Endowing robots with ethical introspection capabilities
Abstract
This proposal aims to develop and implement a computational model of ethical introspection in robots. Building upon existing literature and computational models, this interdisciplinary project integrates insights from psychology, artificial intelligence, and ethics. One focus of the project is analyzing how the robot may take advantage of an ethical introspection to update its own space of ethical values. Then, the robot will start with a small set of basic ethical values. After the ethical introspection, it will improve, change, and even reset its ethical values. We expect that a robot aimed at ethical introspection will develop values that may differ from ours. In this case, we will analyze the situations in which the developed values will be compatible with ours. The other focus of the project is analyzing how a robot that introspects on its own ethical decisions may inspire teammates across domains in healthcare scenarios through self-explanation capabilities (transparent deliberation, inner speech, justification, etc.). In this case, we expect that a robot aimed at ethical introspection can influence caregivers and healthcare professionals. For this purpose, we will analyze the situations in which the robot s self-explanation capabilities will positively augment the ethical awareness of the team.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Feb 05, 2025
- Source ID
- FA86552417379
Entities
People
- Valeria Seidita
Organizations
- Air Force Office of Scientific Research
- United States Air Force
- University of Palermo