· Contributors · Organizations ·
Ethical Artificial Intelligence
Cognitive Engineering & Decision Making
DescriptionWith the adoption of autonomous decision making algorithms in various domains, advanced systems increasing encounter scenarios with ethical dilemmas. In these situations, the current state of the art relies on assumptions about the world and encodes the course of action for a given scenario, i.e. implicit, embedded ethics. This approach is fragile and do not account for discrepancies between design and operations, thus making the system’s trustworthiness questionable. The U.S. Department of Defense recently adopted “5 key ethical principles of AI”, encompassing five major areas: Responsible, Equitable, Traceable, Reliable and Governable representing an important step forward. However, the operationalization of ethical principles to practice in the field is currently dependent on amorphous human judgment. This panel will discuss the need and challenges for generalizable frameworks for artificial intelligence agents to reason about ethical decisions in a wide variety of contexts.