Understanding Partner Selection for Cooperation: Towards Supportive Hybrid Intelligence for Joint Undertakings

Human cooperation has long been recognized as a cornerstone of social success and well-being, serving essential roles from overcoming challenges to fostering communal resilience. For several decades, research across disciplines has examined the mechanisms and benefits of cooperation. Currently, this interest is becoming increasingly relevant in computational fields. As intelligent systems start to be envisioned to act in various roles, including decision-support systems and social catalysts, there is a shift toward designing intelligent systems that not only act in synergy with other humans but also support human-human interactions. For instance, such systems can potentially help us improve the way we interact with each other, play a key role in facilitating and maintaining the quality of joint tasks, and potentially support us by helping us select suitable partners for joint endeavors. However, to develop intelligent systems that effectively serve as social catalysts or decision-making support systems in social settings, these systems must grasp and engage with the core processes that underlie human social interactions. This entails recognizing and interpreting both verbal and nonverbal behaviors and understanding how these cues relate to complex social phenomena, such as partner selection, comprehending human preferences in partner choice, and detecting underlying relational dynamics. The core proposition of this thesis is that, to design intelligent systems capable of fostering cooperation, it is first necessary to deepen our understanding of how people initiate cooperative behavior and choose partners for such tasks. However, there remains a significant gap in empirical research that designers could utilize as a foundational resource to understand how these elements contribute to cooperative initiation and unfold within human-to-human interactions. Therefore, this thesis aims to provide empirical contributions that offer insights and resources to guide the development of socially intelligent support systems.