Duncan Jones on the Moon
Duncan Jones discusses his movie Moon.
Published on July 13, 2009 by Matthew Hutson in Psyched!
...
MH: You studied philosophy in college?
DJ: I did, I was at college in Ohio doing philosophy and then I went on to graduate school at Vanderbilt in Nashville. It was general moral philosophy, but also I was trying to make some arguments for how you might possibly apply ethics to sentient machines when we get to that stage. It was a little premature.
MH: The title of your undergrad thesis was "How to Kill Your Computer Friend: An Investigation of the Mind/Body Problem and How It Relates to the Hypothetical Creation of a Thinking Machine."
DJ: That was pretentious bull****, really. The paper itself was pretty good. It had some good ideas in it but it had a stupid title, sorry. When you're in college everything seems much more important than it really is.
MH: When you mix the mind/body problem with thinking machines something's gotta come of that.
DJ: Absolutely. I was looking into Daniel Dennett's work with the Cog project at MIT trying to create machines that were self-aware and just expanding on that from a moral philosophy point of view, where would we stand if and when Cog ever got to the point where it was self-aware and what are our duties to it? Can we turn it off? Just things like that. Daniel Dennett describes "functional equivalence:" If something always as far as you're able to sense it is doing what you expect a sentient being to do, even if it's completely artificial, it's your duty to treat it as a sentient thing. That's it in a nutshell. I didn't need to write a paper, I could have just told you in a couple sentences.
MH: It seemed like that also played a role in the plot of the movie.
DJ: In a way. Gerty [the computer in Moon, voiced by Kevin Spacey] was my antithesis to Cog. The idea was, Gerty isn't sentient. Gerty is actually very, very simple in some ways. He has one through line which is I'm going to make sure that Sam is safe and looked after and returns home at the end of three years. That's his job, and it starts when Sam wakes up and it ends when Sam is in the return vehicle going home. After having three years of not having anyone around. as far as Sam's concerned, Gerty is his best buddy, someone he can rely on and treat like a human being. For Sam 2, he's just a machine and for the audience he's possibly Hal 9000. Everyone brings their own baggage to Gerty. But Gerty's actually very simple. So I think that's quite interesting from a psychological point of view, that Gerty is the sum of what people bring to him.
MH: I think that goes with the general trend of anthropomorphizing things. People do that with their iPods.
DJ: Absolutely. That's absolutely the case. I do it with my laptop. You hate me, you hate me!
MH: To me the limitations of Gerty came to the fore when I realized how easily he could be deceived. Sam said, "I'm just going outside to check on the thing, I swear." And Gerty's like, "Okay I believe you." In any human, the bull**** detector would go off.
DJ: I think with Gerty, as infinetesimal as the chances were that Sam is telling the truth, he had to give him the benefit of the doubt. Because Sam is the guy that he is there to look after. And if there were micrometeorites or if there were some kind of damage to the base, Sam should go outside to check it. So even if it's incredibly unlikely, he has to let him do it.
MH: Do you think much about what aspects of social interaction are important to simulate in artificial intelligence?
DJ: It's important that a computer system which is going to be able to interact with human beings be more than just like Eliza, you know that old computer program where you would type in things and it would give you crappy answers back. It would not pass the Turing test. Basically if you want to have a computer system that could pass the Turing test, it as a machine is going to have to be able to self-reference and use its own experience and the sense data that it's taking in to basically create its own understanding of the world and use that as a reference point for all new sense data that's coming in to it. That to me is really interesting. My old girlfriend a long time ago who I went to Vanderbilt with, she ended up on a psychology track at graduate school and one of the things I was always asking her is how much evidence is there that a human being is a purely physical system and that what we consider the mind is actually just a manifestation of a physical system, of a self-referencing system where you have sense data coming in and some kind of system in the brain which is referencing that and comparing it to experience and a little feedback loop. It seemed that there was at least some evidence that that might be the way things work. I'm certainly not a dualist, I believe that everything is a physical system, so to me that makes sense and the only way you are going to replicate a sentient being in a machine is to create that same system. I went off on a really big tangent there. To answer your question, yes.