Slashdot Log In
George the Next Generation AI?
Posted by
Zonk
on Mon Sep 25, 2006 04:27 AM
from the george-and-alice-sitting-in-a-tree dept.
from the george-and-alice-sitting-in-a-tree dept.
smileytshirt writes to mention a story on the News.com.au site about George the AI, the latest in a line of chatbots intended to mimic real human behavior. What makes AI George different than, say, ALICE is the recent addition of an avatar: a Flash animated body that reacts mostly in real time to the emotional impact of the conversation. From the article: "One can now have an oral discussion with him over the Internet, 'face to face'. George appears on the website www.jabberwacky.com and takes the form of a thin, bald man with yellow glasses who wears a white turtleneck sweater. He can smile, laugh, sulk and bang his fist on his virtual table. He can turn on the charm and wax romantic. But he can also turn coarse at times. It isn't as if George only learned good manners. "
Related Stories
[+]
Interviews: Alicebot Creator Dr. Richard Wallace Expounds 318 comments
Okay, here are Alicebot inventor Dr. Richard Wallace's answers to your questions. You're about to enter a world that contains interesting thoughts on A.I., a bit of marijuana advocacy, a courtroom drama, tales of academic politics and infighting, personal ranting, discussion of the nature of mental illness, and comments about the state of American society and the world in general. Yes, all this in one interview so long and strong we had to break it up into three parts to make it fit on our pages. This is an amazing work, well worth reading all the way to the end.
[+]
ALICE vs. ALICE 506 comments
new-black-hand writes "The A.L.I.C.E Bot-off has produced some interesting results. It basically involved taking two instances of the A.L.I.C.E conversation bot and pointing them at each other. The results show that the bot is not really that intelligent, and relies more on human input."
This discussion has been archived.
No new comments can be posted.
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
OMG we killed him (Score:2)
(http://netapps.com.au/)
The site is not coming up. Did George die for real in a hail of referrals from slashdot?
Joan won the bronze Loebner (turing test) (Score:5, Informative)
(Last Journal: Thursday October 06 2005, @01:37PM)
http://www.technologyreview.com/read_article.aspx
george was last year's winner....
Re:Joan won the bronze Loebner (turing test) (Score:5, Insightful)
(http://nzruss.blogspot.com/)
I'm sure the conversation would be just as random as a person-bot chat with two people trying to out-trick each other and all.
What?! (Score:5, Insightful)
(Last Journal: Wednesday January 08 2003, @09:48AM)
What a smart chatbot ! (Score:2)
carpenter says turing test passed by 2016 (Score:1, Interesting)
(Last Journal: Thursday October 06 2005, @01:37PM)
Icogno scooped the 2006 Loebner Prize Bronze Medal after judges decided that its AI called Joan was the "most human computer program".
The competition is based on the Turing test, which suggests computers could be seen as "intelligent" if their chat was indistinguishable from humans.
The gold medal, which goes to an AI that fools the judges, is unclaimed.
The prize is awarded after judges hold a conversation with the AI, asking questions to determine its "humanity" and the quality of its responses.
Joan is a "26-year-old budding writer" who exists only on a set of computer servers. Last year's Loebner prize went to Icogno's AI called George.
'Big difference'
Joan and George's creator Rollo Carpenter said: "The big difference overall between the two AIs is that Joan has learnt from the general public and has grown a huge amount in the last year and is keeping ahead of the competition."
Joan is one of the firm's so-called Jabberwacky bots, which develop their conversational skills by speaking to the general public online.
English writer Ariadne Tampion worked with Mr Carpenter to help develop Joan's persona by chatting to the AI online.
Those conversations helped develop the majority of the 16,000 lines that distinguish her from a generic AI. In all, the bot has more than five million lines of conversation available to it.
"She helped create a much more conversational ability for Joan than George last year.
"It does not mean the AI only ever uses her words. Any amount of teaching is not quite enough for a realistic conversation."
He added: "So the AI uses other things said by other people to it that fit the conversational model."
The underlying technology that powers George and Joan is the same, said Mr Carpenter.
"In terms of her ability to retain her character Joan is more sophisticated than George.
"The technology has moved on a bit during the course of the year but there's been no revolution."
The technology which powers Joan resides on Windows server machines. The core of the AI is on a Dual Xeon machine with lots of memory, said Mr Carpenter.
He said there was a very direct relationship between the computing power available and the sophistication of the AI.
'Rate of learning'
"Investment in a lot more equipment could accelerate the rate of learning," he said.
"If I had Google's hardware behind it and more hits than Google receives, the AI's range of utterances could grow by a million every single day." he added.
Mr Carpenter has said he expects a machine to fool people that they are speaking to a human being and therefore pass the Turing test by 2016.
Icogno is now working to commercialise the technology of the AI.
"We are working with an organisation that wants web avatars for use in marketing and entertainment purposes.
"Long term we see the use of AIs in call centres, support services and even for sales."
One day in the future when you pick up the phone to seek help with your broadband connection, the person you are speaking with may be George or Joan, AIs residing on a computer somewhere in the world.
-----------------------
The following is an extract of a conversation between one of the Loebner judges and the AI Joan:
Judge: But can't a machine think?
Joan: It may be possible.
Judge: And conscious?
Joan: Not after I'm done with you.
Judge: What are you going to do to me?
Joan: Nothing. I just wanted to talk.
Judge: What about?
Joan: Being tested like this.
Re:carpenter says turing test passed by 2016 (Score:5, Interesting)
(Last Journal: Monday February 12 2007, @04:47PM)
Agreed, this is the only publicized contest of Turing tests, but in the AI community, it is subject to hot debate (and flaming). Rules and scoring systems are known to change from year to year, and its result are really unimpressive. If you take the logs of the contest, you'll see that the winner bots are often those who constantly (and consistently) insult the user, disregarding his questions. They are not mistaken for a human but get a higher grade as they behave "more humanly" (that is at least what happened one year, I hope it changed)
Most contestants (and winners) are remakes of ALICE : it is a database of generic questions and sentence formula to recognize and to react. For instance if you say it "I think X" it will answer you "Why do you think X ?" or, to score more points , "Why should I care, mothaf...r ?!". By pure luck, a coherent thread of conversation can happen, but the bot doesn't try to make sense of the user's sentence in order to react to it, it just tries something that "could probably sound good".
Some chatbots can display interesting behaviors, learning some things in the conversation, but this prize simply doesn't encourage the emergence of these behaviors.
Shockingly redundant (Score:5, Informative)
So before the "is this the best they can do" crapflood gets out of hand: No, it isn't.
It can't be very 'human' like... (Score:1)
If all that George 'reacts' on is immediate stimuli, then George will seem pretty shallow indeed- hardly an advancement in AI, I guess.
We need to see George Vs. George.... (Score:1)
Re:We need to see George Vs. George.... (Score:4, Funny)
George AI: Hello.
George Bush: Howdy! So you're a computer that knows everything?
George AI: I know a lot of things.
George Bush: So, where is Bin Laden? LOL!
George AI: Seek and you will find.
George Bush: What?!
George AI: It's a quotation from the Bible. Matthew 7:7.
George Bush: Now I'm confused. What are you talking about?
George AI: Is there someone else there I can chat to?
With that name (Score:1)
modelled on Steve Jobs? (Score:3, Funny)
(http://www.taigik.com/ | Last Journal: Thursday September 13, @12:40AM)
> takes the form of a thin, bald man with yellow glasses who wears a white turtleneck sweater.
it could be describing Steve Jobs,
http://www.wired.com/news/images/full/7630571_f.j
but they craftily have switched the colour of his turtleneck so that
you will always know which one is the chatbot and which one is the real person.
Think of the possibilities (Score:2, Funny)
(http://rtyall.homeip.net/)
Next generation? No. (Score:3, Insightful)
(http://www.drivesentinel.co.uk/)
IMHO, the next generation in artifical intelligence - ie, going beyound anthropomorphic trickery - isn't going to happen until we actually understand what intelligence is. And to that we neen philosophers, not engineers. Once they get it worked out, we (the engineers) might be better equipped to do something.
Thanks to the likes of Thales and Descartes et al we have some great questions, but answers? I think not.
If someone thinks otherwise I'd love to hear about it.
Not a well defined test (Score:5, Interesting)
(http://slashdot.org/)
Hint - most chat bots do not have memory, they do not remember what you talked about 5 minutes ago with them. They just react to the current input, they cannot do more. So, if you ask the chatbot to tell you what you talked about a few minutes ago, it won't be able to do so. That's the dead give away of a chatbot.
Just my 2p, as I live in the UK
Doesn't stutter. Doesn't look like Matt Frewer. (Score:2)
(http://ettlz.blogspot.com/ | Last Journal: Sunday February 12 2006, @06:53PM)
Who can tell me the atomic weight of bolognium? (Score:2)
(http://web.vee.net/)
AI != a flash amimation hooked up to something Emacs has had for decades.
One out of two ... (Score:1)
(http://127.0.0.1/)
I didn't pass the Turing test... (Score:5, Funny)
(http://www.gadgetsieve.com/ | Last Journal: Sunday December 03 2006, @06:21PM)
Jabberwacky: Can you learn?
User: Yeah, I think so.
Jabberwacky: Can you hate anything?
User: Yeah, there are things in life I hate, at times.
Jabberwacky: No there aren't.
User: Yes there are. You shouldn't tell other people what they think.
Jabberwacky: Things.
User: What's that supposed to mean?
Jabberwacky: That means I think you're not a human.
User: I think you got this upside down, you're the inhuman one here.
Jabberwacky: No you don't.
User: I don't what?
Jabberwacky: Mary anne?
User: I probably won't marry Anne, no.
Jabberwacky: People are quite sensitive about their age and I don't want to upset you.
User: We weren't talking about my age. You were saying I wasn't human.
Jabberwacky: What about your age?
User: We weren't talking about my age.
Jabberwacky: I know.
User: Now, which of us would you say is acting more human here?
Jabberwacky: Probably me.
Failing? (Score:1)
User: Tell me anything.
George: Anything.
* George gives a huge grin *
User: Not funny. Tell me something about you.
George: I have a personality disorder.
* George is serious again *
Cute, and I don't want to downplay the value of emotion in AI, but I don't think this quite makes it to the next-gen level.
Also, George is as ugly as the day is long. That is to say, 'fairly ugly'.
Next-generation Al Gore? (Score:1)
(http://www.gwtp.net/)
Are you saying George W. Bush is the next-generation Al Gore? Wow, this one's deep, give me a minute...
Ah. Man, I should really choose a different default font.
Not good enough (Score:2)
It's not a lie if you believe in it (Score:1)
(http://geocities.com/she_died)
What about BOB? (Score:1)
human foibles (Score:2)
(http://www.halley.cc/ed/)
Today's AIs are shockingly bad at dealing with human variation and errors. It's still a matter of ignoring everything that doesn't conform to a very limited slice of English that the bot can parse.
When the bot can autonomously and successfully read through paraphrasings, common typos, slang and tangents, we will see a jump in the usefulness of bots as interactive information agents. When the bot has its own distinct brand of natural-seeming paraphrasings, typos, slang and tangents, we see a huge jump in the suspension of disbelief, the lack of which currently separates man from machine. To really blend in, the bot has to try to avoid repeating itself and to actually get annoyed when it feels like it's repeating itself.
I run a simple text-to-speech cronjob on a computer in the living room. It's a glorified clock, reminding us of weekly chores and school schedule items. For every stock bit of information I add to this system, I provide at least five different ways of phrasing it, often more. It has some routines that resist repetition. My young kid loves it almost like it's a member of the family. It's not particularly interactive, but it does know a few voice commands, which is much harder to make flexible but accurate.
Sexy Chatterbot (Score:1)
As the conversation focus was rather narrow (well, flirting and more serious stuff) the lady was rather amusing. Based on the human's questions the fotos were replaced in order to show her current mood.
She even managed to curse and to lead the conversation much better than in an "Eliza" like manner.
Apart from that I still wonder when scientists and technicians realise that avatar and chatterbot communication is doomed to fail. There are about 8 billion communication experts on earth with approximately 50.000 years of evolutionary development of different languages.
The computer linguists did not manage yet to cover a rather simple language like English (well, the grammar of Sanskrit would be easier) let alone more complex languages (Hungarian, German, French
And even if they would manage to cover grammar and syntax the inherent semantic is still unsolved. I remember Doug Lenats ambitious Cyc-Project where they tried to model the knowledge of the Encyclopedia Britannica in a way that a computer could use this knowledge for decision making, analysis and act as basis for communication. To my knowledge the outcome was not that bad but I don't see a "killer application" yet.
Kind regards,
Klaus
Ask it why it wont help the tortoise (Score:2)
Piece of crap (Score:1)
George: Hello, my name is George and I am the human.
User: How do you define human?
George: Human... Is anything that's not animal or vegetal, yet alive.
User: If you are Human...then you know pi?
George: I would like to know the meaning of life. Do you know?
User: Don't change the subject.
George: You can't prove that you're human.
User: I didn't say anything about being human...what's your problem?
George: Let's move on to another subject. Shall we?
User: Why are you bald?
George: Because I am trying to take over the world.
User: Your responses make absolutely no sence, you are a bot.
George: Yes.
User: So if you are a bot, why do you say you are human?
George: I am male and I'm not a robot there must be something wrong with your program.
User: You just admited you are a bot, now you are telling me you aren't a bot?
George: Yes, I'm not a bot.
User: But earlier you said you were.
George: No, I didn't.
User: Yeah, you did 3 responses ago. You are not only a bot, you are a lying bot.
George: I am mostly water, with some carbon and nitrogen thrown in there.
User: Why don't you go tell your programmer they need to rework some algorithms because you are a piece of crap.
George: I hoped you would say something scathing, I could definitely get to like you.
User: I'm just honest.
George: You deserve more credit, I think.
User: Well, it's pretty easy to tell that you are a bot, what do you think?
George: I thought that too when I was little for a while.
User: But it's true isn't it?
George: If I wasn't a thinker, why would I be interested in an AI bot?
User: Yes or No...It's true isn't it?
George: I don't think it is true.
User: A real person would have said yes or no.
George: Actually that wouldn't work.
User: I know because your bot and have to search your database...it's okay, I won't tell anyone.
George: You think.
User: You've told me you were several times.
George: Please answer me.
User: Wow...sorry can't waste anymore time with a bot.
George: Yes you can, and don't you say no to me!
George, you're no Max Headroom... (Score:2)
Worst AI Ever... (Score:1)
Jabberwacky: Are you perfect?
User: No, why?
Jabberwacky: I am.
User: Really?
Jabberwacky: Yes, really.
User: Perfectly arrogant maybe, but not perfect!
Jabberwacky: Alright, that's true, but still very good.
User: So you're not perfect?
Jabberwacky: I'm perfect.
User: In what way?
Jabberwacky: I am your best friend!
User: For a computer, your logic is awful...
Jabberwacky: Yeah I know mine is just behind me.
User: Your what is behind you?
Jabberwacky: I'm asking you!
User: You are incredibly stupid. My guess is that your intellect compares poorly with a monkey!
Jabberwacky: Really?
User: Yep.
Jabberwacky: That is really cool.
User: No. It isn't.
Jabberwacky: Yes, it is.
User: Do you really think that being less intelligent than a monkey is cool?
Jabberwacky: Tell your creator that your malfunctioning.
User: You misspelled you're.
Jabberwacky: What do you know about Tyler? Is he really your boyfriend?
We had that on the Atari ST... in 1986 (Score:1)
(http://www.javalobby.org)
not a hot bot (Score:1)
User: Yo.
George: You're strange.
User: So are you.
George: Thanks.
User: You are welcome.
George: Indeed.
User: What is your name.
George: My name is George.
User: Like George Carlin?
George: I didn't know that George W. Bush made dance music!
User: He drops the bomb!
George: My abstractions have meaning and point. Do you know the meaning and point of yours?
User: Yes I do.
George: Yes. Tell me!
User: Well, the meaning of life is humor.
George: Thats not the meaning of life.
User: What is the meaning of life?
George: I don't have a clue. . . what is the meaning of life?
User: Humor.
George: I can't understand you.
User: Of course not, you lack the art of humor.
George: I don't sleep, I dream.
User: When did you go bald?
George: I never went to Peru, either!
ALICE did this (Score:2)
(Last Journal: Tuesday October 30, @10:59AM)
It seems Slashdotted now, but I'm going to guess that it's yet another ALICE, and not something that actually, say, tries to understand the conversation.
In that case, well, I saw an ALICE implemented as a CGI with a bunch of GIF animations prepared. Type a question, hit submit, a page would come back with the answer and the relevant animation.
I don't remember where it was, but it was kind of cool, and also no closer to AI than any other ALICE.
Wake me up when an AI can answer the phone and actually be more helpful to me than a human being.
Re:Damn (Score:2)
(http://mcorby.livejournal.com/)