With his transmedial series "MERZmory", Merzmensch explores human memories, and also their traces in the Latent Spaces of Artificial Intelligence.

In one of the manifold artistic approaches, he trained various AI-driven visual models on his photographs. Afterward, the archeological (etymological?) research begins.

In Faces, he trained Neural Model StyleGAN2 on his photo dataset "MERZphoto" (photographies he did in the last 10 years). Previously this particular Neural Model was pre-trained on faces (FFHQ), that's why initial generations still depict facial features, even if they disappear with iterative training. Exactly this transgression between the AIs initial knowledge and Merzmensch's own memories is in focus of the artists' newest installment. This bridge between dream and awakening, between fiction and reality. What is reality?

In his essay, you can follow his approach, as artistry and technology merged into one vision.
https://towardsdatascience.com/visual-ai-trained-on-my-memories-a54c55967c28?sk=3e37e2716ccee2101ba07025e5bcfd60

This is human-machine collaboration.