Google Gemini uses user data and Nano Banana 2 to create personalized images for US subscribers

Google has launched a new feature in its Gemini app that enables highly personalized image generation powered by Nano Banana, drawing directly from users' personal data across Google services like Photos, Gmail, Calendar, and Drive. This update to Gemini's Personal Intelligence capability allows subscribers to create custom images—such as a claymation scene of themselves and their family enjoying a favorite activity—without needing detailed prompts or manual photo uploads. According to Google's official blog, the AI automatically pulls relevant context from connected accounts to make generations feel "deeply personal" and reflective of individual tastes and lifestyles.

The rollout began this week for Google AI Plus, Pro, and Ultra subscribers in the US, with the feature expected to appear in the Gemini app over the next few days. As reported by TechCrunch, this means Gemini can infer likes and interests from prior chats and data without explicit instructions in the prompt. Ars Technica highlights how it simplifies the process by digging into Google Photos, using labels for people and pets to guide realistic depictions. For now, Europe and other regions are excluded, though international expansion of Gemini features is underway, per 9to5Google.

At the core is Nano Banana 2, an advanced image generation model built on Gemini's latest architecture, which integrates multimodal understanding to handle text prompts, photo edits, and high-resolution outputs up to 4K. Google's DeepMind blog details recent enhancements like Nano Banana Pro, which adds state-of-the-art reasoning for precise visuals, legible multilingual text, and editing capabilities—from restoring old photos to creating mini figurines. Users can tweak results easily, select different reference photos, or regenerate with Pro mode for even higher quality, all within the app's tools menu.

Privacy and safety remain priorities, with all generated images embedded with Google's SynthID invisible digital watermark for verification—users can now upload any image to Gemini and ask if it was AI-created by Google tools. Visible watermarks (like the Gemini sparkle) appear on free and Pro-tier outputs, but Ultra subscribers and developers in Google AI Studio get clean versions without them. The Next Web notes that Personal Intelligence leverages data from multiple Google apps while emphasizing secure handling, training only on limited info like specific prompts to improve over time.

This development matters for millions of Gemini users seeking quicker, more intuitive creativity tools, potentially transforming how people visualize personal ideas, family moments, or professional assets without starting from scratch. It affects primarily paid US subscribers initially, who gain an edge in AI-driven art and editing, while raising questions about data usage in personalized AI. Google indicates expansions to more languages, Chrome integration, and additional users are coming soon, alongside broader SynthID support for audio and video.