Transfiguration: Art and Machine Intelligence

Self initiated creative research · Begun in 2019 and expanded over time

Long before the current generative AI wave, this project began as a personal investigation into how machines interpret original drawing, color, and form. The page brings the early 2019 experiments together with later animation, drawing, and videography studies to show how the research evolved.

My contribution

Artist and Researcher · Original Drawing · Dataset Creation · Model Training · Animation · Visual Direction

I created the source drawings and training data, trained and tested the systems, and developed the visual experiments. Rather than presenting a single finished artwork, the project records an ongoing creative inquiry into collaboration between the artist and the machine.

Project archive


 

————————————————————
These experiments below involve animation and style transfer. That as well as the copywriting were created in 2019.
 
 

 
 
 
 

In my day to day I am an Art Director in charge of the design of many Public Art projects in the fields of Media Arts and Architecture. Since a child I’ve always loved to paint. While working on my last painting I started thinking “What would happen if I could teach an Intelligent Machine how to paint like I paint?”  That set me on a journey of discovery….

It is crucial for creatives to understand the concept of “training” and how through it we can create intelligent machines. Artist and designers of the future will train machines into magical  instruments of art. After learning how Machine Learning algorithms can be applied to art (GAN, Deep Dream) Style Transfer seemed the most versatile for creating art and for a direct collaboration between Human and A.I. 

For example I started the journey by trying to understand how basic forms and colors are understood by the trained machine. I did this by creating very simple sketches and using them as the departure point.

 

There is a fascinating moment when you start seeing the results of your art training and you can start understanding the complexity behind the Machine Learning. As if by magic new patterns start emerging. Interesting moments of beauty. Your initial artwork has transfigured into a new world of possibilities. This is not just an effect but a reinterpretation of art by an intelligent machine.

 

 

After various training sessions I started seeing an organic expression emerge from the A.I. It seems as machines get more intelligent they start resembling the patterns and language found in nature. Artificial complexity is itself an organic expression. 

 

Finding these organic tendencies inside these Machine Intelligence algorithms to me is incredibly fascinating. It leads me down imaginary pathways to visions of a future where machine induced aesthetic border nature. In other words these algorithms helped me imagine a future were machines would want to design like nature. 

 


It is an incredible moment for artist and designers when you can teach an intelligent machine to reproduce new artworks as if they are children that resemble the “DNA” of the parent artwork. This makes you feel as if  the possibilities of art are truly infinite. When you start experimenting with new technologies it serves as a catalyst for new ideas and inspirations.  

 

 


With these tools there is a potential to shift the focus from one work of art and instead it now becomes a focus on the essence of the artwork, and translating that essence into something an intelligent machine can understand.

 


 

 

These new workflows of the future will rewire our perception of how art is made and what is the true essence of art.  There is now, new fertile grounds for creativity  and exploration  and it us up to us designers and artist to seek ways to use these new incredible tools and contribute to the dialog.  To understand what it means to be an artist and designer alive in this new decade.. And not be afraid of the changes that are coming. 

 

 


 

.

.

.

.

.

.

.

.

.

.