Synthesis and Properties of Optimally Value-Aligned Normative Systems

The value alignment problem is concerned with the design of systems that provably abide by our human values. One approach to this challenge through leverage prescriptive norms that, if carefully designed, are able steer a multiagent system away from harmful outcomes and towards more beneficial ones. In work, we first present general methodology for automated synthesis aligned normative systems, based on consequentialist view second part, provide analytical tools examine such namely Shapley individual compatibility several values under fixed set norms. We illustrate all contributions running example society agents where taxes collected redistributed according parametrised