Google expands Gemini’s personalized image generation using Google Photos

Google is expanding personalized image creation in the Gemini app by letting eligible U.S. users connect Google Photos so Gemini can generate images using personal photo references.

What Google announced

In a post on the Google Blog, the company said eligible users in the United States can use Personal Intelligence with Nano Banana and Google Photos in the Gemini app for free. Google describes the feature as a way for Gemini to pull from a connected Google Photos library to create more personalized images, including images based on the user or people in their photos.

Google’s earlier announcement, also published on the Google Blog, said the Google Photos connection is opt-in. The company said users can connect Photos so Gemini can use images of the user and loved ones as references for image generation. Google also said its models are not “directly trained” on a user’s private Google Photos library through this connection.

The rollout matters because it moves generative image tools closer to personal photo libraries, where images may include faces, family members, locations, events and other sensitive context. Google is presenting the feature as a personalization tool, while privacy questions center on what information is processed, retained, reviewed or inferred when a user connects Photos to Gemini.

Privacy and review details

Google’s Gemini Apps Help page on personalization with Connected Apps says Google Photos can be one of the connected services used to personalize Gemini responses. The help document says connected app data can help Gemini provide information about “relationships and people in your world,” depending on user settings and requests.

The same Google help page says summaries, inferences and generated media from Google Photos may be reviewed in some cases. That language is important because it clarifies that the privacy story is not limited to whether model training occurs. Even when a company says private photos are not directly used to train models, users may still need to consider review, processing and data handling practices around prompts, outputs and related metadata.

Google’s public materials characterize the feature as optional, not automatic. Based on the company’s descriptions, users must choose to connect Google Photos before Gemini can use that library for personalized image generation.

Availability limits

Engadget reported that Google expanded the feature to eligible U.S. users and noted that Gemini’s Google Photos connectivity is not available in Texas and Illinois. Engadget attributed that limitation to state-level privacy laws, which is notable because both states have been associated with stronger rules around biometric privacy and facial data than many other U.S. jurisdictions.

Google’s announcement and support documentation do not frame the product as a bulk collection program for public faces. The documented feature is tied to a user connecting their own Google Photos account to Gemini. However, because personal photo libraries often include other people, the practical privacy question is broader than the account holder alone: images may contain friends, relatives, children or bystanders who did not separately opt in to Gemini-based image generation.

What users should watch

For users considering the feature, the key points from Google’s own materials are that the Google Photos connection is opt-in, the company says private Photos libraries are not directly used to train Gemini models, and Google’s help page still allows for some review of summaries, inferences and generated media in certain cases.

That combination may be acceptable to some users who want more customized image generation. Others may prefer not to connect a personal photo library to an AI assistant, especially if their library includes sensitive images or photos of people who have not consented to that use.

The expansion shows how major AI products are increasingly moving from generic prompts toward personal context. In Gemini’s case, Google is making that context available through Google Photos, while users and regulators are likely to keep scrutinizing how consent, biometric privacy and human review are handled.