as thumbs-down flags to help correct problematic outputs.

Google signaled a broader roadmap for Personal Intelligence. The company plans to expand access to more countries, eventually bring similar capabilities to the free tier of Gemini, and integrate the feature into AI Mode in Search in the near future. It also said additional Google apps and services may be added in the months ahead with user permission.

The rollout underscores the tension between convenience and privacy that has defined consumer AI to date. By drawing together disparate data silos, Personal Intelligence could make planning and discovery markedly faster, helping users plan trips based on past visits or match purchases to specific needs, but it also concentrates sensitive personal signals in a single decision-making interface. That concentration raises questions about transparency, data handling and potential regulatory scrutiny as personalized AI moves deeper into everyday digital life.

For now, the beta will provide a controlled testbed for both the technology and Google’s claims of secure, user-directed linking. How users balance the promise of context-aware assistance against the risks of mis-personalization and privacy erosion will likely shape whether Google expands the feature beyond its early subscriber base.