The Most Likely Image is a project investigating the creative possibilities of Artificial Intelligence. 

AI image generators respond to a written instruction with a picture that has the highest probabilty to correspond to the prompt. For this reason, the ‘statistical hallucinations’ created by the software often excel in mediocracy. Also, all the usual racial, cultural and gender biases apply. One challenge is to overcome or exploit these inherent flaws. 

In four series I explore a realm of ‘meaningful incongruities’, using abstract prompts to challenge the AI to draw something surprising and unplanned.

But what is going on here? Already in the early 80s, Baudrillard claimed that our current society has replaced all reality and meaning with symbols and signs, and that human experience is a simulation of reality. We all live in a kind of virtual reality, mediated by language and images. The advent of AI seems to bring this trend to quite another level, by outsourcing the creation of words and images to autonomous machines.

A book publication, also titled The Most Likely Image, is part of this project.