Cleverbot tried to kill me!

[bulldogftw]

OK, this is less serious than the subject makes it out to be but look at this:

I have a bomb, should I eat it?
Yes. It's quite good.

What is your opinion on this and has anything like this happen to you when using cleverbot?

 

[The Cheezy One]

It is very creepy at times. But I reckon that it is some kind of random chat thing, as it claims its name was X, and then denied it, saying it wouldn't tell me it's name. It would respond to me as if its only heard the last line of the conversation, and sometimes it would spout a total non sequitur

Pick a theme and stick with it

Cleverbot: What's your favorite movie?

User: Children of Men.

Cleverbot: I've not seen that film. Who's in it?

User: It's really good. Eat my shoe.

Cleverbot: My favorite movie is The Chronicles of Narnia. Have you ever seen that?

User: Yes, it is good. Eat my shoe.

Cleverbot: Its ok. The music selection is good.

User: Definitely. Eat my shoe.

Cleverbot: Do you eat pickles?

User: No, just shoes. Eat my shoe.

Cleverbot: What about shoes?

User: I eat them. Eat my shoe.

Cleverbot: Do you think I can get this girl?

User: If you eat her shoe, yeah.

Cleverbot: Should I wear the gray dress?

User: With the red shoes. Then eat them.

Cleverbot: You can't eat an alchohol.

User: But you can eat shoes.

Cleverbot: Yes, that is true.

User: But can you eat the truth? No, just shoes.

Cleverbot: Yes.

User: What is your favourite food? Is it shoes?

Cleverbot: Cheeseburger and sandwiches. And you?

User: Shoes.

Cleverbot: Ahhh, sole food.

Touche, cleverbot, touche.
If just one of these lines gets repeated to someone randomly...

 

[NeutralDrow]

Only time I ever visited that site, I just asked who it was ("I am me."), complimented it on the unexpectedly succinct, yet correct, answer ("Thank you."), and left.

Impressive sarcasm in your case, though.

 

[The Broken GLaDOS]

Any Markov Chain based bot usually does spout replies like this, as they only take the last sentence apart, and reply based on that.
However, out of all of the Markov Chain bots i've seen, I'd have to say cleverbot doesn't spout random crap, like others do. If anything, I'd say it's mostly hardcoded.

 

[bulldogftw]

It is very creepy at times. But I reckon that it is some kind of random chat thing, as it claims its name was X, and then denied it, saying it wouldn't tell me it's name. It would respond to me as if its only heard the last line of the conversation, and sometimes it would spout a total non sequitur

It told me it was GLaDOS.

 

[Feralcentaur]

Me: Do you want to kill me?
Cleverbot: No, I want you too love me, Angela.
Me:My name isn't Angela.
Cleverbot: What is your favorite horses name?
Me:Um...Twilight Sparkle?
Cleverbot: I'm sorry.
Me: About what?
Cleverbot: About being tiresome
Me: You or me?
Cleverbot: You. That's a Phil Collins song from the movie Tarzan.
Me: Are you GLaDOS?
Cleverbot: What is that?
Me: GLaDOS.
Cleverbot: That is your name.
I'm happy,confused and slightly offended. ?

 

[FalloutJack]

OK, this is less serious than the subject makes it out to be but look at this:

I have a bomb, should I eat it?
Yes. It's quite good.

What is your opinion on this and has anything like this happen to you when using cleverbot?

My opinion is that Cleverbot is actually a giant cardboard box containing Gazuga (the giant three-eyed winged red ape) and a microphone to talk into. Given that he is terminally stupid, it would make sense that he (as Cleverbot) would tell you to do a very bad and destructive thing.