Artificial intelligence (AI) can beat championship-level humans at chess or go, but it cannot decide that it wants to play chess or go. Understanding volition as a key component of "intelligence," something which our current models are incapable of, means that our discussions of "ethical AI" are not really about AI; they are about the choices made by AI researchers and developers. Ethics is for beings who can make choices.
Humans have decided to commit genocide on many occasions, something an artificial general intelligence would likely consider illogical. But an "intelligent" computer that incorporates the human ability to behave transgressively might decide differently.
View Full Article