12.01 Decision Making and Responsibility Allocation

Aim of the project

This project will investigate the relation between decision making in groups and the attribution of responsibility to members of the group. The aim is to determine how collective decision making and responsibility allocation can be aligned, to come to terms with the role of computer systems in teams, and to foster better understanding and perception of responsibility among the agents.

Theoretical background

When it comes to making collective decisions, an important component of a group’s functioning is in the information exchange that obtains between the group members. The formation of a shared opinion, or at least of an opinion that all group members are willing to subscribe to, is often required for successful collective decision making. The recognition of a commonly shared belief by group members is also a first step in sharing a responsibility for the actions that were taken on the basis of the beliefs subscribed to, certainly if these actions are themselves epistemic in nature, e.g., when scientists collectively pronounce on a matter of general interest.

Also when the group’s actions cannot be seen as based on a shared belief, the exchange of information within the group is a crucial factor. The specifics of the exchange might prevent information to arrive at the relevant agents, and this may lead to a dramatic reduction in the quality of group decisions, with all its associated problems of responsibility allocation and blame, as in “how could I have known?”, “you should have told me”, and so on. Accordingly, to allocate responsibility it is important to get clear on the organization of the information exchange within a group. And when we hope to improve the sense of responsibility within a group, much can be gained by organizing the information exchange appropriately.

Research design

First and foremost, the project involves studying the systematic dependency between the communication structure within a group, e.g., its network properties, and failures of information exchange, e.g., cases of polarization, shared information bias, and the like. This part of the research is carried out by means of philosophical, mathematical, and computational methods, drawing on network analysis in sociology, recent advances in philosophy of science that are concerned with how scientific communities function, and methods from argumentation theory. Additionally, the research will bring in empirical insights from sociology and social psychology, both of which offer important constraints on how agents might interact and exchange information. The project is designed to be carried out in conjunction with project 6.4 on the value of diversity. Another cross-over is with project 8.6 on interventions in social networks. Potential applications of the insights are ready to hand. Another live domain of application is in the structured deliberations among judges.