What you need to know
- Google’s Gemini gets a big personalization upgrade this week, as artificial intelligence uses Personal Intelligence and the Nano Banana 2.
- Now, users can tell AI to create an unusual photo of them and their family.
- Depending on your simple command, Gemini will access your photos, history, video history, and more to create what it believes you’re looking for.
- Gemini’s personalized image creation is rolling out “over the next few days” to subscribing AI Pro, Plus and Ultra users.
Twins‘s popular photo generator, Nano Banana, gets a big, personal upgrade to recognize who you and your family members are.
In a press release, Gemini teased the personal touch that using the Nano Banana can give the images you want to create. To put it bluntly, Google announced today (April 16) that Gemini Personal Intelligence can now use your memories stored in the Photos app and the Nano Banana 2 to create an AI version of you and your loved ones. Users can connect their Photos app to Personal Intelligence, so the AI can “use real photos of loved ones to guide the photo creation process.”
To do this, Google said Gemini will bend a lot in labels you posted on your photos. This allows the AI to complete the image creation task without needing to say much. One of the examples provided was: “Imagine a favorite hobby of me and my family.” While it may seem vague to an outsider, to a Gemini (and tagged photos), he should know exactly what you’re talking about.
Article continues below
Android Central’s Take
I can see this being helpful for parents and their children. If they have a vision of being a princess or a spacewalker, parents can do it now. It’s just something quick and fun that will make them laugh. It should by no means be used in formal statements, but I don’t think that’s the point. The point here is to use your own memories and put a fun, fantastical spin on it. In this case? I like it.
Moreover, users can paint the image of Gemini in a clay style or with watercolors, as an oil painting, etc. they can be specific by saying to create. We’ve been down this road before, as Google reminded users that Gemini — artificial intelligence — can make mistakes. Furthermore, it may not even select the correct photos to use for the generation process.
If this happens, users can tap the plus sign to select a more suitable photo. Google adds that it does not use your photos as training material for Gemini. This feature is also completely optional, meaning you don’t have to give Gemini access to your photos or engage with Personal Intelligence if you don’t want to.
Who you are makes Gemini go further
Image 1 of 2
This integration for Personal Intelligence and Nano Banana 2 goes beyond mixing you and your family into the content being created. Google says that Personal Intelligence can learn what it knows about you—your tastes and lifestyle—and create images from simple queries. If you told a Gemini to “design my dream house,” he would do his best using the information he knows about you. This can come from your YouTube history or even your most basic Google searches.
Android Central’s Take
Nano Banana really caught people’s attention when it debuted. It has a fun name that may account for its popularity, but it has its uses. One of the abilities users have is to turn a command into a pretentious sculpture, Google said. Now it’s grown to include your memories, who you are, and more. This may bring back some people who previously got tired of typing long instructions over and over again.
Gemini’s new imaging capabilities will roll out “in the next few days” to eligible Google AI Pro, AI Plus and AI Ultra subscribers in the US. The post says there are plans to bring this capability to desktop Chrome users and more “soon.”
Nano Banana was incredibly popularbut its second iteration took it a few steps further. Nano Banana 2 was launched in February and quickly became AI’s default image generator, replacing the former. This version is faster and more accurate when listening to the user’s query. In addition, Nano Banana 2 now offers its content generated in resolutions ranging from 512px to 4K in various aspect ratios.
