About the project
In this project, we consider the preferences of various stakeholders over a set of possible outcomes. Because stakeholders might be in disagreement over their preferences on the outcomes, we want to establish a notion of a middle ground, i.e., some set of preference statements that all might agree with. This middle ground is thus not a compromise, nor does it give a complete specification of preferences over all options.
We formulate this notion in mathematical terms and analyse it for the specific case of lexicographic preferences. That is, outcomes are specified through features or variables and stakeholders' preferences are assumed to be some (unknown) lexicographic orders over such outcomes wrt the variables. We exemplify this setting on the Moral Machine Experiment, where users (stakeholders) are asked to provide decisions in variations of the classic trolly problem: deciding which out of two groups of individuals to save while the other group dies.
The main goals are:
- A mathematical formulation of a middle ground.
- An analysis of its properties: Existence, uniqueness, etc.
- Identifying the complexity and algorithms for finding a middle ground.
Project team
-
Associate Professor Ana Ozaki, University of Oslo, Department of Informatics
-
Postdoctoral Fellow Anne-Marie George, University of Oslo, Department of Informatics