Emotional machines: composing for unstable media

thorolf thuestad, university of bergen, faculty of fine art, music and design

This research explores the idea of anthropomorphism as an approach for generating artworks, with a specific focus on machine empathy.

This project is particularly concerned with the tension between anthropomorphism and
mechanomorphism: the way we see human characteristics in our machines, and conversely in
how we expect our surroundings, including fellow humans, to act in a deterministic or
mechanistic way.

I postulate that anthropomorphism is fundamental to how humans interpret and interact with the world, providing a resource for artistic expression. As such, the project relates to social, political, economic and psychological sciences with the rational choice theories, microeconomic models, and belief in people as rational actors found there.

By developing, categorizing and implementing methods for tapping into the irrational in people in order to evoke art experiences, this project intends to explore areas where the deterministic ideas of our current mechanomorphistic society fail. These methods will then be investigated through the construction of sound producing robotic and kinetic  installations, musical compositions and other time based art-forms. These works will  attempt to evoke what may be considered an irrational human trait; machine anthropomorphism. It will then investigate how a machine/computer interpretation of human irrationality may manifest itself by exploring how generative systems might be used to mimic (inconsistent/irrational) human decision-making processes to create time-based inter or transmedial compositions with activities generated and shaped depending on the different «moods» of the generative control system. The kinetic sculptures will be made environment sensitive allowing the actions of each sculpture to be influenced by the presence of the others, or by the presence of audience or human performers. Finally, the project will create hybrid man-machine transmedial time based performances focusing on interaction between human and machine performers.