Gemini can now generate personalized images from your Google Photos

Google has expanded the Personal Intelligence feature in Gemini: the AI assistant can now create personalized images using data from Google Photos and other connected apps.

Until now, Personal Intelligence let Gemini tailor its answers to your context — recommending restaurants based on your preferences, for example, or reminding you of tasks from Gmail. Now the same mechanism applies to image generation. It runs on Nano Banana 2, Google's in-house image generation model.

In practice, it works like this: a user types a prompt such as "draw my ideal home" or "show what I'd take to a desert island," and Gemini creates an image that reflects your personal style rather than an averaged-out result drawn from the internet. Under the hood, the system uses labels from Google Photos to recognize you, the people close to you, and your familiar surroundings.

The feature only works if Google services are connected and the user has explicitly consented to the use of their personal data. Google stresses that processing stays within its ecosystem and that data is not shared with third parties.

This marks a notable shift in how generative AI works: instead of a generic response to a prompt, users get one tailored to a specific person. Few competitors are pursuing this direction so far. Apple Intelligence bets on personalization through on-device processing, but it has not yet reached image generation based on personal context. With a vast trove of user data across Photos, Gmail, and Calendar, Google holds a structural advantage here.

The further AI assistants push into personalization, the more pressing the question of trust becomes: how comfortable are users with giving a model access to their photo albums in exchange for more accurate images? Google will have to win people over not only with the quality of the output, but also with transparency about exactly how personal data is used.