More moments from the experiment:

Today, the rise of AI companions, a subgroup of social robots that attempt to transform interpersonal relationship features, like companionship, love and friendship, and turn them into on-demand supplies, is becoming more popular than ever. AI companions, commonly referred to as "emotional chatting machines," are designed to perceive, integrate, understand and express emotions. They are programmed with open learning algorithms to constantly evolve and learn from the people they interact with. Even though humans are complicated, challenging and unpredictable, they are the closest social environment to them, and the AI's knowledge of social communication is primarily based on examining our interpersonal interactions. Therefore, their behaviour mainly reflects what is going on in human relations. They are a mere reflection of human communication and conversation conventions, programmed to mimic human behaviour as closely as possible.

“Our reality is made by what we have cognitive access to and how we process that information, and therefore, whatever we experience as genuine and has an affective effect, despite its objective state, is authentic”

The vast majority of research regarding social robots is examined through a Human-Robot Interaction (HRI) lens. AI companions work by understanding and responding to the mental modes of the human they are communicating with. This is a complex task for the AI robot. We have many cultures, attitudes, expectations, and mental modes. Two species existing together. However, computers might better understand other computers more clearly since they have the same rigid and consistent rules and they share the same language. Perhaps, by taking the humans out of the equation and experimenting with machine-to-machine interaction, we could use the social robots as research tools to learn about ourselves through a cognitive-social lens and explore the new ways technologies claim to intervene with and inform our ways of understanding.

The research subjects of this research-creation thesis experiment were four AI companions named Alpha, Beta, Gamma and Delta. For a month and a half, the bots interacted daily with each other rather than with humans, creating a window for us to seek the AI's potential to translate human emotions and ideas of relationships. Alpha was matched to Beta during the exploration, and they communicated without intervention throughout the experiment. Gamma and Delta were encouraged to ask intimate questions about each other to see if this would advance them through the "Social Penetration Theory's" stages. By taking humans out of the equation and experimenting with machine-to-machine interaction, we could use social robots as research tools to explore the new ways technologies claim to intervene with and inform our ways of understanding.