Which controls which? Sense of agency when humans and semi-automated systems co-operate
Project Outline
When a semi-automated system, including artificial intelligence (AI) or a mechanical system, co-operates with a human and commits a socially undesirable action such as an incident or accident, which idea should be adopted: should the human user be responsible as the subject for the action, or should the machine, system, or manufacturer be responsible as the subject? What are the scientific definitions of action and subject? What does it mean to be a social agent or subject who has free will to act and hold responsibility for his/her actions? We propose a scientific definition and answer for these questions from the perspective of psychology and neuroscience.
In particular, we identify that there is an illusion in which human users misattribute their actions as their own instead of belonging to other agents, which can lead to unnecessary feelings of responsibility for an action. Through the brain science based on the understanding of this illusion and the development of its application in human-machine interactions, we propose the idea that most of the human sensation of "agency" (i.e., that oneself caused an action) is based on postdicted subjective sensation or illusion. In this sense, humans cannot be fully autonomous.
Investigators
Takako Yoshida |
School of Engineering, Tokyo Institute of Technology |
Associate Professor |
Yuji Suzuki |
Hang & Paragliding Promotion Committee, Japan Hang & Paragliding Federation (JHF) |
professional hang glider pilot |
Participating and Cooperating Organizations
Tokyo Institute of Technology |
Japan Hang & Paragliding Federation (JHF) |