Should We Become Emotional With AI?

Abstract

Affective computing algorithms are becoming increasingly prevalent in devices and services with which we interact daily. Despite positive intentions behind the development of this technology, its impact on our society is still unclear. This paper describes four art performances using an affective algorithm designed to infer emotions based on facial expressions. These performances propose alternative scenarios which imagine possible entanglements with this technology. One performance invites people to consider how their bodies are seen through the lens of an affective algorithm. Another performance displays a dystopian scenario in which this algorithm is deployed as an instrument of the law. The third artwork invites participants to reclaim their agency when confronted with such algorithms. The fourth artwork considers a utopian scenario in which a symbiotic relationship between humans and emotional AI yields higher capacities for both. The purpose of these artworks is to encourage a critical discussion regarding the possible futures of our engagement with affective algorithms.