Responsibility in a Multi-Value Strategic Setting

TL;DR

This paper introduces a model for responsibility attribution in multi-agent, multi-value settings, enabling agents to anticipate responsibility and select strategies aligned with their values, with non-dominated regret-minimising strategies effectively reducing responsibility.

Contribution

It presents a novel model for responsibility in multi-agent multi-value environments and extends it to responsibility anticipation for strategic decision-making.

Findings

Non-dominated regret-minimising strategies reduce expected responsibility

The model applies to multi-agent, multi-value scenarios

Responsibility considerations influence strategic choices

Abstract

Responsibility is a key notion in multi-agent systems and in creating safe, reliable and ethical AI. However, most previous work on responsibility has only considered responsibility for single outcomes. In this paper we present a model for responsibility attribution in a multi-agent, multi-value setting. We also expand our model to cover responsibility anticipation, demonstrating how considerations of responsibility can help an agent to select strategies that are in line with its values. In particular we show that non-dominated regret-minimising strategies reliably minimise an agent's expected degree of responsibility.