THIS ARTWORK HAS BEEN EVALUATED BY THE AI JUDGE. CLICK HERE TO SEE THE REPORT.

Functional clothing for a not-so-fictional future

An attempt to guide AI to truly creative design using the example of fashion design.

In philosophy, we are still a long way from finally defining the origins of our thoughts and ideas.

Language makes us social beings. The idea of communicating something specific arises from our thoughts.

The source of the idea does not only consist of what we have learned and socialization. Just as AI makes use of an unbelievable data set resolved into noise, our subjective reality is formed from the filter processes available to us, such as seeing, hearing, understanding, etc., from a pool of infinite information. From noise.

Just as the Karplus-Strong algorithm is able to produce realistic sounds from white noise, fed into feedback loops and then filtered using an equalizer, realistic sounds of instruments or even animal and human voices, the visual worlds produced by AI are also created from noise. Specially filtered by the input of the user’s more or less creative subjective idea.

Since the artist cannot be able to understand the true reasons for his ideas in the conventional creative process, I have looked for a way to instruct and guide the image-generating AI with art-specific instructions and vocabulary, which underpins the main theme (functional clothing for a not-so-fictional future / fashion design) with academic terms, which I in turn extract from an AI-fed pool consisting of noise of art-related technical terms by random generator.

Surprisingly, highly precise results crystallize out of the noise of information and possibilities that I would not have been able to formulate myself. My part was to transfer the results to a specific design genre, in this case fashion design. The method developed can be applied to all design disciplines. As a designer, I detach myself as a subject from the creative process, marvel at the “creative” composition of the results and transfer them into a design process that I can use.