How does artificial intelligence work? Is it smarter than us? Is it going to take our jobs? What data is it using? The answer to all of these questions is no one knows. Big tech companies compile different AI systems' data, and no one can see their inner workings. AI systems can't think of new things, but they can predict the most likely outcomes based on the data it was trained on.

I wanted to find a way to see inside an AI model and create a window so we could understand some facet of how it generates material.  For this work, I took photographs of various fruits and fed those as reference images for an AI to generate another image. I created a one-word prompt, simply the type of fruit given to tell AI what to generate.

Said simply: when I gave it a photo of an apple, I told it to make an apple. I expected it to provide me with the same image. However, it created an image with a slight variation from the original. I then used the new photo as the reference image and told it to make an apple again.

Then again, and again, and so on. This process created 250 to 500 recursive iterations of different fruits. The results are provided as time-lapse videos served on dinner plates.

This is Farm to AI-Table. This work is a window into the AI system's thinking. Through the window, you can see a need for humans to monitor this technology and keep it on track.

It can quickly go wrong, but at the same time, the window provides a black mirror of who AI thinks we are.