Google adds personalized image generation to Gemini app

Google said it is rolling out personalized image generation in the Gemini app in the US, letting the tool use data from connected Google apps, and optional Google Photos libraries to tailor results.

The update lets users create images with shorter prompts, include labeled people or pets from Google Photos, switch reference photos, and view which source image Gemini used.

Google said Gemini does not directly train its models on private Google Photos libraries, though it uses limited data such as prompts and model responses to improve the product over time.

🔗 Source: Google

🧠 Food for thought

Implications, context, and why it matters.

This feature fits into Google’s wider Personal Intelligence effort

Personalized image generation sits inside a larger beta tool called Personal Intelligence. Google is rolling it out to eligible Google AI Pro and AI Ultra subscribers in the U.S. 1.
Gemini can connect to a user’s Gmail, Photos, YouTube, and Search to use account data for more tailored replies 1.
It can pull a car’s license plate from a photo. It can also pin down the trim level (the specific version of a vehicle model) by searching emails in Gmail 1.
The feature is opt-in and starts turned off. It does not work with Google Workspace business, enterprise, or education accounts right now 2.

Deeper integration raises stakes for privacy plus competition

Google is keeping pace with rivals such as Microsoft Copilot and Anthropic, an AI company, that are building assistants with deeper access to personal and system data 3.
That access can make results feel more personal. It also raises risk since a hacked Google account could reveal a combined record of someone’s life, including financial details in emails plus location history drawn from photos and Maps activity 3.
As these assistants spread across Google plus rival products, regulators and privacy advocates may take a closer look at how sensitive information gets used or protected 3.