Within the past decades, Social Autonomous Systems (also referred to as Social Robots, SR) have seen a dramatic improvement in their functions and possibilities for interactions. In contrast to earlier generations of artificial systems, whose programmed range of behaviours was predictable and limited, new generations of social robots can spontaneously develop and adapt their behaviours by learning from interactional experience with humans. For this reason, social robots have been increasingly employed in a wide array of contexts, from educational to healthcare and clinical.

This project (also referred to as SEEDS: Situated and Embodied Approaches to Human-Robots Interactions) endorses an embodied, situated perspective to investigate the interactional dynamics between humans and social autonomous systems. Embodiment has become an important concept in many areas of cognitive science; however, while many would agree that humans are embodied cognisers, there is much less agreement on what kind of humanly crafted artefacts could be considered as embodied. The project seeks to explore this issue by integrating theories and methods from cognitive science, philosophy and robotics (but also drawing on sociology, ethnomethodology and linguistics). An array of micro-analytical approaches will be adopted and adapted to study how humans engage and share meanings in interaction with social robots starting from real life situations.

Two main questions will be addressed over the project course with particular attention:

1. How should social robots be designed to be perceived as intersubjectively-meaningful partners by humans; and what kind of embodied and agentic features should they possess?
2. How can we improve the intelligibility of behaviours and gestures within a framework of cooperative actions involving humans and SR? How can we implement fine-grained interactional skills in SR, such as expressing and perceiving actions, or establishing and maintaining mutual behavioural attunement?

The major breakthrough to be expected from SEEDS is a comprehensive framework for embodied intersubjectivity applicable to the humanities and AI research, showing how artificial systems can be successfully integrated into human everyday life; but also helping us gaining unique understanding of so-far unknown aspects of human behavior.

---

Principal Investigator(s)