Google adds image generation to Gemini that draws on your email, photos and calendar

Google is taking another step toward truly personal AI: Gemini can now generate images based not only on the text of a request but also on data from the Google ecosystem, from Gmail and Photos to Calendar and Drive. The update adds image generation based on Nano Banana 2 to the Personal Intelligence feature.

The logic is simple. Instead of writing a long prompt that lists your tastes, habits and personal details, you can type something like "make a visual of my dream home," and Gemini pulls the relevant context from connected services. If Google Photos is linked to Personal Intelligence, the model can use not only general preferences but also real photos, as well as labels such as "family" or "pets," to build more accurate and personal images.

For Google, this is an important extension of the strategy it launched in January 2026 with Personal Intelligence. Until now, the feature worked mostly in text scenarios: it helped find trip details in email, assemble plans from the calendar, factor in purchase history and make Gemini's answers less generic. Now the same idea carries over to visuals. Personalization is no longer an add-on to the chat but part of the content generation mechanism itself. This matters especially amid competition in which users already expect AI to deliver not just a pretty picture but a result that reflects their habits, interests and surroundings from the first request.

Technologically, Google is betting on Nano Banana, the family of image generators built into Gemini, rather than on the Imagen line, which is geared more toward standalone visual and professional tasks. The Nano Banana line already has several versions: the base Nano Banana for fast everyday scenarios, Nano Banana 2 with faster iterations and expanded capabilities, and Nano Banana Pro, which uses the full reasoning capabilities of Gemini 3 Pro. The key bet is that the image is created not by an external module but inside the same system that understands the text, the conversation context and the user background tied to the account.

Google separately stresses transparency and control. If Gemini picks the wrong reference or gets details wrong, you can refine the result with ordinary text or manually add another photo with the "+" button. Personalized images also get a Sources section, which shows which data or photos were used as the basis. The company says it does not train its models directly on the private Google Photos library, and that connecting Personal Intelligence remains voluntary and is configured by the user. This is a crucial point: the more deeply AI relies on personal data, the more sensitive it is to errors, privacy and trust.

The rollout is phased. On April 16, 2026, Google announced that the feature is starting to appear for Google AI Plus, Pro and Ultra subscribers in the US. It promises to add it later to Gemini in Chrome on desktop and extend it to other audiences. The launch is not available everywhere: Europe was not in the first wave, and Google's help materials also list the European Economic Area, the United Kingdom and Switzerland among unavailable regions. For the company, this is a typical cautious launch scenario, especially for products closely tied to personal data.

The main takeaway is simple: the next battle in the AI market is no longer only about generation quality but about the depth of personal context. Google holds a nearly unique position because it already has email, photos, calendar, search, maps and video history inside a single account. If the company can balance convenience and privacy, Gemini will gain a very strong advantage: less manual description, more useful results from the first request.