Moral Uncertainty and Artificial Intelligence
Artist: Janusz Jurek
Moral Uncertainty and Artificial Intelligence
Pamela Robinson presented ‘Moral Uncertainty and Artificial Intelligence’ to Effective Altruism UQ (University of Queensland), 2021.
Abstract: The decisions and actions of future artificially intelligent systems have the potential to affect us in significant ways. Because of this, we want to build ethical and value-aligned AI. But designers have plenty of grounds to be uncertain about which moral facts and theories are true, about the status of moral facts, and about how best to handle this in building AI. I discussed the morally risky design choices that designers of artificially intelligent systems face, and what makes them especially challenging.