The problem here, I think, is that although humans can be studied without knowing the structure of their brains, (psychology, sociology, etc...) machines of today cannot.
The reason for that is that humans are humans and barely differ from each other biologically (the structure of the brain), but ANNs are very diverse and their behavior changes depending on their "biology," so machine psychology or machine sociology cannot be built on stable grounds
gowld|8 years ago
why?
Or do you mean that study of some humans is generalizable to most/all humans, whereas machines lack this consistency and generalizability?
Anyway, why would this concern prevent "feature detection" tests that apply to all machines?