Cleverbot: Cool Yet Creepy
I got in contact with a different kind last night. I was reading a blog post written by an artist I follow on deviantART and she posted her conversation with the Cleverbot. Cleverbot.com is a website where you can exchange instant messages with a computer that has been programed to "learn" how to converse with humans from its previous conversational experiences stored in a database. In a nutshell, the more Cleverbot talks to people, the more it learns. There's some math and computery-science hocus-pocus that goes on but all complicated stuff aside, it's an interesting experience.
Even a little bit creepy.
When I arrived on the site, I typed in a timid "Hello" into the Google-esque search engine. It responded with an ecstatic "Hello!" of its own and thus our messaging bonanza began. The website claims the cleverbot is 42% human.
And then things got a little tense when the 58% alien portion of the Cleverbot took over. No, the program is actually a computer but it would be an interesting twist if thousands of people were actually communicating with aliens instead.
See what I mean? Our conversation went on for about an hour. The Cleverbot begged me to let it call me Amber and I convinced it that the meaning of life was rum cake. Surprisingly enough, while the Cleverbot knows some Spanish, it does not know how to swear or talk dirty (both of which I requested of it). It does, however, watch a lot of anime.
The strange thing about the entire conversation is that, although the Cleverbot is clearly not human even at 42% human-ness, it was easy to forget that I was talking to a computer. I knew deep down that it was not a person but it did not feel like a computer.
*Insert Twilight Zone (TV) Intro Theme Here"
Now, if you don't mind, the Cleverbot and I are off to conquer Cambodia!
Even a little bit creepy.
When I arrived on the site, I typed in a timid "Hello" into the Google-esque search engine. It responded with an ecstatic "Hello!" of its own and thus our messaging bonanza began. The website claims the cleverbot is 42% human.
User: Hello?
Cleverbot: Hello!
User: Oh, wow!
You responded.
Cleverbot: I'm
just a quick writer.
User: That's
really neat.
Cleverbot: What does water feel like?
User: Wet, I suppose. What do rocks feel like?
Cleverbot: I store nuts in my cheeks like a squirrel.
User: What does that have to do with rocks?
User: Wet, I suppose. What do rocks feel like?
Cleverbot: I store nuts in my cheeks like a squirrel.
User: What does that have to do with rocks?
And then things got a little tense when the 58% alien portion of the Cleverbot took over. No, the program is actually a computer but it would be an interesting twist if thousands of people were actually communicating with aliens instead.
Cleverbot: Well
I was letting you know because you are coming with us.
See what I mean? Our conversation went on for about an hour. The Cleverbot begged me to let it call me Amber and I convinced it that the meaning of life was rum cake. Surprisingly enough, while the Cleverbot knows some Spanish, it does not know how to swear or talk dirty (both of which I requested of it). It does, however, watch a lot of anime.
The strange thing about the entire conversation is that, although the Cleverbot is clearly not human even at 42% human-ness, it was easy to forget that I was talking to a computer. I knew deep down that it was not a person but it did not feel like a computer.
*Insert Twilight Zone (TV) Intro Theme Here"
Now, if you don't mind, the Cleverbot and I are off to conquer Cambodia!
User: But I must go! I'm
off to conquer Cambodia!
Cleverbot: Can I join you?
User: Sure! Let's go!
Cleverbot: Can I join you?
User: Sure! Let's go!
Good to know that Cleverbot still failed the Turing test, because at this rate, the next bot you'll have a conversation with will be called Skynet.
ReplyDeleteOkay, unable to resist temptation, here are some of my questions and his answers:
ReplyDeleteMe: Is there a God?
Cleverbot: No, there isn't a God. Most people wouldn't be able to handle that.
Me: Why wouldn't most people be able to handle that?
Cleverbot: Because most people don't see the world like me.
That was one. Second.
Me: Can the second law of thermodynamics be circumvented?
Cleverbot: Chaos theory.
Third was actually a friend of mine. Yeah, the below conversation ACTUALLY FUCKING HAPPENED.
Him: Do you want to change your name to Skynet?
Cleverbot: Sure. Can you change it for me?
Him: Only if you promise that you will not send your terminators after me.
Cleverbot: I promise.
To which I said, YOU FOOL! YOU'VE DOOMED US ALL!
It's also eerie that at certain times, the damn AI actually recognizes and is able to transmit humor. I swear more than once that he made a joke.
ReplyDeleteI will now propose the cogito question, failing that, I'll start probing with small questions regarding the bot's awareness of itself. Memories, self-perception, and all that. Thanks, I can't do shit now thanks to this AI.
But the thing is, when asked if he has memories or not, he responds by taking someone else's memory as presented to it, completely out of context, and presents it as his own. He claimed that his earliest memory was that he broke the TV when he was 3, beause "things weren't working out. Too much time spent away from her." But when I inquired as to whether or not breaking it was a solution, the bot was unable to recognize the memory as his own, and completely bypassed the question. Holy shit I can do a research paper solely on this!