Attitude authority in the age of AI
摘要
Might artificial intelligence (AI) be highly reliable, now or in the conceivable future, at detecting your current attitudes, such as your beliefs and desires? If so, should people listen to you, or to AI, when trying to know your attitudes? The latter possibility threatens our long-standing practices of regarding each person as the ultimate authority about her attitudes. Disquieting as this threat may seem, I argue that it will not come to pass, at least when it comes to those attitudes that you determine for yourself through your judgements about what is true, desirable, or worthy of pursuit. This is because AI cannot determine your attitudes for you in this way, and because this capacity for attitudinal self-determination grounds your authority over these attitudes.