Issue No.04 - July/August (2006 vol.21)
Bruce M. McLaren , Carnegie Mellon University
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/MIS.2006.67
Computational models of ethical reasoning are in their infancy in the field of artificial intelligence. Ethical reasoning is a particularly challenging area of human behavior for AI scientists and engineers because of its reliance on abstract principles, philosophical theories not easily rendered computational, and deep-seated, even religious, beliefs. A further issue is this endeavor's ethical dimension: Is it even appropriate for scientists to try to imbue computers with ethical-reasoning powers? A look at attempts to build computational models of ethical reasoning illustrates this task's challenges. In particular, the Truth-Teller and SIROCCO programs incorporate AI computational models of ethical reasoning, both of which model the ethical approach known as casuistry. Truth-Teller compares pairs of truth-telling cases; SIROCCO retrieves relevant past cases and principles when presented with a new ethical dilemma. The computational model underlying Truth-Teller could serve as the basis for an intelligent tutor for ethics.This article is part of a special issue on Machine Ethics.
machine ethics, casuistry, artificial intelligence, case-based reasoning
Bruce M. McLaren, "Computational Models of Ethical Reasoning: Challenges, Initial Steps, and Future Directions", IEEE Intelligent Systems, vol.21, no. 4, pp. 29-37, July/August 2006, doi:10.1109/MIS.2006.67