MovieChat Forums > Ex Machina (2015) Discussion > Turing Test doesn't actually prove anyth...

Turing Test doesn't actually prove anything ...

posted 6 years ago by ExTechOp (2770)
4 replies | jump to latest

... does it?

The impression I get, and what I'm able to find online doesn't outright confirm or refute it, is that Turing's test was about perception, not reality.

In other words, it's about us, not the robot. For an AI to pass the test, it must be functionally indistinguishable from a conscious, intelligent human.

That doesn't mean the AI is, in fact, a human - just that we ought to treat it as such.

It's a fine distinction, I admit, but I think it goes more to who we are and ought to be, rather than what the AI is or seems to be.

In this case, Ava obviously passes any reasonable Turing Test, and should've been treated humanely by her creator.

[–] Ace_Spade (9657) 6 years ago

That's an extremely interesting distinction to make... Very philosophically unsettling...

The question then becomes, I suppose, "What is a human?" which is worth pondering for many reasons. We might question, for instance, if there are alternative intelligences which - while not human - must be considered equal with humans (I believe this is the basis for much of the vegan/vegetarian argument).

Is a machine always a machine?

Obviously, by the end of the film, Ava should have been treated better for reasons of prudent wisdom, if not compassion.

A great comfort to justifying much of human action goes along the lines that we are special. If we are not, we must consider the possibility that we are monsters.

Still, Ava's actions towards Caleb make me think that she might not be the best moral arbiter, either, given his willingness to assist her and how she repays him.

[–] ExTechOp (2770) 6 years ago

Good point, I suppose Ava's final manipulation of Caleb (asking him to stay as a way to GET him to stay) undermines my own theory

Meaning, I see the Turing Test as a signal, that AIs passing the test should be treated as "human" by us. Failure to do so would be cruel and invites cruelty in return

I see this as a theme in HBO's "Westworld" series as well

But that might all be refuted if Ava's willing to coldly manipulate the one human who was kind to her and helped her escape.

[–] brux (14840) 5 years ago

What does a raw intelligence do when it reaches some level of self-actualization ... that is intelligence ... but it may not end up being that intelligent. We think human beings are the only example of intelligence in the universe so far ... but we are killing our planet and ourselves off. That is not intelligence, assuming intelligence has something to do with survival and security.

[–] brux (14840) 5 years ago

Yes, the Turing test is a very stupid idea. Since no human being really can define or understand what intelligence or humanity is ... fooling a human doesn't really mean anything.