Artist uses AI for design aesthetics

An artist started using artificial intelligence to create new design aesthetics for everyday household and fashion items.

December 1, 2020 - Moscow, RU — Artist Igor Dunaev launched a new collection of items designed using artificial intelligence. He uses his personal computer to generate seemingly random images from data sets. These sets are systematically chosen to bleed several images into each other, yielding unexpected results in the process. The art rendered by the computer is then given final touches by the artist and applied to everyday products.

The benefits of such a process include the speed at which it is done. Modern graphics cards render images increasingly fast - much faster than the average artist is capable of. The other benefit is the mitigation of human error due to technical skill, mood or physical disability, albeit a new set of skills is required. This procedure also reduces the cost of creating art and designs due to the low cost of energy the modern computer uses to run. This method of design in many ways automates the process.

Due to the chaotic nature of artificial intelligence, the computer needs a human direction otherwise the images make no sense from a human perspective. The human brain perceives the world in a very subjective manner whilst the computer sees it only as data. In this artistic process, the computer and the artist complete each other and merge their frames of reference.

Items created include dresses, skirts, phone cases, pillows, curtains, backpacks, clocks and more. Part of the collection can be seen here.

The technology making this possible is a class of machine learning known as generative adversarial networks. The codes are available as an open source at a GitHub repository. It is written in the Python3 programming language.The original authors of the architecture are cited here.

###