Google's Gemini AI gets a personalization boost, letting subscribers generate family photos that draw on images stored in Google Photos. By linking the Photos app to Gemini’s Personal Intelligence, the model can pull labeled pictures and produce a quirky scene from a brief prompt such as “my family at the beach.” The feature arrives this week for AI Pro, Plus and Ultra plans.

The system leans on the tags users already applied to their photos, so a vague request still yields a recognizable result. Users may refine output by selecting a different source image via the plus icon. Google stresses that none of the personal pictures are used to train Gemini and that participation is entirely optional.

Nano Banana 2, the engine behind Gemini’s image generator, now supports 512 px to 4K resolutions and a range of aspect ratios, making the output suitable for everything from phone wallpapers to prints. Because the tool taps into a user’s own visual history, it promises a more playful, family‑centric experience than generic text‑to‑image models. Availability expands to desktop Chrome soon, but only for paying subscribers.