The feature uses connected Google apps and labeled Photos to create images without detailed prompts, and is rolling out to paid U.S. subscribers
Thomas Trutschel / Getty Images
Google $GOOGL's Gemini app is rolling out a personalized image generation feature that draws on users' connected Google apps and Google Photos libraries to produce AI images without requiring detailed prompts. The capability is available to Google AI Plus, Pro, and Ultra subscribers in the U.S. and will expand to Gemini in Chrome desktop and additional users in the coming weeks, the company said.
The feature works through Gemini's Personal Intelligence system and its Nano Banana 2 model. The underlying mechanism relies on account-level context Gemini has already built up — through connections to services like Gmail — so a short, unadorned prompt like "Design my dream house" can return results shaped by a user's tastes without requiring any additional personal detail.
Linking Google Photos to Personal Intelligence unlocks an additional layer of specificity: Gemini can read the labels and named groupings — such as "Family" — that a user has applied within the library, enabling prompts that call up recognizable people without requiring a user to manually upload or identify any photos. Style options include watercolors, charcoal sketches, and oil paintings.
For situations where the output misses the mark, TechCrunch reported that Google has built in several corrective options: direct feedback to the model, a "+" icon for swapping in a different reference photo, and a Sources button that surfaces the contextual basis Gemini used to generate the image.
On privacy, Google said the Gemini app does not train its models on users' private Google Photos libraries. The company said it trains on limited information, such as specific prompts and model responses, to improve functionality. Connecting Google apps to Gemini remains an opt-in setting that users can change at any time.
The feature lands as Gemini's image tools have drawn competitive attention. OpenAI upgraded ChatGPT Images with GPT Image 1.5 — framed as a response to the attention Google's Nano Banana models attracted — with the company emphasizing faster performance, lower API pricing, and improved consistency across iterative edits. Google's own developer documentation maps Nano Banana to Gemini 2.5 Flash Image and Nano Banana Pro to Gemini 3 Pro Preview.
Personal Intelligence debuted earlier this year, reaching all U.S. users in March before Google extended access internationally to markets that now include India and Japan. The new image generation capability has also drawn scrutiny in the context of broader privacy concerns around Gemini: a proposed class-action lawsuit filed in federal court in San Jose, California accused Google of activating Gemini by default across Gmail, Google Chat, and Meet without user consent, alleging potential violations of California's Invasion of Privacy Act.
Join 500,000+ readers who start their day with Quartz.
By subscribing, you agree to our Terms of Service and Privacy Policy.
