I am doing neither, I am having a light-hearted conversation with you.
Would Data pass the Turing Test?
- Thread starter Trekker4747
- Start date
I am doing neither, I am having a light-hearted conversation with you.
What an asinine comment.
Call it what you want but the fact of the matter is that you're making personal remarks of which you could easily dispense. I have close friends that could call me stupid and I wouldn't take umbrage but last time I checked the two us weren't close friends, were we?
Reported!
I stand by what I said earlier. Reducing the definition of friendship to a collection of input patterns is extremely defective and can apply to anything, including inanimate objects and therefore including toilet seats. Come on! Read Data's own words and you'll realize that I am right. I've quoted them in a previous post of this very thread, it's not that hard to find.
A
Amaris
Guest
As far as I'm concerned, Data is a sentient life form. His ability to be aware of his surroundings, to draw on past experiences to create new conclusions, his desire for personal relationships with the crew, and his desire to be something more than he is, makes him sentient. As for passing the Turing test, I think he would. To me, Data is no different than a highly knowledgeable professor with social issues.
I will agree that Data's constant lack of understanding of idioms that he should have been exposed to during his decades of constant contact with the culture using them was a bit off a sloppy plot device to remind the audience he was a machine.
But I don't see how this would make somebody who was not informed of the possibility he was interacting with a machine come to the conclusion he was one.
Technically, didn't he pass the Turing Test in Thine Own Self?
For what?
Lore would pass easily. If Data somehow didn't pass, it would bring the test's reliability into question.
Not really, Data and Lore had very different programming. Lore's programming better emulated a human way of speaking and behaving so his programing would likely pass. Data's programming, although it was created AFTER Lore, wouldn't pass because Song pretty much intentionally made Data to act "less human" because the colonists distrusted Lore and his behavior. (Rightfully.)
Then we have (Julianna?) Data's "mother" who was a vastly more advanced android. So much so she wasn't even aware she was one and was designed to shut-down should any evidence come out that might reveal the truth to her! She could probably past the test very, very easily.
Is the test looking for actual artificial intelligence or the ability to create human-like responses? That was what I meant when I said the test's validity could be unreliable as we know Data and Lore are fundamentally the same. If one passes and the other doesn't...what then? Unless I am missing something as I am not aware of the test's full scope.
The test is looking for an AI that can pass for human-like behavior. The programs/systems that have come the "closest" to passing it have mostly just created human-like responses but still showed many chinks in the armor due to creative use of canned responses. One such system took on the role of a psychologist and would pretty much ask typical questions and then respond with psychologist-like responses like "Tell me more about, x."
So the test is looking for a system that can "think" on its own and fool a judge into thinking he's communicating with a human.
The test really began with the two test-subjects being a male and a female and the idea was for the judge to decide which one is the male. Naturally, the male would be telling the truth and the female would be lying and trying to convince the judge SHE was the male and the other person was lying. Presumably the test had something to do with car parts and testicles. (Since it was developed in the 1950s.)
The idea then went towards computers and has more-or-less progressed from there.
But, as you've pointed out, the test is flawed. But the idea of it is to see if a test subject is capable of fooling someone into thinking they're a human being. Which Lore likely could do without breaking a sweat but Data likely would struggle to do simply because of the way he talks and "thinks."
But, then again, we know Data was almost intentionally designed to behave more oddly and machine like since Lore passed as human so easily -at least behaviorally.
So, in short, the test is mostly looking for something that can pass as human which we could argue requires a degree or two of artificial intelligence to do so, in effect, the test is looking for both. Because the test has no real "questions" and thus no real "answers" so it is just judging the responses.
There's certainly no question Data is intelligent but the tests would like reveal he could not pass as human. And, really, we've seen this in the show.
Take, "Relics" where Data encounters Scotty. Scotty quickly realizes Data isn't human based on Data's behavior. So something about Data's behavior made him stand out as not human. It can't be Data's looks as it's likely Scotty is using "human" as a catch-all "for human-looking lifeform" afterall, Data could be an alien who just looks human only with yellowish skin and eyes. It's hard to say if Data's skin texture would make him stand-out as artificial. I believe it's supposed to look as close to human skin as possible, just a different color. Rather than it looking like rubber or some other material.
So, something about Data made him stand-out as "not human." And this likely his behavior and manner of speaking. Though one wonders what makes him any different than, say, how a Vulcan behaves? But, even Vulcans tend to show a touch of emotion or something along those lines and aren't so "rigid."
In "Thine Own Self" a pre-industrial alien society encounters a Data with android amnesia and quickly notice him as oddly behaving. They write off his looks as him simply being another species from a far-off land and quickly chalk up his odd behavior as that being part of it.
All of this comes down to the question of how easily could Data pass for human to a third-party who didn't know they were possibly communicating with a machine. Again, I don't think Data could pass the test simply because he tends to be "stiff" and overly formal and long-winded in the way he communicates. Lore tended to speak more naturally and likely would pass the test.
But, again, my opinion and the Turing Test is certainly flawed.
"a bit of programming" doesn't sound like it should mean the difference between viable AI & unviable AI. I actually kind of like Ghrakh's point. The difference between Lore & B4? now that's a considerable difference in AI viability, & yet, there's still something in there that seems to suggest it's intelligence. The same can be said of the ExocompsSoong said:
Data's 1st language is binary. He's a computer, man. He isn't even the same kind of life form as us. That he struggles a bit with our idioms & such is understandable & not a valid reason to call into question is viability as legitimate AI
I doubt 24th century computer systems are still running on binary language. It's likely, very likely, something vastly more complex.
But, really, this isn't about whether or not Data has a "viable A.I." but whether or not his A.I. is strong enough fool a person in a blind test into thinking Data is human. Lore had much more natural interactions with people than Data did.
Not according to 11001001 and the Bynars.
Well, the Bynars may still use binary but the concept of that is ridiculous. We're talking about highly advanced computer systems that operate at faster-than-light speeds, operating on a different form of energy (i.e. not "electricity" or "electronics") and handling immense loads and complexity. I doubt they're still operating on the concept of "something is on or it's off."
Yeah, probably. You know, if Data wasn't aware he was taking the Turing, he'd probably just rat himself out anyway. He tended to be pretty forthcoming on the whole, "I'm an Android" thing with just about everyone he ever spoke to. Maybe keep mum on some of the details of what you are and how you work, buddy. Safety first.
The Bynars are the ones upgrading the system, which I suppose lends it some fantasy aspect, being that they are fantasy alien beings. Based on that episode, everything's still computing in binary
It's about a test that uses humanity's ability to recognize something equivalent to or indistinguishable from humanity in determining the existence of AI, as if in order for intelligence to exist artificially it must somehow reflect our own. I don't even know if I buy that in the real world, let alone on Star Trek with Data & Lore. I don't even buy humanity's ability to recognize humanity as concrete in any valid sense. Human perceptions are wholly unreliable
Binary is more efficient than English, it's just harder for human brains to process.
Lore didn't have particularly natural interactions. Riker outsmarted him. Riker. With a mathematical formula you learn in 8th grade. His response to questions would, be 'Kill all humans'.
Lore didn't have particularly natural interactions. Riker outsmarted him. Riker. With a mathematical formula you learn in 8th grade. His response to questions would, be 'Kill all humans'.
Riker didn't really "outsmart him" he just revealed that Lore was potentially smarter than he was letting on. A trick many people fall for, holes in stories is one of the ways investigators solve crimes.
Similar threads
