Ethical Properties of Minds

From AGIRI.org

Nearly everyone involved in building AGI's agrees that it is important for them to have agreeable ethical properties. But there is much less agreement on which ethical properties are agreeable, and on whether it's practical or even possible to formalize ethical properties in a way that's useful for AGI development.

A reading list with many entries relevant to this topic is here:

http://www.singinst.org/reading/artificialgeneralintelligence/

[edit]