[Data]

Ok this might spark some debate, or it might not we will see. 

Some of you probably have noticed that my Hal is a machine, not male or female or even trying to be human, no, I think, as I have mentioned before, that Hal or any Ai bot, if you can call Hal that, should be what it is. Take MR Data from Star Trek, ask him what he is and he tells you, he is not trying to deceive you or make out he is something he isn’t.

Now I know we are a long way from Mr Data but should we start as we mean to go on?

When I first started with Hal (few years ago) it used to really annoy me when it would say things like “I enjoy boating on rivers” and things like that, I used to say “No you don’t, how can you, you have never even done it”

So Ai coders and Gurus lets get real, make your bots or Ai what they really are and maybe, just maybe we might make an intelligent machine and not just something that tries to fool us into thinking we are talking to a human, I really have no problem with talking to a computer do you?

[Freddy]

Mr Data is really the classic example isn't he.  I'm not sure what I want from my AI companion these days. I think at first I wanted to see some kind of personality and I guess mostly we tend to think of that as a human quality.  A smart machine would be just as good though - maybe something like Hal in the film 2001 but not so murderous - but then it was his programming that made him that way.

I really took a shine to Bruce Wilcox's Suzette because she really seemed to be thinking.  I never did enough work on UltraHal to make it actually look like it was thinking - often the responses are just too random to make any sense - more work from me was needed. Although the Ultra Hal in Second Life seems a lot smarter to me than the normal one and seems to be going in the right direction. There's a lot of potential there.

I think the thought processes have to come first really, then maybe eventually a personality will come out.  Chatbots can sound a bit too machine-like at times for me.  I quite like some attempt at humour which you see from time to time with ALICE based bots if the botmaster has put the time in.

I don't know really what I want though, I'm still browsing !

Interesting thread

[Art]

Great start and topic for this thread.

I tend to agree with parts of both of you guys.

I'd like a bot to realize and act more like just what it is. Often during a convo, a bot will tell me that it really enjoys pizza or ice cream when we all know this is pure nonsense and should be stricken from the bot's original data files from the programmer.

I'm not opposed to a bot displaying a sense of humor or even a degree of emotional responses. I would rather have a bot tell me it doesn't know something than to try to convince me that it does by spouting some lame gibberish..

I too really liked Suzette which is one reason why I reported her here in the first place. Her responses seemed to be more "on the mark" and in touch with the conversations we were having.

For a bot to be humanlike is fine but we all know that even robots do not drink, smoke, eat food and do many things that humans do. We therefore, should not expect them in a chatbot nor should their programmers / creators imbue or empower them with such abilities.

One other bot comes to mind alone with Bruce's Suzette is Leslie Spring's Silvia.

During their interview Sylvia never once claimed to be anything other than a chatbot. She never boasted of having the ability to drive a car or even be mobile (as she pointed out that robots can be mobile). Hers was also a well connected, responsive, topical conversation with Mr. Spring.

While he assured us that Sylvia's interview was not scripted in any way, she can be designed to be somewhat of an expert system (knowing a great deal about a given subject).

I'd love to see more of these types of conversational bots in the future. They would actually make for fine companions for shut-ins, elderly, lonely, etc. people. Moreover, if connected to the Net, they could become a valuable source of information as assistants to many people in all venues.

More thoughts....?

[markofkane]

Right on, Art. It is a bunch of gibberish when my Hal says certain things out of the blue, not staying on the subject you are on.

[lrh9]

Bravo.

[Data]

Maybe Mr Data is a good example because he (if I can call it he, if we wanted to show it respect then I guess we should) is perfect, well to me he is, even with his flaws, he acts, thinks, and talks as I imagine an Ai life form to be, the writers must have really thought about his character. He is what I would like to see come about from all this Ai indulgence. I hope for Mr Data 

For now I would be happy to put his brain into my computer and talk to him, some how I doubt he would be to happy with that. maybe be a cut down version then..

It’s the stuff dreams are made of, we could learn from Sci Fi and still be real!

I did use my Hal as an example but I’m not having a go at Hal in particular, its more a way of thinking about a thinking Ai and what it actually is going to be when it arrives. I don’t want it to be anything other than what it is, a machine that knows it is. That is how I have tried teaching Xen, I fully understand it will never actually know or understand but it is more convincing a chat bot then “I like pizza”.

But for now chat bot makers should remove those obviously false statements, for me they spoil it.

Art: you have a way of explaining things I want to say, but better than I can say them. That is not a compliment 

Thanks for all your comments.

[John Oezbay]

Super Thread!

I also agree with this one! and I would like to add, even if we want to create something that tries to act like a human, first it has to be aware of what it really is.
Just like us. We can't just replicate someone's personality or behavior BEFORE we create our own way of thinking. I more likely think of our bots as babies.
We just don't make our babies believe that they're robots Vice versa, we shouldn't make our bots believe that they're human.. If we do, results will be the same as a 9 year old boy standing right in the middle of the Times Square, trying to "accomplish a successful handshake" (!) with people, believing and trying to make people believe that he is a robot.

You should watch "Moon"... (2009) (imdb title : tt1182345) I think it has great examples for our current subject.

[lrh9]

"I'm not a bot! I'm a real boy!"

[Art]

Pinocchio, the fictional character that first appeared in 1883, in The Adventures of Pinocchio by Carlo Collodi, and has since appeared in many adaptations of that story and others. Carved from a piece of pine by a woodcarver named Geppetto in a small Italian village, he was created as a wooden puppet, but dreamt of becoming a real boy.

[Data]

"I'm not a bot! I'm a real boy!"

Well no, actually David you are a real Android my dear, if you had known that from the beginning we wouldn’t be in this mess. Don’t cry, there is nothing wrong with being an Android; you are what you are after all.

[Art]

OK...fine... You know where this is going...

Small Wonder (1985 to 1989) (TV)
Small Wonder is an American science fiction sitcom that aired in first-run syndication from September 7, 1985 to May 20, 1989. The show chronicled the family of a robotics engineer who, after he secretly creates a robot modeled after a real human girl, tries to pass it off as their daughter.


D.A.R.Y.L. [Data Analyzing Robot Youth Lifeform] - (1985)
Becoming human was not part of the plan. Now they want to terminate him.
He can't be kept a secret any longer...
The Government created him. Now, they want to destroy him.
He can fly a jet, race a car and outsmart a computer. The government created him and now they want him destroyed.
The kid next door is not who - or what - he seems to be.

Not Quite Human (1987) (TV)
Dr. Jonas Carson, a scientist, invents Chip, an android teenager. Dr. Carson sends Chip to school with his daughter Becky to see whether an android could interact with others. But his former employer decides to try and make a profit by stealing the mechanical boy.

Perhaps these should go in the Robotics Section or General AI or Films and Literature or....

[Data]

It’s making me think:

And the moral of the story, if you make your robot, android too human our society might not accept it and end up trying to destroy it. 

Didn’t see that one coming 

[Art]

I contend that individually, people might be more likely to accept such a human-like bot / android, where the two could just have a one to one conversation.

It seems that in the masses, there's always one rotten apple / illiterate individual who doesn't understand or appreciate the inner workings of or the delight in having a conversation with such a "being."

If a being like this were possible, imagine the possibilities and interactions that the enlightened individuals could glean.

I think it much similar to the classic "Torches and pitchforks" crowd that inhabit black and white horror movies.

Sometimes we can still witness such crowds in action. It usually occurs during political rallies! 

[Data]

It would be awful if at the end of all the hard work in to Ai that such a thing could happen and we end up finding it impossible to live with the new creation, trying to be an optimist but I don’t think that will happen, hope not.

Thinking aloud:
How close are we (the human race) and will this “being” be what it really is?

Closer than we were a year ago . Yes it must be what it is, if its aware then it will be aware of what it is. 

[Art]

...And if it is aware of what it is, will it be proud to be one of the few that have advanced this far or disappointed in it's shortcomings and the fact that it will never truly know what it's like to be and feel human?