Yeaaah... but how would you know in advance that it's a bot? That's the whole point of the test. You're screened off. You type in responses and get responses back but don't know who or what is typing them.
Printable View
Yeaaah... but how would you know in advance that it's a bot? That's the whole point of the test. You're screened off. You type in responses and get responses back but don't know who or what is typing them.
You're able to tell that it's a bot in this case since that's what Cleverbot is advertised as. The Turing test requires that you do not know whether you're talking to a human or a bot.
Trying to approach intelligence as something happening within the bounds of a skull, that can be replicated or even just simulated by reproducing the activity therein, is too reductive. Both establishing and maintaining consciousness, or individual intelligence, relies as much if not more upon feedback with the environment, and in all remotely intelligent Earth species, feedback with other individuals in a social context. The mysterious, immaterial component of consciousness is culture, and it is there, not in the brain, that the bulk of consciousness resides and is transmitted. Brains, or processors of some sort, may be necessary to intelligence, but a single brain is very likely not sufficient to produce it.
Uhhhh..?
Yeah, the Turing test requires you don't know which one it is. As I just said. You're screened off. Your only interface is somewhere you type messages and receive them. Obviously when doing the Turing test you won't receive a popup every thirty seconds saying "THIS IS CLEVERBOT LOLOLO". What are you even talking about please?
Question is also addressed to tommo, who has a habit of thinking that thanking inane posts relieves him of answering a question.
I thanked that post because it relieved me of typing the exact same thing.
I don't understand how you don't understand what I'm saying.
You probably don't understand what I was replying to.
Stormcrow said try the Turing test on cleverbot. The Turing test requires that you don't know whether you are talking to a bot or a human.
To see if you can tell whether you are talking to a bot or a human.
When you are talking to cleverBOT you know you are talking to a BOT.
I'm terribly confused.
Where am I?
Why am I sticky?
Have you ever even used cleverbot? The name is in the title, all the time, for fuck's sake.
Edit: I understand what you're saying, but you can't currently do a Turing test on cleverbot. The only place it resides is on a website. You can't do the Turing test if you go to cleverbot.com to do the Turing test in the first place.
I also confuz
But as part of the prerequisites for a Turing test you would first have to hide what you are talking to, and then double blind yourself by also doing it on a hidden person.
There is literally zero reason why you could not administer the test on this software, any more than you couldn't administer it to any other software.
In order to do a Turing test, OBVIOUSLY, you have to first conceal that information. You're not allowed to look at the damn thing before you decide if it's a human or not.
But that's NOT what I was replying to. Ffs.
I was replying to THIS "Also if you want to give a Turning test to AI go to cleverbot.com"
Not "If you want to give a Turing test a go, download the source code for cleverbot, compile it on a computer, go to another computer and get your friend to either personally talk to you or use the clever bot."
But all of that is contained in 'giving a Turing test'. If you just go on cleverbot.com, by definition you're not actually doing a Turing test, are you?
Exactly. Therefore you cannot do a Turing test on cleverbot. That's what we've been saying this whole time.
Okay, I guess this is all just a semantic discussion, then, based on confusion between 'do a Turing test by going to cleverbot.com and nothing else' and 'do a Turing test by going to cleverbot.com and actually doing the stuff you need in order to do a Turing test'. Terminated.
Except stormcrow said "Also if you want to give a Turning test to AI go to cleverbot.com"
Just read better.
I just did a turing test with Cleverbot, and a piece of cheese. I have determined that the piece of cheese has become self-aware.
www.cleverbot.com is not a website where you can administer the Turing test to an AI tho.
It's just a website with a chatbot on it.
I find it interesting that it is possible to discuss the subject of making a "thinking" computer program, and can even conceive it, but when one tries to sit down and actually write this program, it's hard to know where to even start?
What would be the fundamental rules that a "thinking" computer program must have?
Uhh...
How else would you interpret that? He clearly states that if you want to give "a Turning test to AI" [sic], then you should go to cleverbot.com; implying that either cleverbot.com was some sort of Turing test experiment or that chatting with the bot is equivalent to conducting a Turing test on it.
Most thinking computer simulate (i.e. not emulate) a mind by using lists of some sort that would resemble a brain, with the contents including symbols for whatever the program would be learning to use, and with connections running in between the items to represent how strongly connected arbitrary symbols were, based on the input the program receives.
e.g. A simple AI program would take purely textual input, delimited by linebreaks, or total input, and make simplistic connections between them as the user typed into it. The program may simply store input as structures in a linked list, each composed of the string of characters that makeup the input, and with another linked list which stores the connections and their strength between each symbol and the symbols it is connected to. The program would then scroll through its mind to search for a symbol which is most related to the input it receives, and output that symbol, which in this case is a sentence or so typed in earlier by the user. The program would simulate intelligence by repeating input the user has already given it and it would learn by adding each response the user would give it to its mind and associating responses that appear adjacent to one another.
If you say so.
No Turing tests are not given to AI. Turing tests are conducted with humans interacting with either an AI OR a human. You don't give a Turing test to AI.
Any whatever. You misunderstood. Done.
Sloth - I've thought of a way to make a bot that could pass a Turing test. I lack enough programming skills to make it atm, although it's really quite simple.
But for AI, as in it thinks.... I'm assuming you mean it could figure something out.
For example if you tell it to read a wiki article on thermodynamics, it would understand it.
This is a lot more complicated. I agree that it is hard to even know where to begin.
I really think that you would need to create a program which is just like any program we have now, it responds to input. Much like what you would call an "instinctual animal"; they are pretty much the same thing. Like you ask it what colour apples are and it reads a wiki article which says "Some apples produce [whatever chemical which makes them red] in their skin, other apples produce [green chemical] etc...... *unimportant stuff*" But say that the article doesn't mention the colour, just the chemical.
Then you need to create a program, or a separate part of that program, which observes what is happening in the first program, or part of the program, and decides which information to use and what to disregard in it's answer and figures out that the chemicals would produce certain colours. It then prints "Apples can be green or red".
It is sort of a simple concept. But the programming needed to create this may be very complicated.
The most difficult thing, I think, would be to figure out how to get it to discern which information is important and which information is useless (regarding whatever it is that it needs to figure out).
Now you have failed again ffs. -_-
Turing test - Wikipedia, the free encyclopedia
The Turing test is a test of a machine's ability to demonstrate intelligence. A human judge engages in a natural language conversation with one human and one machine, each of which tries to appear human. All participants are separated from one another...
Ya it was my blunder, I had a misunderstanding of the Turing Test.