In this project, I explored an idea of co-creation by an Artist (me) and an Artificial Intelligence (AI) algorithm. I wanted to reduce the artist’s handmark in the final artwork to see if the AI has its own creativity or if it’s just a tool in the artist’s hands. Looking ahead, I will say that I have not managed to get rid of the artist’s hand. The algorithm was not autonomous, and I had to work hard to get what you see below.
Here’s how it all happened. I found some of my analog photos and decided to feed them to AI to see what I would get back. The image was not significantly different from the original. The algorithm got the landscape and replicated it accordingly. So I tried to add something unusual to the image to confuse the algorithm. I drew some sketches over the image. Again, nothing came of it. AI seemed to remove everything it did not comprehend and restore the appearance to the original. But I kept adding more hand-drawn sketches, and eventually, the algorithm generated something interesting. We made our moves turn by turn and repeated it 10-15 times for each image. Below is the result of our game as separate slides of steps.
While creating this project, I often realized how similarly the algorithm and the human recognize the image. The difference was in the amount of experience. The algorithm was trained on thousands of photos, and I was exercised by the infinite number of visuals I have seen throughout my life.