Here's an Example of What Google's Gemini 2.5 AI Can Do

Over the past few years I have experimented a bt with various AI graphics programs and writing programs. I thought it my be fun to share some images showing what Google's Gemini AI can do (version 2.5 Flash)

Most of my pantings begin with a background. Backgrounds vary based on intentions. Some are very intentional while many are exploratory, in which I let the brushwork dictate the direction the painting goes. This is a medium sized canvas, 24"x 36". I began with these brushstrokes and colors, unsure where it would take me. A few years ago when I visited the studio of abstract painter Elizabeth Kuth, I saw how she would paint in layers, periodically rotating the canvas 180 degrees, bringing new elements into view, which she would highlight or bury beneath more visual elements. After a few rotations and layers I painted the following incomplete form of a man. Unsure of where to go with it, this has been siting on one of my easels since late last fall.

Last night I shared the above image with Gemini and 

asked that the AI to paint Vincent Van Gogh's face on my 

incomplete piece. Two minutes later I had the following,

which jumps fromthe canvas while retaining 

the original abstract aspects.

It's a collaboration.

This second sequence shows a common technique I had used for years. As anyone knows who knows me I am fascinated by faces and have been drawing them since  pre-school. 20 years ago or so, I began using Photoshop to break faces into contrasting patterms of light and dark.

I used to draw high contrast faces in high school. I loved the way the contrast would amplify the features while simultantously simplifying the elements. You can see how this works here on the cover of Meet the Beatles.

So I took my original painting and fed it to Gemini. [Gemini is a family of multimodal large language models (LLMs) developed by Google AI that can understand and operate across different types of information, including text, code, audio, image, and video.]  My prompt was: Paint this woman in the style of Rembrandt but in this color scheme. Have the face itself emulate that of Marlene Dietrich.

This was the result.

My next prompt was:

Change the background to a Dutch interior 

from the time period of Vermeer.

My final prompt was:

Change the background to be a starry night 

in the style of Van Gogh.

 Each image took less than two minutes to generate.

To try it yourself, go to https://gemini.google.com/

Press the little + symbol to upload an image and 

write an instruction prompt for Gemini to work on.