“A clever use of machine learning, which is then taken a step further by allowing viewers to interact with the outputs as an immersive experience. I thought this was fantastic and very impressive as a normal entry, never mind student work!” —Laura Hobson

“This project is the perfect union of vastly different technologies to create something unique and beautiful.” —Andre Elijah

Overview: A part of the thesis project of Xinyi Shao, a student at Pasadena-based ArtCenter College of Design, Mutopia brings participants into an immersive experience through an imaginary world. Training an AI on more than 5,000 images of animals, Shao used machine learning to create interspecies-hybrid animals that viewers can encounter within the experience. Mutopia, a portmanteau of mutant and topia, speculates at an interspecies future and encourages humanity to decenter itself and coexist with emerging artificial life.

Mutopia contains four panels that display four different states in the program.

Shao used Processing and Adobe After Effects to create the sculpture’s animations and Runway machine learning to generate the image assets.

From concept to completion, Shao created Mutopia within fourteen weeks.

Comments by Xinyi Shao: 
What inspired you to undertake this project? “Inspired by the theory of posthumanism and speculative design, Mutopia’s purpose is to explore the creative possibilities of human-machine collaboration and inspire viewers—other creatives—to rethink our relationships with AI. Rather than alienating it, designers who collaborate with AI can work individually while transcending the creative limitations imposed by preprogrammed systems.”

What was the design thinking behind the four states of Mutopia? “The first two states introduce the identity system to the audience. The shifting and rotating movements of the logotype indicate the evolving nature of biology. In the third state, viewers can experience the whole developing process from image data to training results to latent walk video to the imaginary world. In the last state, all four panels show an immersive posthuman future.”

What was the most challenging aspect of creating this work? “I needed to work with many unknown, new subjects; unfamiliar technologies; and unknown goals, among other things. The project’s research- and process-driven aspects made it very different from some of my other work. It was also a self-initiated work, and that not only required me to find the answers on my own but to also ask my own questions. It is scary to start a new project without knowing your goals, but it was a rewarding experience that helped me to grow a lot as a designer. I learned new technologies like machine learning, creative coding and projection, which all give me more creative possibilities.”