I would argue that there's a point when it stops becoming strictly a machine and you have to start treating it like a human.

When you can't tell the difference anymore.

It's the Turing Test.

Dogs can't pass a Turing Test. Does that make vivisection okay?

Can the robot feel pain from over-work, humiliation from being subjugated to someone else's whims? Does it want to do anything other than work? If not, what's the problem? The human mind is a product of billions of years of evolution. We've evolved to survive and procreate. Humans have always been social animals and because of that we have complex social emotions - which essentially give us our personalities.

Read Blindsight (found here for free from author http://www.rifters.com/real/Blindsight.htm ), Spoiler: it is a science fiction book about non-sentient (biological) intelligence. It posits that such intelligence could be greater than sentient intelligence. Well, I am spoiling it already, but it is great (and disturbing, I probably will never read it again). He was a biologist, but left the field.

It was a Hugo nominee.

JM

The webcomic Freefall has been diving into this headfirst for the past half dozen years.

I should really make a post about two technoreligions geeks adhere to.