In creating ‘A Detour, Invalid Light by Mistake’ I used Stable Diffusion and Stable Audio models as expressive tools to convey the personal emotions I was experiencing at the time. These AI models serve as one part of a multifaceted composition process, with each model’s input and output carefully pre- and post-processed to ensure alignment with the emotional and aesthetic vision of the work. The experimental music is composed with recordings I made from found objects and noInput mixer sounds.