Acting for the Right Reasons: Creating Reason-Sensitive Artificial Moral Agents
This addresses the challenge of creating reason-sensitive artificial moral agents, which is an incremental advancement in AI ethics.
The paper tackles the problem of enabling reinforcement learning agents to make moral decisions based on normative reasons, resulting in an architecture that restricts agents to morally justified actions and includes an algorithm for iterative improvement through feedback.
We propose an extension of the reinforcement learning architecture that enables moral decision-making of reinforcement learning agents based on normative reasons. Central to this approach is a reason-based shield generator yielding a moral shield that binds the agent to actions that conform with recognized normative reasons so that our overall architecture restricts the agent to actions that are (internally) morally justified. In addition, we describe an algorithm that allows to iteratively improve the reason-based shield generator through case-based feedback from a moral judge.