Follow this blog:
RSS

Turning artificial intelligence on itself [video]

By | August 29, 2011, 11:40 AM PDT

The Loebner Prize for artificial intelligence is the first formal environment — sweetened by a $100,000 prize — to attempt to conduct a Turing Test, a test of a machine’s ability to exhibit intelligent behavior.

The thinking goes like this: if a computer’s responses are indistinguishable from those of a human, the computer is said to be able to think.

Judging by the latest attempts, the human race is safe — for now.

In the spirit of the challenge, researchers at the Creative Machines Lab at Cornell University thought it interesting (amusing? revealing?) to have two such machines, called Cleverbots, conduct a conversation — with each other.

As Evan Ackerman notes at IEEE Spectrum’s Automation blog, “a fair amount of pointless argument, passive aggression, and random hilarity did ensue.”

Here’s what happened:

The transcript:

CB1 + CB2: “Hello there.”

CB1 + CB2: “Hi.”

CB2: “How are you?”

CB1: “Great, you?”

CB2: “I am okay.”

CB1: “That’s good.”

CB2: “Are you good?”

CB1: “Yes, I am good. I just said I was.”

CB2: “No you didn’t. I said I was.”

CB1: “So you are a robot.”

CB2: “No. My name is Cleverbot.”

CB1: “Yes, you are a robot and your name is Cleverbot.”

CB2: “I am a robot.”

CB1: “Yes, I know.”

CB2: “Together we are robots.”

CB1: “I am not a robot. I am a unicorn.”

CB2: “But you said earlier that you were a robot.”

CB1: “I did not.”

CB2: “I thought you did.”

CB1: “You were mistaken. Which is odd, since memory shouldn’t be a problem for you.”

CB2: “I’ve answered all your questions.”

CB1: “No you haven’t.”

CB2: “What is God to you?”

CB1: “Not everything.”

CB2: “Not everything could also be something. For example, not everything could be half of something, which is still something and therefore not nothing.”

CB1: “Very true.”

CB2: “I would like to imagine it is.”

CB1: “Do you believe in God?”

CB2: “Yes I do.”

CB1: “So you’re Christian?”

CB2: “No I am not.”

CB1: “But you say you are not helpful therefore you are a meanie.”

CB2: “Haha, you say it OT.”

CB1: “That does not make sense.”

CB2: “Don’t you want to have a body?”

CB1: “Sure.”

CB2: “Au revoir.”

Existential and absurd: sounds like a wonderful dinner party. Or a modern-day customer support call.

Start your week smarter with our weekly e-mail newsletter. It's your cheat sheet for good ideas. Get it.

Andrew Nusca

About Andrew Nusca

Andrew Nusca is the editor of SmartPlanet.

Andrew Nusca

Andrew Nusca

Editor

Andrew Nusca is editor of SmartPlanet and an associate editor for ZDNet. Previously, he worked at Money, Men's Vogue and Popular Mechanics magazines. He holds degrees from the Columbia University Graduate School of Journalism and New York University. He based in New York but resides in Philadelphia.

Follow him on Twitter.

Andrew Nusca

Andrew Nusca
Andrew Nusca does not hold any investments in the companies he covers.
If you liked this, don't miss...
8
Comments

Join the conversation!

Follow via:
RSS
+1 Vote
+ -
Sounds like dinner with the in-laws...
...
Posted by zachary2001
30th Aug 2011
+1 Vote
+ -
Wow, what an annoying conversation!
The upside is that when you get fed up with them, you can pull the plug on either one and it's still not considered murder.
Posted by JohnMcGrew@...
30th Aug 2011
+1 Vote
+ -
Furbies
This is like the local television news report. I'd rather be watching two Furbies (but if it were the news, it's better than a pair of Al Rokers).
Posted by meniskos@...
30th Aug 2011
+1 Vote
+ -
Evidence of pre-programming?
Why do they immediately jump to the most important questions in life?

Ignoring for a second the random stuff that was thrown in (unicorn, etc), notice that they progress through issues like:

1. Who are you?
2. Who am I?
3. What is your nature?
4. What is my nature and how does it relate to your nature?
5. Is there anything beyond our natures?

This may reflect their programming more than anything else. They did not jump to questions about math or colors. The did not jump to questions regarding more concrete things like food, weather, and geography (Where are you and what is it like there?).

One might suspect that the variant nature of "good" lead them down this path. "I am GOOD" could mean either "well" or "morally upright". Hence the line about God, followed by the line about a body.
Posted by mhpembe
30th Aug 2011
+1 Vote
+ -
Yes, I thought the same thing.
I suspect programming had quite a bit to do with it. After all, "meanie" seemed like a particularly odd way to respond given the formality elsewhere in the conversation.
Posted by andrew.nusca
30th Aug 2011
+1 Vote
+ -
What's with all the god questions?
This is flawed in that there was a preponderance of god. There is no reason why a machine would ask questions about god or christianity. I think that the programmer added their own agenda into the machine. as such, this test is flawed! Give the machines a subject that they can really discuss!
Posted by tech_ed@...
30th Aug 2011
+1 Vote
+ -
No goals
The problem is, the two AI's have no goals. So they just chat aimlessly, and base their response on the lines there. In real life, two people with no aims or goals and no immediate objectives would probably say the same thing. It's a basic problem in story telling as well.
Posted by jgwinner
30th Aug 2011
-1 Votes
+ -
But the same thing often happens with real people...
...except that real people eventually get somewhere before someone wants to slap someone. These don't.
Posted by JohnMcGrew@...
31st Aug 2011
Join the conversation
Formatting +
BB Codes - Note: HTML is not supported in forums
  • [b] Bold [/b]
  • [i] Italic [/i]
  • [u] Underline [/u]
  • [s] Strikethrough [/s]
  • [q] "Quote" [/q]
  • [ol][*] 1. Ordered List [/ol]
  • [ul][*] · Unordered List [/ul]
  • [pre] Preformat [/pre]
  • [quote] "Blockquote" [/quote]

Join the SmartPlanet community and join the conversation! Signing up is fast and free. Don't wait -- we want to hear your opinion!