The feature, once locked behind a paid subscription, now draws on Google Photos, Gmail, and Search to generate images tailored to individual users, provided they opt in
Google has expanded access to one of Gemini’s more advanced features, making personalised AI image generation available to eligible users in the United States at no extra cost. The capability had previously been reserved for select paid subscribers.
How the Feature Actually Works
Google is rolling out an integration of their Personal Intelligence system with Google Photos and their Nano Banana Image Model to provide users with images that are based on their own preferences and past experience. It provides an alternative to simply inputting text for generic images and now incorporates the user’s experience to create an image that the user would have wanted to see.
What Data Gets Used
To make this work, Gemini taps into information from connected Google services, including Gmail, Google Photos, YouTube, and Search, to build a clearer picture of a user’s preferences and memories. Google has been explicit that none of this happens automatically. The feature only activates once a user explicitly grants permission to connect these services, meaning the personalisation is opt-in rather than a default setting applied without consent.
Why This Matters
Many companies in the AI industry have begun to place more value on personalisation than generative power so that they are able to keep the user engaged. Image generation has quickly gone from being unique to being a commodity as more and more companies are producing be similar capabilities. With Google relying on the wealth of personal data available to them in Google Photos, Google Search, and Google Mail, Google’s position in the market is strengthened by being able to utilise this data as a potential product benefit, provided the user is comfortable with the giving permission to utilise this amount of data.
The extent to which users will be willing to give Google’s AI full access to their historical data to generate images will ultimately depend on whether they are comfortable allowing the AI to have full access to their image data and will dictate just how widely Google will be able to roll out this service beyond the initial offering of free images to users in the US.
