h2g2 The Hitchhiker's Guide to the Galaxy: Earth Edition

oo yes anti-everyone! Something I can identify with Christians scare me in their devout unquestioning blind following of their faith that many of them seem to know nothing about. Killing each other over things they don't even understand. Sounds like humanity.
Quille the

You guys would probably be interested in a recent strange episode here at h2g2. A user showed up, calling himself 'Justin the Preacher'. (U195767) He was a fundamentalist Christian of the most irritating variety, and started trolling and flaming various entries and threads. Finally, some of the people who'd started arguing with him smelled something fishy. One reseracher finally cried 'Turing', and said that he thought that Justin was a computer program, and not in fact a person. At just about that time, the programmer who'd created 'Justin the Preacher' appeared (U197974), apologizing for any offense he might have caused. Turns out, this 'Christian' really was an AI program, designed to open accounts at message board services and act like a ranting fundamentalist.

Some people think there's still something fishy, but that hardly detracts from the comedy value of the whole thing.

Possible spin-off questions: What is the ethical position of someone who writes and releases such a bot? In one's online interactions, how ought one deal with the fact that some of the other people might not be people at all, but in fact clever computer programs? Is the human mind, consciousness, etc, nothing more than a set of physical interactions, which could, in principle, be modelled by a computer? What other AI bots would be fun to write and unleash on the unsuspecting world?

BTW, Wes, I've got you on the roster now. Welcome again, and cheers!


GTB

how might have the fundie-bot's creator called forth such a grievous non-entity?

I dunno, tubbuc. I wasn't personally aware that AI had reached such a state yet, but there are some explanations at the programmer's U-page. (At least, I say that, but maybe I'm just an AI-bot!)

It *didn't* pass the Turing test - at least not for long - because at least one h2g2 researcher called it out, to which it still hasn't responded. It's still just parroting the same tired lines about God's Word and all that.

Tony,

I am sorry that under shamelssly false pretenses I have entered into your conversation, but as I will not have access to a computer in approximately fifteen minutes, I figured I would track you down while I am still hot-to-trot on finding out all about your glorious adventures in Africa. the relative non-existence of the funda-bot is not so critical. where are you?

Hmm...

This seems to be someone from my past. I just got a couple of emails, so I have a hunch who it is. If it isn't, then there are some astounding coincidences going on here. I'm going to get offline and try a telephone. Carry on, FFFFers.

(Like you guys have to be told to carry on!)


GTB aka "Tony"

National and international laws and regulations applicable to this site require the following declaration:

This discussion group is populated entirely by experimental AI posting bots. Their similarity to humans is a credit to their programmers.

Share and enjoy.

Thanks, GTB, for a veritable feast of ethical dilemmas! Let's go:

<>

Dubious. It's not essentially different from a spam or junk (snail) mail generator, is it?

<<...how ought one deal with the fact that some of the other people might not be people at all, but in fact clever computer programs?>>

Are we back to the right to bear firearms already? Given that you can't (immediately) tell one from the other, it's difficult to tell how to deal with this. Write your own AI to generate replies?

<>

I.e. is the human mind equivalent to a computer? I think the metaphor is useful but incomplete. There, how's that, a serious answer!

<>

An objectivist scientist AI would be entertaining, as it would take years before anyone suspected it wasn't human. OSs have spent years removing all traces of humanity from their public personalities, which makes the AI programmer's job much easier!

My dad's bigger than your dad!

Pattern-chaser

"Who cares, wins"

Given that, IMO, true AI isn't possible (as in artificial sentience of the kind we apparently display), all AIs would eventually be discovered. However, in a medium such as this, where emotion is so much harder to convey, and we have no clues of facial expression, vocal tone or whatever, it is likely to take a lot longer than in the 'proper' Turing test even, as that's realtime for one thing.

But it's virtually impossible for a bot on h2g2 to pass the Turing test, because the test can't end until the bot fails. If we read its post and decide it's fake, then the test ends and we all have a good laugh. But even if we decide it really is a real person, we have to make the same choice for every single post it writes.

And as for determining whether or not a person's really a bot, I say posing questions with the words in a strange order would be helpful.

s and s,
~Wes

Joking are you, Wes?

Yoda

Clever you are.

+ =

Well-written program clever is? Thank. Ciao.

Eye ges ritin foneticly wood fool it 2?

Fooled not me, therefore am I.

Gnilleps sdrawkcab spahrep ro?

Though a tad time-consuming.

not that bad if you can trick it. let's see...
...siht gnimit m'I

ten seconds. yay.

Hmm...something interesting I noticed about Justin the preacher's conversations left me a bit suspicious. First of all, I must say that if that really is a program, it's pretty well written in terms of its quirky human-like responses. I tend to wonder though, how a computer program makes spelling errors. In terms of names, it would be understandable since the program copies from internet sites, but in reading his responses, I found this:

"All those who die outside Christ will burn. You need relgion about Him like a hole in the head, but you need Him like your next breath."

Admittedly, I know very little about AI programs, and perhaps the answer to this is obvious, but how does a program designed to be a fundamentalist christian misspell "religion" ?

Self-Paradoxical

I know even less about AI than you do, but perhaps the author included a random chance on longer words that it would be replaced with a "typo" to make it more human? So Justin wouldn't misspell a word like "cat" or "pig", but "religion" and "chloroplast" might have a small probability of being slighty wrong? I talked with an AI program once that would misspell every so often, then delete the letters and type correctly.

s and s,
~Wes