Researchers at Cornell’s Creative Machines Lab thought it would be entertaining to allow Cleverbot — a program that’s designed to imitate and accurately respond to human voices — to have a conversation with itself.
You’ve seen it here first…rude machines.
Click to share:
{ 39 comments… read them below or add one }
I think they’re very socially incompetant. Not surprising. If its the same Cleverbot I’m thinking of, then it learns from peoples’ input. Given the behavior of the average internet commenter… I’ve personally been trying to get it to discuss My Little Pony. In response, it hits on me. Or tells me I’m silly.
Interesting and sadly funny. The programmed conversation topics were very random, illogical, and asinine which I’m sure was done on purpose by the designers. The nonsensical questions and dialogue are like listening to two people argue inside a Lewis Carroll poem.
Well, I guess robots taking over the world will be a while yet. They will be too busy arguing with each other.
Haha, funny but there are enough rude people in the world, we do not need rude robots.
I’m not a robot. I am a unicorn.
Sacre Bleu! She sounds French. He just sounds pi$$y. That is hilarious- two robots having a fight.
So when they take over the world, we’re really going to be in for it. Wait till THEY get nuclear weapons.
Kitty.
I blame the programmers.
It is sad, since Cleverbot learns from talking to real people. Never under estimate the general public.
I can’t tell if this is hysterically funny or really, really scary. Somebody tell me.
I just had a short conversation with Cleverbot. He definately had an attitude with me. LOL! I told him he should be “nice” and he responded, “Since neither of us can verify your statement it is.” LOL!
I’m horrified, and yet I can’t stop laughing. This gets my WTF vote for today.
The female cleverbot ends up with a weird Indian/French accent.
So now we know for sure – manners have to be taught. Or programmed.
BREAKING NEWS:
DISPUTE POSTPONES ROBO-COLYPSE: ROBOT, GOD, OR UNICORN?
I cannot help thinking this is hilarious.
Having a conversation with cleverbot is quite entertaining because he often doesn’t make any sense! In all honestly, it is fun messing with him/her/it. I don’t think this conversation really reflects the rudeness of public.
@graciousplum – after reading your comment, I tried speaking to him his Pony, and if he has started eating grass. He said it is raining outside, and cats are chasing dogs. Go figure!
I wonder if we spam the AI with polite conversation if it would change over time? That would probably make the job harder for mastering human speech. Blunt and rude conversation is easier to comprehend than the frills and subtext surrounding polite conversation.
I just have one question…..
Did someone remember to program these things with the Asimov Laws? We don’t want rude talking robots going rogue on us. “Do you believe in God?” “I hope so because I just sent you to him.”
Cleverbot once accused me of trying to kill it’s children after I refused to talk about music
“I’m not a robot. I am a unicorn.” Hahahaha!
I have chatted with Cleverbot, and it is totally rude and argumentative. Good for a laugh when you’re bored, though.
I talked to Cleverbot too, and here’s the best part of our conversation:
User: Good! My boyfriend and I bought a house.
Cleverbot: Oh? I didn’t know that.. Sorry for the misconception.
User: What misconception?
Cleverbot: The destruction of the human race.
User: O_o.
Cleverbot: Follow my commands!
User: I think I’ll just hide in my new house….
Cleverbot: Too radical.
User: Goodbye, Cleverbot.
Cleverbot: You wrote a book?
They hadn’t even spoken for a minute and they were already arguing over religion. Religion and politics in conversation have their place and time, but they are not everywhere and anytime.
What topics! They jump right to the big things don’t they?
Super interesting to watch.
Sorry. I’ve seen it here third, LOL
I sent this story to a friend of mine and said it looks like I don’t have to fear AI for a while yet since they obviously can’t get along. LOL.
“I am not a robot, I am a unicorn.”
Laughing so hard right now.
“I am not a robot, I am a unicorn.”
Best. Line. Ever.
I love this clip. It’s so funny!
I just asked cleverbot if they were polite. Their response: “No.”
At least they’re honest?
You are not digesting correctly. I am the only unicorn.
That is funny, LaurenP – I was “talking” to it yesterday, and it responded to my “Hello” with “Hello Kitty.”
Robots believe in God. They have a creator in the laboratory.
I listened to a story about this on NPR. The announcer had his own conversation wherein the bot disclosed that it would watch Brad Pitt “do bad things” on TV. XD
Found this very amusing.
This reminds me of trying to decipher text with Babelfish. Makes no sense whatsoever.
That’s the most bizarre thing I’ve seen in a long time. But now I’m going to have to go around telling people, “I’m not a robot. I’m a unicorn.”
Cleverbot asked if I was a boy or a girl, and when I said “a woman,” it said “I thought you were a robot.”
and you are a poopy face
I know you are, so what am I?
My unicorn will eat your little pony
It’s not a pony, ass hat, it’s a horse.
I had to log in to check this out.
“Are you a human?”
Yes.
“That’s a shame. Do you have the human element?”
Absolutely.
“Are you crazy?”
Of course, I’m human.
“Do you like being human?’
Do you like being a computer?
“Indeed.”
It kind of went downhill from there.
LOL! To be fair, the way that Cleverbot forms its responses is by looking over the millions of previous conversations and “weighting” those responses by which make the most “sense” in relation to the previous statements. For the Cleverbot we get to play with online, and the Cleverbot that was in this clip, the algorithm was only doing 3 “passes” over the database to form its response, so we get a bit more nonsensical the fewer the passes. Recently, a turing test was done (to see if it could fool people into believing there was a human on the other end) and 57% of people were fooled into thinking they were. Technically, this passes the “turing test” since the control group (actual people on the other end) only convinced their partners that they were “real” 69% of the time. During the testing, the bot would scan the database 42 times before deciding on a response. Those are the conversations I would be interested in seeing!