Google is expanding the role of its AI assistant by allowing Gemini to draw deeper context from a user’s own apps, introducing a new capability it calls Personal Intelligence. The feature, announced Wednesday, is designed to let Gemini reason across multiple Google services to deliver more relevant and personalised assistance.
Powered by Gemini 3, Personal Intelligence enables the assistant to connect information from various apps, including Gmail, Google Photos, YouTube, and Google Search. Rather than introducing a new interface, the capability runs quietly in the background, enhancing how Gemini understands user queries by factoring in context from across the Google ecosystem.
According to Google, the goal is to move beyond generic answers and provide responses shaped by a user’s actual activity. For example, Gemini could reference a detail from an email, connect it with a photo in a gallery, or relate a YouTube video to a recent search—only when such connections are useful and relevant. When it does so, the assistant will indicate the source of the information.
Josh Woodward, vice president of the Gemini app and Google Labs, said the company is aware of potential pitfalls during the beta phase. He acknowledged the risk of inaccurate responses or “over-personalisation,” where the system might incorrectly link unrelated information.
In such cases, Google is relying heavily on user feedback to refine the feature. Users can also correct false assumptions directly in chat or request a new response if personalisation misses the mark.
Privacy controls built in
Given the sensitivity of personal data, Google emphasised that Personal Intelligence is turned off by default. Users must explicitly opt in and can choose which apps, if any, they want to connect.
The company also stressed that Gemini does not train directly on the contents of Gmail inboxes or Photos libraries. Instead, training relies on limited data such as prompts and model responses, not the underlying personal details found in connected apps.
Google added that the system is designed to avoid making assumptions about sensitive topics, including health-related information. Users retain the option to disconnect apps or limit personalisation at any time.
Availability and rollout
Personal Intelligence is currently available in beta to paid subscribers, including Google AI Pro and AI Ultra users, with plans to expand access to free users in the future. Eligible users can enable the feature through the Gemini settings menu under Personal Intelligence and Connected Apps.
Once activated, the capability works across web, Android, and iOS, and can be used with any model available in Gemini’s model picker. Currently, it is limited to personal Google accounts and does not apply to Workspace business, enterprise, or education plans. Google also confirmed that Personal Intelligence will soon be integrated into Search through AI Mode.
With this move, Google is positioning Gemini closer to the vision of an always-contextual assistant—one that quietly observes patterns across daily digital life to offer smarter help, while leaving the final level of access firmly in the user’s hands.
