Choice Machines, Causality, and Cooperation
3:00-3:30 pm, October 3
Can a purely mechanical entity make genuine choices in pursuit of its goals? Can AIs do so, and can we ourselves do so if we are mechanical? Can choices occur even in a universe that's deterministic, where an initial state, and some causal laws, already specify everything that will ever happen? And can a choice-making machine somehow find reason to cooperate with others, to help achieve their goals, even if doing so causes net harm to its own goals?
I argue that these matters intertwine in deep and sometimes surprising ways. And although some of the ideas in these questions sound contradictory, I argue that the answer to each question is yes.
Biographies: Gary Drescher
Gary Drescher received his doctorate at the MIT Artificial Intelligence Laboratory in 1989. His book Made-Up Minds: A Constructivist Approach to Artificial Intelligence (1991) proposes an empirical-learning and concept-inventing mechanism inspired in part by Piaget's observations of the development of intelligent activity in infants.
In 2001-2002, Drescher was a Visiting Fellow at the Center for Cognitive Studies at Tufts University. His 2006 book, Good and Real: Demystifying Paradoxes from Physics to Ethics, explores how a purely mechanical, deterministic universe might implement conscious, choice-making, ethics-bearing beings such as ourselves.
