Chang Zhou

This PhD research proposal will investigate how proactive behaviors in social robots will foster collaboration and prosociality. In hybrid societies of humans and agents (especially embodied agents such as social robots), collaboration is the key ingredient for successful task execution. Humans and robots will work in teams, complementing each other, due to their different competencies. However, for robots to successfully be part of hybrid humans-AI teams, they need to exhibit some degree of pro-activity, and decide by themselves when and how to act, when and how to collaborate with the humans. For years, the main approach for creating proactive agents or robots has been rule-based approaches, which are used to ensure reliable and predictable behavior. Yet, this research will investigate new ways to implement proactivity in social agents, relying on data captured in collaborative tasks, and study its impact on the degree of collaboration and prosociality achieved within hybrid teams. The work will seek to place itself in health case scenarios, investigating its applicability for hybrid teams in that application domain. This research will contribute to the field of human-AI interaction and social artificial intelligence.