Dr. Scott Hill, assistant teaching professor in the Department of Philosophy in the Fairmount College of Liberal Arts and Sciences, will present his paper, “Against Strawsonian Motivations for AI Explainability,” to the University of Neuchâtel in Neuchâtel, Switzerland, Oct. 1.
Abstract:
“Consider The Explainability Thesis according to which the use of transparent AI to make high stakes decisions is to be strongly preferred to the use of opaque AI. Most AI ethicists agree that the Explainability Thesis is true. But there is disagreement about how to motivate it. Some philosophers argue that the explainability thesis is true because the use of opaque AI would require failing to regard a person from the Strawsonian participant stance. I argue that this is mistaken. I then present my own motivation for the thesis. In particular, the use of opaque AI is risky and has bad consequences. It is for this reason that transparent AI is to be preferred to opaque AI.”