Google's Gemini now generates personalized images using your Google Photos library

Google's Gemini can now create personalized images by accessing users' Google Photos libraries through its Personal Intelligence feature. The capability, powered by Google's Nano Banana 2 image generation model, allows Gemini to produce images that reflect individual users' tastes and lifestyles based on their photo collections.

How the integration works

Users who opt into Personal Intelligence can prompt Gemini with requests like "Design my dream house" or "Create a picture of my desert island essentials." The system uses labels within Google Photos to identify people including users, friends, and family members, according to Google spokesperson Elijah Lawal. Nano Banana 2 then generates images incorporating these personal elements.

The generated images "automatically reflect your specific tastes and lifestyle, gleaned from the Google apps you've connected to," according to Google's blog post announcing the feature.

Privacy and training data

Google states it will not "directly train" its AI models on users' private Google Photos libraries for those who opt into Personal Intelligence. However, the company does train on "limited info" including "specific prompts in Gemini and the model's responses."

Users must explicitly opt in to the Personal Intelligence feature to enable this functionality.

Availability

The feature is rolling out "over the next few days" to AI Plus, Pro, and Ultra subscribers in the United States. Google plans to expand availability to Gemini on Chrome desktop and "more users" soon, though no specific timeline was provided.

What this means

This marks Google's first integration of personal photo data into AI image generation, leveraging its extensive Photos ecosystem to differentiate Gemini from competitors. The approach mirrors broader industry trends toward personalized AI assistants, but raises questions about how companies balance personalization with user privacy. The limited training on prompts and responses, rather than raw photo data, suggests Google is attempting to thread this needle—though the distinction may prove difficult for users to understand fully.