A brief sketch of interconnections among these disciplines might look as follows: more and more phenomena are becoming global in scope, from climate change to the behavior of economic markets; many if not most of these phenomena have the characteristics of social dilemmas, which implies that improving problems associated with them requires a high degree of coordination; however, most decentralized coordination problems entail a degree of complexity that requires cooperation on a very large scale, which often defeats even skillful governance. Computational contraints, human social behavior, and the limitations of current knowledge all circumscribe the capability of governance to deal with complex problems effectively.
The following paper, written for a class in computational learning theory in spring 2005, explores some of the formal connections among artificial intelligence and social science disciplines, by way of game theory. Download paper on adaptive learning (235KB)
Many more relevant publications can be found at the below link for the Collective Dynamics Group, led by Duncan Watts, a professor of sociology at Columbia University. I am indebted to Prof. Watts for his guidance, and would like to credit his book Six Degrees for crystallizing the present form of my ideas.
See also my past research interests and publications, in particular the page on social dilemmas.