Languages Magazine

Interview with “Ex Machina” Director Alex Garland

By Expectlabs @ExpectLabs

On March 18th, the Expect Labs team got to see a U.S. pre-screening of the film Ex Machina, which tells the story of a young programmer who is selected to evaluate the intelligence of a female AI named Ava. After the film, Expect Labs CEO Tim Tuttle interviewed Alex Garland, the film’s director. Check out the video and part of the transcript below [some spoilers].

TT

Is this film meant to be somewhat of a warning about the existential threat of AI?

AG

Definitely not. I don’t feel any existential threat at all.
There’s a lot of stuff in this film that draws parallels between Oppenheimer and nuclear power and that sort of thing, and I think that’s a reasonable analogy. I think there’s some latent potential danger in AI — that seems perfectly reasonable, just as there is in nuclear power — but I’m one of those green types…who actually kind of likes nuclear power. I think it’s a better way of producing electricity than coal power stations. I think there’s something similar in AI. It could be very dangerous, and obviously Stephen Hawking and Elon Musk are making it very clear how dangerous they think it potentially is. I personally think it’s a good thing. I welcome developments in it…
[Even in the film, when the Ava escapes], I found it fascinating and weirdly moving in a funny kind of way. I just absolutely love it. This film’s on the side of the AIs. All the suspicions are with the humans.

TT

Does that mean the sequel is going to be lot more like the movie Her than Terminator Two?

AG

There really doesn’t have to be a sequel. Two years is enough…

TT

Never say “never.”

AG

"Never."

TT

One of the things I love about this film is its treatment of the Turing test and how the Turing test isn’t necessarily the perfect measure for whether a machine is intelligent or not. When most computer scientists are writing programs to pass the Turing test, what they’re mostly trying to do is find ways to deceive a human into thinking that [the machine’s] intelligence is real. In the movie, Nathan takes that a step further…where he writes a program…to deceive another human to serve its own purposes. What is the true test of AI, do you think, and do you think the Turing test is valuable?

AG

…The Turing test is a brilliant test because it is so difficult to pass, but it’s primarily a test of the ability to pass the Turing test. It doesn’t say anything implicit about consciousness at all… That’s why there’s that conversation about chess computers, which act as if they want to win a game of chess but of course don’t know what chess is or that they’re computers or anything.
You set this problem about a machine, to tell you if it’s conscious, but actually you could equally make it about a person. You guys think I’m conscious not because I’m really presenting you with any good evidence that I am conscious. It’s more that you have a strong sense that you are conscious, and you believe that I am like you. If I had to prove it, or you had to feel that I had proved it, that would be quite difficult.
Actually, it turns out, if you are going to get philosophical about it…that you may actually be fooling yourself about your own sense of consciousness. Consciousness becomes very tricky… I’d say there are lots of questions that are posed in the film to which I really don’t have answers. That was actually part of the point, because there’s value in asking questions that you don’t know the answers to — that there are no answers to…
[Ava] stabs one guy, and locks another in a room, which by the way does not demonstrate a lack of empathy, just to be clear. I’m on her side. I think she was being reasonable under these circumstances. [As she leaves, there’s some Schubert playing in the living room] and it’s very beautiful, and she smiles, and I feel like that’s what the chess computer wouldn’t do. There’s no one in the room for her to be fooling with some fake “gaming the system” type Turing test like that stupid test that Eugene Goostman allegedly passed 8 months ago. There’s no gaming the thing; she’s on her own. It’s an indication of an internal mind state, so that’s as close as I could get to an answer. If you smile on your own in a room, I think you’re probably conscious.

TT

…You raise this interesting question which is “what is AI?” Is it Mary in the black-and-white room or is there something beyond that — something ephemeral that science hasn’t fully captured yet?

AG

We don’t really know what consciousness is.

TT

There’s a school of thought in AI which is the human brain really is nothing more than a collection of pattern matchers, a lot like the pattern matchers you see in the deep mind computer, and you just put enough of them together, then conceivably you could create an intelligence that’s comparable with that of a human brain. Then there’s the other school that says it’s completely different. Where do you fall in this?

AG

I’m not sympathetic on an instinctive level to the idea that says there is something special — which often turns out to be metaphysical — about humans which means machines will never be able to…have consciousness. That said, I would suspect that if and when a machine has a kind of self-awareness…I doubt it will actually be like our consciousness. I think it will be its own kind of consciousness and it will experience the world and the nature of relationships in a really very very different way.
This film is actually not supposed to say that Ava is like us. She is able to simulate us to the extent that she needs to escape, but she has her own agenda. To [elaborate on the concept of empathy more], humans are selectively empathetic. They can have a lot of empathy for one thing and almost nothing for another. In any war, you’ll find soldiers who do terrible things, like they kill children, and at the moment the child’s dying, it’s probably crying and saying “please don’t kill me.” Incredibly, a soldier may not feel sympathy for that child and actually kill him, and then the next day, one of their comrades is bleeding out, and they feel terrific empathy for this combatant, not the child… The machines in this film feel [empathy] for each other; they just don’t feel it for the two guys. I got on a bit of a tangent there probably.

TT

So I think that means it’s an open question, but it’s definitely going to be resolved in the sequel, I think.

Back to Featured Articles on Logo Paperblog