What are some problems in philosophy that are related to AI safety?

From Stampy's Wiki