Researchers develop FULL logic framework to formalize Kantian ethics for building safer, more principled AI moral agents

New multi-sorted quantified modal logic called Formula of the Universal Law Logic (FULL) formalizes Kant's categorical imperative for machine ethics

Addresses limitations of current approaches by considering agent intentions and purposes, not just action axioms like 'do not harm'

Framework formalizes concepts of causality and agency to move beyond simple enumeration of human moral intuitions

Tested on three Kantian ethics cases to demonstrate reasoning capability for evaluating moral actions