From Pluralistic Normative Principles to Autonomous-Agent Rules

Abstract With recent advancements in systems engineering and artificial intelligence, autonomous agents are increasingly being called upon to execute tasks that have normative relevance. These directly—and potentially adversely—affect human well-being demand of the agent a degree normative-sensitivity -compliance. Such norms principles typically social, legal, ethical, empathetic, or cultural (‘SLEEC’) nature. Whereas this type often framed abstract, as high-level principles, addressing concerns concrete applications requires refinement into explicitly formulated practical rules. This paper develops process for deriving specification rules from set norms, thereby bridging gap between operational practice. enables select most normatively favourable action intended context premised on range underlying relevant principles. In translation reduction SLEEC rules, we present an iterative uncovers addresses concerns, identifies resolves conflicts, generates both preliminary complex normatively-relevant guiding development better positioning them SLEEC-sensitive SLEEC-compliant.