Collective-Intelligence and Decision-Making

Abstract

Decision-making, while performed by humans, is also expected to be found in most (artificial) intelligent systems. Usually, the cognitive research assumption is that the individual is the correct unit for the analysis of (human) intelligence. Yet, the multi-agent assumption is that of a society of interacting individuals (agency) that collectively supersedes individual capabilities. Therefore, the entire society of agents is, itself, an additional locus for the analysis of this collective- intelligence. In this paper we propose models that explore the agent-agency mutual influence from the decision-making perspective. We outline three case study scenarios for the models’ experimental evaluation: i) large-scale disasters, ii) electricity markets, and iii) Web-empowered knowledge and social connectivity. The scenario-driven evaluations are being used to guide our research efforts.