Towards an experiential ethics of AI and robots: A review of empirical research on human encounters

摘要

The past few years have seen a profound re-acceleration of interest in artificial intelligence (AI) and robotics, accompanied by intensifying debates about ethical regulation. Yet, less attention has been paid to how people experience AI and robots in practice. This paper explores the potential of an experiential approach to AI and robot ethics. Specifically, we review empirical studies on human experiences with AI and robots and argue for the value of assembling and analysing findings from studies that inquire into the everyday encounters with AI and robots. Following a hybrid approach that combines systematic review with narrative social inquiry, we identify six key dimensions of human experiences with AI and robots : appreciation of imperfection, formation of affective relationships, discomfort with lack of transparency, addition of invisible work, shifting responsibilities, and readiness to trade off privacy for other benefits. By placing these dimensions into dialogue with ethical AI governance, pragmatist philosophy and Science and Technology Studies, we argue for an experiential approach to ethics, i.e. an approach that grounds ethical reflection in lived encounters, where abstract principles often take new, context-specific meanings. Thereby, we invite attentiveness to ethical concerns that might otherwise become sidelined in extant AI and robotics policy frameworks. • Develops an experiential ethics framework for AI and robot governance • Reviews human encounters as central to ethical considerations • Highlights preferences for imperfection and privacy trade-offs to refine principles • Identifies affective entanglements and hidden work as added ethical concerns • Calls for integrating lived experiences into policy