Building an Artificial Conscience: Prospects for Morally Autonomous Artificial Intelligence

Abstract

Discussions of ethics and Artificial Intelligence (AI) usually revolve around the ethical implications of the use of AI in multiple domains, ranging from whether machine learning trained algorithms may encode discriminatory standards for face recognition, to discussions of the implications of using AI as a substitute for human intelligence in warfare. In this chapter, I will focus on one particular strand of ethics and AI that is often neglected: whether we can use the methods of AI to build or train a system which can reason about moral issues and act on them. Here, I discuss (1) what an “artificial conscience” consists of and what it would do, (2) why we collectively should build one soon given the increasing use of AI in multiple areas, (3) how we might build one in both architecture and content, and (4) concerns about building an artificial conscience and my rejoinders. Given the increasing importance of artificially intelligent semi- or fully autonomous systems and platforms for contemporary warfare, I conclude that building an artificial conscience is not only possible but also morally required if our autonomous teammates are to collaborate fully with human soldiers on the battlefield.

Citation

Casebeer, W.D. (2020), "Building an Artificial Conscience: Prospects for Morally Autonomous Artificial Intelligence", Masakowski, Y.R. (Ed.) Artificial Intelligence and Global Security, Emerald Publishing Limited, Leeds, pp. 81-94. https://doi.org/10.1108/978-1-78973-811-720201005