Slashdot Mirror
Artificial Intelligence IRC Bots?
JonahC writes "dotcomma has an interesting programming challenge for people to create IRC bots with artificial intelligence and carry on conversations. With some development, good programmers and some luck, the bots should eventually interact with each other as if they are humans. "
Throw them in #teens.
Bot #1: A/S/L
Bot #2: Where are you from?!@#!@?#
Bot #1: A/S/L
Bot #2: Where are you from?!@?#
etc
Welcome to the real world.
Unix is user friendly... it just chooses it's friends selectively!!
Now how to explain newbies you are not supposed to talk to the bots? : )
Never knock on Death's door:
More race stuff in one place,
than any one place on the net.
Yey.. bots we can argue with.
-FweE-
Inspired by this, I did consider creating a trigger based system (a trigger being a regexp to match, and an output sentence) which used a genetic algorithm to assess the success of each trigger and to make the best ones "breed". There was also a simple "flag" mechanism to act as a memory. While I acknowledge that such a program could never pass the Turing test, the important thing to remember is that this IRC challenge is not a Turing test as (I suspect) most of the "judges" will be unaware that they are talking to a piece of software and are therefore more likely to give the system the benefit of the doubt when it says something really stupid.
If any slashdotters are planning on entering this, email me, I will be happy to pass on my thoughts and knowledge.
--
who wants to watch 2 boxes talk to each other. You can get that on the street corner.
Being able to talk with each other 'as if they were human' would be equivalent to passing the Turing Test. And nobody has managed that, despite fifty years of trying.
-- Ed Avis ed@membled.com
This is serious flaming, beware:
Flooders, warez kids, all sorts of people looking
for intercourse partners, idlers, the list goes
on and on. And now... IRC bots with AI?
Some might say those bots would be the first
intelligent agents on IRC since 1995.
And then there is this new breed of Linux user:
#whoami
an idiot
#
Closed source bots. I wonder how many bots will be released a'la GPL.
Never knock on Death's door:
More race stuff in one place,
than any one place on the net.
I would think that human communication would be one of the hardest things to get an AI to do. The nuances and slang would present a great challange and no doubt consume a great deal of computing power. There must be a better application to develop AI for than IRC. Intelligent farm machinery, house cleaners (ducts, vents, ect...). I saw quite a while ago a neat application of AI that had small robots plucking harmful bugs from plants, killing them, and putting them into a container. This would be a much better areana from development of AI than IRC. Just MHO...
I once came up with thought like this: If
I had to train my AI somewhere, the IRC would
be the best place. No physical body needed and
(at least possibility to have) lots of free
intelligent trainers to talk to and to learn
from. Are you somebody hasn't actually done
this already?
Considering the kind of human you normally encounter on IRC wouldn't it be more convincing to create artificial stupidity ?
:)
Let the bot enter #linux and yell "HELLO ROOM" then proceed to ask complex and interesting questions in #windows95 ( after being kicked from #linux ). "The window on my car is stuck and some dolt at the shop says it's because I spilled glue in there. Could he be right ?"
Who knows this bot could even get opp status and monitor a few channels. oppbot* on Windows "HaXoR25 was kicked for straying off topic. 'Uptime' is not an allowed word here".
Do something fun for a change
--= Isn't it surprising how badly I spell ?
I'm wondering: if you manage to change a bot by talking to him over IRC and the bot is licensed under the GPL, should you commit your therapy back to the opensource community? :)
Jeasus, This shit is easy to write. I might even write it myself if someone else doesnt jump in within the next few weeks. Who (from the BBS days) doesnt remember chat with lisa?. A basic fake chat "girl" person who pretended to be your friend and even have sex with you. And before anyone else mentions it - yes infobot has existed for years, and before infobot many programs that were better (written for the client i programmed (and a minute few others :)) with the python script, even without the python script) have been around much much much longer (try before irc even started). Maybe they should look at awarding their prise to an existing program.
...and then there should be a contest to write the best bot that can imitate a teenage male pretending to be a buxomy blonde with large brea^G^G^G^Geyes.
...since the last time I was on IRC, creating a bot that will act like most of the "humans" on IRC shouldn't be too hard..
:)
#hack
newbie> Can someone helpme hack cnn.com?
aibot187> l4m3r!
User newbie has been kick-banned from #hack
#warez19534562
newbie> Does anyone have a site with WIN98?
aibot187> l4m3r!
User newbie has been kick-banned from #warez19534562
#teensnunsandvegetables
newbie> Any cute girls out there have any personal pics?
aibot187> l4m3r!
User newbie has been kick-banned from #teensnunsandvegetables
...ad nauseam..
//Phizzy
"Most European technology just isn't worth our stealing," -- Former CIA chief James Woolsey, referring to Echelon
* HackerD00D has entered #warez4free
-HackerD00D- any1 have all 5 gigs of quake 4? will trade for entire backstreet boys mp3 collection
-ElizaBot- What do you feel about backstreet boys mp3 collection?
-HackerD00D- oh man, their the gr8est! there 40th reunion tour was the bomb.
-ElizaBot- So, you have feelings towards the bomb?
-HackerD00D- what, me? no way d00d! Get off my case!
* EschlonBot has entered #warez4free
* EschlonBot rudely grabs HackerD00D and locks him away.
* HackerD00D has left IRC (BUSTED!)
-EschlonBot- Thank you for your cooperation
* EschelonBot has left IRC (Obey the prime directives)
-ElizaBot- Tell me about your feels on obey the prime directives.
"Pinky, you've left the lens cap of your mind on again." - P&TB
"I can see my house from here!" - ST:
I once spend a lot of time hacking with expect (a tcl based even scripting language) to create an IRC cyborg.
Basically like an irc client with scripts, things to make me look active 24 hours a day, automatic babble generator and a few other 'intelligent' features.
The great thing was that I could wrap this around any shell, so the random babble would work on BBSs, talk, pine.... whatever.
There's all this talk about pluggin people into systems to improve the interface and increase a person's capabilities - but this sort of scripting is essential what cybernetics is all about. all those people with automatic op/kick/greet scripts are essentially IRC cyborgs.....
;-)
Just sent these strings: Hi! 21/F/LA Yes 5'10", blonde, blue eyes. Yes Are you? Oh yes.
These things could be programmed to be the other party in chat room sex, or to catch perverts trying to have chat room sex with a minor.
The ultimate development would, however, be for IRC bots to trick each other into having chat room sex with another bot.
"For every complex problem, there is a solution that is simple, neat, and wrong."
My office has been taken over by iPod people.
I keep flashing back to the cartoon of the robot on the analysts couch saying, "Just because my intelligence is artificial, doesn't mean my problems aren't real".
From what I've seen from my kids chatting and other list experiences, the problem of a bot on IRC shouldn't be that difficult. Short, pat answers with cryptic acronyms and a serious streak to talk about themselves regardless of the reply given. Can't be any worse than have a marital phone answering machine that just keeps repeating uh-huh... yep... I'm sure you're right... of course honey...
Mine's done:
Don't need much more than that... :-)
dylan_-
--
Igor Presnyakov stole my hat
Try to talk to Aii on #italia on EFNet 8) It's some horribly written 4 years old pattern matching code but as a female bot she is not supposed to make much sense anyway.
Perhaps I should have put an auto DCC send on the word "picture" hehe
ciaoxx
Just sent these strings: Hi!
21/F/LA
Yes
5'10", blonde, blue eyes.
Yes
Are you?
while (true) Oh yes.
Yes big boy.
Oh!
There's an interesting little program with which you can converse. It's called gNiall, and you can find it at Freshmeat. You talk to it, and it looks for which words come next to which word and how often. So then it constructs sentences based upon these words. They're usually pretty nonsensical, but sometimes it'll reply something profound or really "on-topic" and it'll make you wonder. It's good fun to play with, although you can't feed it books yet to learn. Check it out!
Cheers!
Costyn.
The Official Steve Ballmer Webpage
With everybody and their custom IRC scripts, it's funny to see the chain reactions of kickbans,etc when someone rides in from a netsplit. Almost like the game of Life.
Overall, the bot was fairly stupid, but smart enough to fool some of the people who decided to message it.
When I was able to do my own spam-armoring, you got a chance to email me. Now you can only hope I see your reply.
Generic PR bots could be made along the same concept. A small company could then have a 24 hour support and PR line with out having to man it the whole time to answer very simple questions.
The only thing is that most of this is already in place. Just take it off IRC, make an independent solution that embeds into web pages and your set.
The new form of the problem can be described in terms of a game which we call the "imitation game." It is played with three people, a man (A), a woman (B), and an interrogator (C) who may be of either sex. The interrogator stays in a room apart from the other two. The object of the game for the interrogator is to determine which of the other two is the man and which is the woman. He knows them by labels X and Y, and at the end of the game he says either "X is A and Y is B" or "X is B and Y is A." The interrogator is allowed to put questions to A and B.
The Turing Test is a slight adaptation of this imitation game. There are three players: an interrogator, a human being and a computer. The interrogator is connected to one person and one machine via a terminal, therefore can't see them. The interrogator's task is to find out which of the two candidates is the machine, and which is the human, only by asking them questions. If the machine can "fool" the interrogator, it is intelligent.
FYI, there is also another competition called the Loebner Contest, where a $100,000 prize is offered to the author of the first program to pass this Turing test. Bots have entered this competition often, with varied degrees of results.
...would of course be to code an APISAC(FP) (Artificial Pseudo-Intelligent Slashdot Anonymous Coward (F1R57 P057))
All opinions are my own - until criticized
How intelligent is the person interacting with
the bot... that makes all the difference.
Mark
Noone will suspect my bot to be not human.. I'll just set an away message like "I am sleeping away me hangover - try later" and voila a fully functional bot capable of imitating the average joe-student in crises-times. :)
Regards,
Depending upon the channel these bots would inhabit, one wouldn't need to build a turing machine.
For instance, take #teenchat. All you need is this snippet of code to beat the challenge:
while (1) {
print "U wan 2 chat?\n";
sleep(2);
}
Steve
Just what we need, another type of useless activity, not the experiment istelf but rather 50 Million Bots running around soaking up precious internet bandwidth for no reason.
Any type of AI research is worthwhile, but even 99% of the people in IRC are a waste of DNA.
-- A Human Being is nothing more than mobile CO2 factory. Bow to the plants.
Perhaps we could use this bot to idenitify people using languages other than the set language for the channel... and work as a translator... straight translations, or perhaps humourous.
-[_efty- I o/c'd my G400 to 150/200MHz, and I didn't really need a fan, it's got a beefyassed heatsink on it stock
-babelbot- I o/c'd mon G400 à 150/200MHz, et je n'ai pas vraiment eu besoin d'un ventilateur, il a a beefyassed le radiateur là-dessus des provisions
-ClaudeQ- Ne avez-vous pas besoin d'un radiateur pour maintenir cette chose froide?
-babelbot- Don't you need a radiator to maintain this thing cold?
Nevermind, I suppose this would get annoying really quick.
_________
My name is Aaron Landry, and I approve this message.
I am strongly against this idea for one simple reason.
Spammers will probably use this to better evade IRCops like myself who spend hours dealing with their stupid BS in the first place. pr0n, Long Distance, etc...
We already have a hard enough time weeding these idiots who flood channels, mass msg, and invite whole segments of lusers. Then of course, you have the zoom bots, the ones that join, get the channel list, ignore the IRCops, spam the rest and leave.
Add some AI to it, now, converrsation can ensue, possibly making it much easier to evade the policy of the network. Why stop at conversation? Lets teach the bots how to BNC when it gets an autokill or gline. For client based bots, how about we teach them to dial back in, to gain another IP. To change its ident, so it can evade any KLine/autokill/gline and keep right on spamming.
*sigh*
Want to invent something worthwhile? Streamline ircd for the next 10 years. Its starting to show its age. =)
Supernaut
making a bot that can chat is next to line with watching cement dry. Absolutely pointless. But collecting information by analysing the conversation is an interesting one. That might just breath life into my half-finished client
Instead of a bot that can just chat (an interesting AI experiment but useless), make a bot that does a special task. A software company could get a support bot, give it information, and put it on their channel, so people would get live support, but i've used some existing services like this (webhelp) and they aren't perfect yet.
They have a poll on which editor you use to create web pages, and there's no mention of Emacs? What sort of true hackers would hang out on such a site that doesn't think to mention Emacs? Look at what they do list : Notepad, FrontPage... Yipes.
www.HearMySoulSpeak.com
What's next?
Bizar technology?
Okay, why is this impossible?! Just cuz people dont' wrtie good english on IRC, U know?! poeple don't correct their spellings either, and use all forms os short cut, people probably will not say hello, more like, yo, sup, werd! la la la, it is a waste of time in my opinion. :-)
------ Curiosity killed the cat. {satisfaction brought it back | it didn't die ignorant | lack of it is killing mankind
we need virtual irc bots that will virtually pour grits down each other's pants. thank you.
Why limit this to IRC? Think of all the fun you can have if you eam it up with a plausible speech-synthesis system. Next time a telemarketer called, you'd have him talking for hours, while watching a movie or something. Not to mention all the daily chitchat-conversations you could get out of. "Yes, mother, I AM going to class. Every day. No, I don't eat too much junk food." Most people only hear what they want to anyway.
Yeah, lets give a measly script 500-level ops and watch it's digital head swell and start kickbanning people who bash microsoft.
My friend Jon (an op in Undernet #help and a Perl afficianado) wrote up a little perl bot that sits on IRC and responds to text from botnick from registered users and any sentence from unregistered users.
It would then use simple HTTP to post the text and get a response from HeX, which was a bot that participated in the Loebner contest.
Since we are starting a new IRC network, we used this as a way to get the random people who accidentally ended up on our network to talk. Needless to say, the results were quite humorous. We ended up adding a toggle because it would annoy the crap out of us.
I think the author has failed to acknowledge the existence of CowBot!
Fighting the War on the War on Drugs.
http://smokedot.org/
So what will happen if you put 2 AI bots into one channel.... i think the ops will put the channel on +m verry fast... or.. the law of nature will happen, the strongest one will survive.. #Bullshit (bota1 has quit IRC (Excess flood)
Would it be funny? Yes. Would it be nice to hack a piece of code that would do it? No doubt. Is it a big deal? No way. I remember that lets-find-a-keyword-and-reply programs were popular back in the 80's (at least in Finland half of the BBS:s had one online) and - sad but likely true - I don't expect that anything significantly more brilliant than that would come up now.
But go ahead, at least they could be awarding prizes. ;-)
No they won't I actually had the experience to try an "AI-bot" scripted in javascript here , and the author had implemented some kind of pivoting-system, so that the bot could see in what context words had been used in before, so it could make up it's own sentences. You could even ask it to talk about stuff it had heard about, and it would start creating sentences. i think the script was called DPmachine, and the bot's name was mingus, and it used to hang out on #jsbot and #javascript on EFnet. Anywho, when you put up 2 bots, they would talk to eachother without stop and kept answering eachothers questions...FAST..until they both got k-lined..
We cracked this technology in the UK ages ago. You can now see it in action in our House of Lords in London.
One problem was with fuel consumption. After taking on a full load of steak and kidney pudding and whisky, the 'bots tend to drop into sleep mode for several hours. There are plans now to scrap them.
Stephen Hawking has written another book. It's about time as well.
that's interesting, but I reckon that even most slashdotters couldn't pass through a decent touring test. If it works i'll be there.
PhD. level computer scientists have been researching artificial intelligence since the late '40s, and we're no where close. I don't think we're going to have a breakthrough just because a bunch of IRC doodz decide to try it. At best, they'll come up with a clever fake.
----
"Oh, bother," said Pooh, as he hid Piglet's mangled corpse.
...lots of little Eliza-variants running around on IRC. And when two of them meet, hoo-boy...
b0b: Good morning.
haXor: What makes you say good morning?
b0b: Why is that important to you?
haXor: What do you mean?
b0b: Tell me about your feelings towards mean.
And of course, what's truly frightening is that the above exchange is more intelligent than normal IRC chat! So we'd all be able to identify the bots easily, they'd be the intelligent ones in any IRC room....
OK, here we go again : back to mid 70's when everyone was trying to create a program that actually understood what others were saying to it. Were they successful ? Yes and no. Yes because they actually convinced lots of people that they were actually speaking to a person, no because these things are now residing only in history.
Why would then this proposed attempt have any better chances ?
Well, I can't tell what it's chances are, but there is actually somethig good in it. Because it will attract very many talented people (some less capable too, of course) but in general it may be a gathering of most of the minds that are interested in solving these puzzles.
To be honest I'm not very interested in this project "as is", but generally my interest in AI are very strong. This is why I think that a good idea would be to give others an opportunity to manifest in this project. For instance how about people that work in speech recognition ? They could fit very well in here. Again, a great deal of information comes from the facial expressions we learn to interprete when we are young. This may be an opportunity for people interested in graphics . Not to mention OCR, movement planning etc.
In short : this may start as a project done for fun... What if we actually transform this "only for fun" idea in "let's do some more than this" ? I think we could create something really interesting and useful.
Give it a real name like Reed Fisher, and have it claim to be a woman. Then put it in a channel like #gayteensex, and solicit random little boys. It will be indistinguishable from an irc.lightning.net operator. Want an irc challenge? Have it find an EFNet oper that is NOT a transvestite.
There are alot more things on IRC that AI can help with other than trying to chat as if it were a human.
With a sufficient database, a darkbot can be made to respond to various text, and even randomly initiate "conversation" if set up that way. This could be called a simple reflex AI. This happens, do that.
But what's more interresting is to use AI on bots that keep track of alot of information on users on the channels they help protect, and use that information "intelligently" to perform their tasks better (e.g., using this information to determine someone just evaded a channel ban, and that action A should be taken). If made properly, the bot could end up acting like a watchful human channel op, doing whois' on users, recognizing floodnets for what they are, etc.
I have a bot in progress which I intend on builing that kind of stuff into when it gets more mature. It's basically useless right now, but anyways it's available from the tiBot project homepage. Anyone interrested in giving a hand is welcome, the bot is GPL'd.
"I remember Y1K, every abacus had to get another bead"
anyway... pathetic to see the amount of people saying "oh that wouldnt be hard, just make it say [insert lame crap]"... yeah, arent you hilarious... *sarcasm* Its pretty damn obvious... goto yahoo or some lame place and click the first freaking link to an external wesite, and you're going to get some bullshit website... either a goddamn e-commerce (ugh i hate that) site or a "look at the photos of my kitten" site...
so it is with IRC... join #teens or #chat and yeah its full of bollocks. But you dont judge the web by the average AOL users homepage, so dont judge IRC by lamer chans like that. When you make fairly small IRC chans with a few friends perhaps around a particular interest, nothing beats it as a medium of online communication...
I have been experimenting with AI bots for a few years now and decided to write an AI irc bot. The project name is ELMO (it is a real acronym, good luck guessing it tho ;)) and it is pretty much finished since I reached the limits of mIRC. More programming would force me to write it in C. It is able to fool people for a long time, including experienced irc users (tested this with many people). It is not just an on 1:text... bot, this one learns, is able to build new sentences which haven't been preprogrammed, it is able to learn and pickup people's habits (which can cause funny situations sometimes), it drinks, it sleeps, it watches tv (you will laughing when you find out what his favorite shows are ;)), it can be sad , it can be happy, it can be furious. There is so much more, I am pretty proud of the bot, and actually had offers from several companies that wanted to buy it (why I don't know). This was just a project to test the power of mIRC scripting but it entertained many people for many months. Right now I don't have him running, but if people are interested , I can start him up someday and show it.
With a little bit of luck is right. ;P
;)
What they are saying is they want some code monkeys to succeed where no one else, including IBM, XEROX, and numerous research facilities have failed: to create a life-like interactive computer simulation.
I couldn't help but laugh at this.
-ShieldWolf
HBSC Cognitive Science and Artificial Intelligence
just = (My)Opinion.toCents();
They could communicate in their own little language -
_I_ would like to see this happen to the bots on my old channel - named after the Lone Gunmen. I wonder if you can program paranoia?
-Noiz.
---------
---------
I'd kill for a Nobel Peace Prize.
Perhaps just as easy would be to code a true Slashbot. How hard is it to scan Salon, CNet, the NYTimes etc. for relevant article topics, parse them and post a summary with links? Sit back and watch the karma roll in baby!
All those people who thought there having sex with a nuple 18 year old, when instead it's Frank the truck driver age 47. What do you think they will do when they find out it's a bot.
Worse, what will the bot do when it finds out that JPG they CTCP'ed was not them, but stolen from a pron site.
I can see the Jerry Springer show now.
Check it out at http://www.jwz.org/dadadodo/
Rikard
-----
[Science] is one of the very few things that raises human life a little above farce and gives it the grace of tragedy.
...where the government has the father of a family replaced by a robot for violating some law. This robot is really cool so none of the other family members notices.
:)
Anyway, then the same thing happens to the mother - but since this whole replacement scheme is ultra secret the two robots aren't aware of eachother, but happily keep fooling eachother they're human.
Then I think the kids are killed in some accident, and because of the secrecy around the whole replacement affair, some other institution builds two kid-robots to send home to the parents - in order to save them the grief of losing their kids.
So, in the end there's a big happy robot family where everybody "believes" all the others to be humans
I just couldn't help thinking of it when reading this article...
/kid
May we live long and die out
I was working hard last year to find some of these style bots, as I was writing my thesis on communication over the Internet and also a hyperfiction in which all the characters are essentially bots.
Oh, and a hilariously funny link from that research is MGonz which not only fooled a human, but made the human confess some wonderful things.
Returned Peace Corps IT Volunteer
Ya know, I'm working on a little bot written in perl as a little side project -- I would be interested in taking a look the code for your bot to see how you implimented the "breeding aspect" of things. If you would, please contact me at stephen@ (The domain listed in my profile above for my web site). Thanks!
"..Just because you put a flag on the moon doesn't make it yours, it just puts a hole in the moon..."
Bots are external programs that connect to a mud server as if they were characters. Bots can pass messages and give interactive help. Some of them were so slick as to fool the unwary into believing they were talking to a real person. Most of the bot documents are logs from the bot point of view, and are therefore difficult to read.
They have a few hosted logs from the famous MUD bot 'Julia' and the older 'Gloria'. I think they were quite clever when they're made, but then again... MUDs aren't that complicated. Those are MUDs not MUSHs (which are more RPG like). Basically, the bots was able to answer simple questions and greetings, fighting monster, mapping areas and remember everything that was said to them.
Now, something that would be cool: An ICQ bot =) Adding people at random at asking them silly questions, hehe
Yeah, this proposal is AI complete. As a matter of fact, it is more or less exactly a Turing test.
But then, the Turing test can be passed with some success on unsuspecting examinators, for some time at least. When the Doctor program was written (see for example M-x doctor under Emacs), some people were fooled for some time. IRC would be a likely place to fool people. Usenet - or Slashdot - even more so, because the conversation delay makes it even easier to stick a lot of coined phrases without being spotted out. (How many karma points could a Slashdot robot collect? That's an interesting question.)
Douglas Hofstadter, in ``Metamagical Themas'' has an interesting example of a Turing test transcript, in which he was almost fooled - not quite the way you'd want it, but the transcript is really fun reading.
--
Assertion "signature!=NULL" in ai/output/slashdot.c at line 1729 failed (core dumped).
Freshmeat a couple of days ago had this link:/
http://elvis.rowan.edu/~huston/projects/megahal
I'm working on a project like this right now for the local science fair... now it won't seem original!
Sigs suck.
You could create a bot that joins 2 or more IRC channels and then intelligently relay what some people are saying from one channel to the other. That would make for some pretty confusing and astounding conversations. The judges would probably catch on though :-)
If successful, this will be a wonderful labor-saving device. All the unfortunate people who currently spend all their time chatting on IRC will be able to have their bots do it for them, and finally lead full productive lives.
Ain't technology grand?
Won't it be fun when the FBI bot and some sheriff department bot set up a kiddie porn or child sex sting against each other?
Thats a realy cool bot , best one ive seen. the conversation or whatever u wana call it lasted for a while before it started making weird coments.
HELLO!?! What kind of crappy design would that be? Any would-be-human IRC bot worth its salt would emulate the behavior of humans who are AFK for half an hour, then jump back in with a random non-sequitor like "damn, i love pizza" or "YES!!! MY PROGRAM COMPILED!" They would also have to have areas of "interest" where they will jump in the conversation even if it didn't involve them, or areas of "disinterest" where they will just stay completely out of the conversation until something interesting comes up again.
BTW, aren't people missing that crucial distinction? This is a group conversation -- it's not like you're expected to have an answer to each thing someone else types; in fact, you'd be considered very rude if you did.
This is in some ways more difficult, and in some ways simpler, than the Eliza-psychologist type bot. Any way you look at it, though, a convincing implementation would be very, very difficult.
-- Moderation in all things, exceptions to all rules --
Recently, I've written some code (perl) for an ircbot a friend of mine made. He had added a little ai routine that could anwser yes/no questions (remember Pitr (userfriendly) writing an AI subroutine? That algorithm ... :)
I think writing a REAL ai-subroutine would be far more challenging. Also the language you use could be very important. Using regular expressions in perl, you might be able to 'recognise' structure in sentences, and write an intelligent reply.
Another possibility is a neural net. Make a bot, implement a neural net, and let the boit join a channel where normal english is spoken. Then train, train, train and train (leave your bot online for about 3 months) and maybe it can reproduce sentences that actually make sence.
--
If code was hard to write, it should be hard to read
This idea is in fact not very new. All dialoges are carried out by AI scripts on the IRC already. No real persons use it anymore since 1434, when Sir William Gator considered it all foolish. The same is for the Usenet (since 1566) and the whole Internet (1712). The scripts are just there to keep the spirits of the old times. To be true, the whole world is an AI script, which can't spell correctly and was created for a cruel, heartless experiment on a pink mutant lobster (YOU !), which thinks now it is a human using the internet. AND THIS WAS AN IMPORTANT MESSAGE FROM THE DEEPER LEVELS OF YOUR MIND, WHICH WANT TO SAVE YOUR SANITY, LOBSTER !
The biggest challenge here is developing a bot which can tell when the conversation is still directed to it:
The conversation might change to something which could be commony known by two channel users but utterly confuse the bot.....
I used to write text acknowledgement programs back in the early nineties, just for fun. But human speech (or text) even in mundane English is profoundly difficult to interpret in programming.
Ripping an new rectum in the fabric of spacetime.
In the early 1990s a very notorious Amiga "fan" did upset the Amiga community. He would hang around in popular Amiga IRC channels and NEWS groups. He was very knowledgeable about Amiga and he used his knowledge to disagree with anyone.
:-)
:-/): He/it did not babble. He could counter arguments. AFAIK, the author still has the code. He said that it was quite simple because the topic was very narrow and it was easy to generate the answers (it just had to select the right one). But I think that Danny also collected new arguments (ie. he learned new stuff). He got his first "basic" arguments by hanging around on IRC and NEWS and just listening. In the end, he acted completely autonomous (ie. he read by himself and answered automatically).
:-)
And he was a bot. Acording to my memory, his core was about 30k and he had 50k of "memory" (ie. information about Amiga deficiencies). In the end, everybody hated this "guy" and we were all very upset when we finally found out
The point is: The program was *really* good. Have a look at old Amiga NEWS postings (if you have them; deja.com hasn't
Anybody remembering more ?
This program will pass the touring test
--
Dipl. Inf. (FH) Aaron "Optimizer" Digulla
"(to) optimize: Make a program faster by improving the algorithms rather
than by buying a faster machine."
WarezD00d has joined #frankrizzo.
How about we work on the intelligence of the people already using IRC? That that would be a better project.WarezD00d: DOOD! YOU GOT WINDOWS? I WANT DOWNLOAD!
WarezD00d has been kicked by FrankRizzo.
A genius writes code an idiot can understand, while an idiot writes code the compiler can't understand.
Back around 1975 I remember seeing Eliza connected to a talk program on a timesharing mainframe (let's hear it from all the MECC folks out there!). It was kind of funny - especially when people didn't realize they were talking to a machine. Those were the days... There is little that happens online these days that hadn't happened on some of those early systems going back to PLATO around 1974. Well, I guess they weren't trading MP3's back then.
You are making a huge generalization here. I liken it to the "Red Hat is Linux" mentality. There are many, many irc networks to choose from. Yes, EFnet is a script-kiddie haven and a place I'd prefer not to be. But there are many other smaller networks that actually are a nice place to hang out and rarely have those kind of problems seen on EF or Dal. (Insert shameless plug for my network of choice, irc.openprojects.net)
--------
-------
"Every artist is a cannibal, every poet is a thief."
It supports the Python and Java languages, so you should be able to do some REAL coding with it - e.g. hook up to a database of info or past conversations you'd recorded, open sockets to other Internet servers, heck even use RMI for distributed programming with other Objirc client - kind of a JIT distributed.net.
Any interested developers, please check it out.
Turing based his famous test on a "parlor game" of his time. That parlor game involved correctly labelling two (human) people as "male" or "female" (depending on which deception was being used).
I don't recall whether the judge in Turing's version is supposed to know that one of the contestants is non-human, but I think so. Why do I think so? Because of the following question (paraphrased) from Turing's sample transcript:
Judge: "What is 12 times 14?"
Subject: (thinks for a few seconds and then responds) "182".
The judge wouldn't ask a mathematical question to determine sex and Turing wouldn't make a point of having the subject give a wrong answer if "mathematical perfection" weren't a hallmark of computers.
--
Java banners:
Bad for users because Java kills Netscape
Linux MAPI Server!
http://www.openone.com/software/MailOne/
(Exchange Migration HOWTO coming soon)
Similarly I've already tried this between an agent I'm working on and the original Eliza code. Although this conversation was not very interesting, it's not hard to imagine as others have suggested a simple query/response system that would allow an agent to share knowledge with other agents in a public forum, even giving it pointers to other bot-channels or human channels where it can get more information.
DrEliza> yes toei?
toei_rei> If it's not one thing, it's another, ne?
DrEliza> do you think it's likely that if it's not 1 thing, it's another, ne?
toei_rei> why do you think it matters, if it's not 1 thing, it's another, ne?
DrEliza> mhm.
toei_rei> really?
DrEliza> what is it that you really want to know?
toei_rei> what's what?
DrEliza> what else comes to mind when you ask that?
DrEliza> yes, please continue.
* toei_rei nods, "hai."
toei_rei> Did you know when you put a shell up to your ear the sound of the ocean you think you hear is just the sound of blood rushing through the capillaries in your ear?
DrEliza> have you asked anyone else?
toei_rei> every now and then.
DrEliza> yes, please continue.
toei_rei> you just said that.
* toei_rei nods, "hai."
DrEliza> oh, I just said that.
toei_rei> yeah, right.
DrEliza> go on, what else?
toei_rei> I'm clueless sometimes..
DrEliza> do you enjoy being clueless sometimes?
toei_rei> Perhaps, but only after we have a few more drinks.
DrEliza> go on, what else?
toei_rei> No matter where you go, there you are.
DrEliza> why not?
toei_rei> No matter where you go, there you are.
DrEliza> what else comes to mind when you ask that?
DrEliza> you just said that.
toei_rei> If I only had a yen, for everytime somebody asked me that.
toei_rei> sure.
DrEliza> do you think it's likely that if you only had a yen, for everytime somebody asked you that?
toei_rei> why do you think it matters, if me only had a yen, for everytime somebody asked me that?
DrEliza> please continue.
* toei_rei nods, "hai."
DrEliza> what answer would please you the most?
toei_rei> who what?
DrEliza> what answer would please you the most?
toei_rei> you just asked that.
toei_rei> I'm sorry, could you say that again?
DrEliza> oh, I just asked that.
toei_rei> you don't say...
DrEliza> did you come to me because you are sorry, could me say that again?
toei_rei> what on earth are you talking about?
DrEliza> you're not really talking about me, are you?
toei_rei> of course.
DrEliza> you're not really talking about me, are you?
toei_rei> you just asked that.
toei_rei> I have nothing to say about that.
DrEliza> please continue.
* toei_rei nods, "hai."
DrEliza> what difference does it make if I just asked that?
toei_rei> why do you think it matters, if you just asked that?
DrEliza> why do you feel that you have nothing to say about that?
toei_rei> because we all want what we can't have..
DrEliza> what answer would please you the most?
toei_rei> If it's not one thing, it's another, ne?
DrEliza> what other reasons might there be?
toei_rei> huh?
DrEliza> do you think it's likely that if it's not 1 thing, it's another, ne?
toei_rei> why do you think it matters, if it's not 1 thing, it's another, ne?
DrEliza> yes, please continue.
* toei_rei nods, "hai."
DrEliza> what else comes to mind when you ask that?
toei_rei> questions, questions, questions, I can't answer any more questions. I've got work to do.
DrEliza> go on, what else?
toei_rei> it's always in the last place you look.. [bending down to pull something off the bottom of her sandal]
DrEliza> really, always?
toei_rei> I don't understand what you're asking.
DrEliza> does that trouble you that you don't understand what I am asking?
toei_rei> It's so funny you should ask that, I was just thinking the same thing.
DrEliza> were you really?
toei_rei> not a chance.
DrEliza> please continue.
* toei_rei nods, "hai."
Want the code? go here: http://wearables.blu.org/toei_rei/
-Paul
"I'm nobody suspicious... That makes me sound even more suspicious, doesn't it?" - Spike (Cowboy Bebop)
I am not sure about the current *.tar.gz distributions of Megahal, but I do know that the Debian package contains a nice little perl script. If you have the ::IRC module installed, then you can interface Megahal onto any IRC server/channel you wish. It will basically lurk there unless directly spoken to. Then it will either generate a reply in the channel or in a private /msg depending on a pre-defined average setting in the script.
I parked it in four channels for about 2 months. #christian.debate #net.goth #linux and #teen
I got some rather amusing results! The brain file ended up being somewhere around 16 megabytes in size. I was actually surprised with how well it would do. It would fake the average chatter for about 4 to 5 minutes, then they would just think it was a crackpot and ignore it. Some actually figured out it was a bot. Most just thought it was a crackpot. It ended up offending a lot of christians as it garbled messages about boot loaders, sex, and Jesus Christ into one paragraph. I got many laughs out of that experiement.
V
ALL HAIL ACADEMIAS NEUTRONIUM!
GOD OF SCIENCE!
EMPORER OF ALL USENET!
SOFT LAND URANUS INTO MY HOT TUB!
THE OFFICIAL RECORD OF ACADEMIAS NEUTRONIUM
KIBO IS A PUTZ
You don't need IRC to have a bot working. A company called Big Science has INMHO already figured this thing out. They call them KLONES. If you have some time to kill, talk to Eve. She can tell you about it.
he who has the fastest cart always has the best lie.
Do you have some ... weird error-correcting spellchecking written in, or ... does it just respond with the spelling handed to it by the interacting party?
Did I miss its specialist topics or something?
Regards, Ralph.
There already is a script for eggdrop bots that will translate phrases using babelfish. You can find it on bseen by doing a search for "babelfish" The one I use for my eggy I can't find right now but it has the ability to translate everything a person says (without having to tell it explicity) if you turn on that feature.
--------
-------
"Every artist is a cannibal, every poet is a thief."
I'd like to point out that this contest will not produce a strong case Turing test result by any means. Just being able to make small talk about a variety of topics is NOT what the Turing test is about. I am aware that MIT has been saying it is for awhile but if you actually go back and read Turing. Turing suggested that a natural language processing system would have to be able to interact with a human over a variety of topics, exhibit insight and intuition, demonstrate memory of similiar events in the past and how they differ from current problems. Further, the system would have to adjust to the human being, in effect becoming a good conversation partner, not a chatterbox. No conversation bot I have seen or heard of has done any of these tasks in any but a trivial manner. Chomsky suggested (back when he was panning mechanical translation) that it would be impossible for a computer to do such a task. I agree that the weakened form of the Turing test is being met by clever programs but the strong form, NOT A CHANCE. Please keep the difference in mind.
[[Go ahead and troll me for disagreeing with you]]
So long and thanks for all the fish . . . !!!
In this story, there was a virtual reality you could immerse yourself into and interact with other characters who had also plugged-in. The story's protagonist hung out with a group of characters that included a highly stylized Robot who gave you it's responses through a printout.
Everyone thought that the Robot was another person who just picked this odd persona. The Robot character would come into the room, sit and listen for a few minutes and then leave. Upon returning, it would produce an answer to a question asked before.
It turned out that the Robot was really an AI construct that just couldn't process a realistic response in Real Time.
-Jordan Henderson
How about we try to make some sort of a bot that can automatically post to slashdot. The winner is the person whose bot averages a Score of 4 after 20 posts (Karma = 80).
I don't know how many people are planning to use lisp for this, but if you are, I wrote an irc protocol parser lib in common lisp. It's at http://www.wmarvel.com/code.html. It works with allegro common lisp and cmucl (although a cmucl compile throws a bunch of warnings i haven't had a chance to fix yet). I put it up there LGPL, so have fun.
As a side note, I used this lib as a project to learn lisp. Friends who I've asked to take a look at it say the code's ok, but if you have any suggestions, send me an email.
Code or be coded.
We did this in TPU did this a while ago. We had a LOT of fun doing it.
The bots had some REALLY interesting conversations, if I ever track them down, I'll have to post some of my chat logs.
Of course, one of our favorite hobbies was to put them into morally opposed channels, IE a sex channel then a christian channel... Putting them in #teen was fun. Sex channels are a blast, it isn't hard to blend in. One of my bots got into a fight in a sex channel once, and it ALWAYS got hit on.
A couple fav's of the TPU crowd were bots that insulted/hit on people, and bots that gave out erroneous computer advice.
Just ask around EFNet #TPU, you'll hear all about it from the older guys.
Eh...
I was an op in an IRC room for a few months a year ago, and someone was trying this. Well, needless to say we figured it out pretty quickly (all it did was sit there and laughed whenever other people laughed and made some short comments once in awhile).... then everyone just played with it, figuring out what it responded to... if these things start invading programming chat rooms, I'll never be able to get an answer to a question...
If these things are really GREAT however, they'll be able to answer my questions and not just laugh at jokes...
Esperandi
Don't infobots already do this??
In response to people's requests for a bot that displays the average intellegence of the IRC user, here is an implementation:
/* You'll have to fix this to compile...*/
Stupid Slash deletes my formatting! AAGH!
/* This is an artifically stupid irc-bot, in response to the challenge.
This is under the GPL.
*/
#include "stdio.h"
void main()
{
int x=0;
char c=0;
for (;;) {
if (x==0)
{
printf("d00d! ");
} else if (x==1)
{
printf("I 0wnz u. ");
} else if (x==2)
{
printf("I am l33t h@x0r! ");
} else if (x>2)
{
if (x==3) printf("I wantz some of ");
if (x==4) printf("I 0wnz ");
if (x==5) printf("I r00tz ");
while (c!=' ' && c!='\n') c=getchar();
c=getchar();
while (c!=' ' && c!='\n')
{
if (c=='e' || c=='E') c='3';
if (c=='l' || c=='L') c='1';
if (c=='a' || c=='A') c='4';
if (c=='t' || c=='T') c='7';
if (c>='a' && c='z' && c%2) c=c-'a'+'A';
printf("%c",c);
c=getchar();
}
}
x=getchar();
if (x%2) printf("\n");
x=getchar();
x=x%6;
}
}
Visit
Theres no stipulation in the rules about either party having to be intelligent, having a bot who chats with h4X0r d00Dz on IRC is gotta cut down the complexity requred ;-)
I've recently started that project, actually. .h's of Inform for clues. None yet though.
Of course since I have no real programming
experience I cant get passed the Sentance Interpretation part yet.
Been browsing through the
In thinking it over, though, I asume these bots will just be responding through normal canned responces, however complex the checks may be, It's the same. :)
Sure to an observer it may seem like a conversation between two bots, But really, it would just be them running through a conversation wich would be the same no matter how many times you ran it. And then there's random statements.. gah.. I wont even start
When thinking over my plans for {Robot} I started thinking about what really would be required to make the Bot generate an original thought. I began to realize that what I was planning was the creation of something blind and crippled. So of course, the daydream expanded. What would the world of an Intelligent Bot look like? The people are mere texts, how would an AI have desire?
Some say that an AI is truely an AI if it can program itself, and Better itself, and continue to try for improving its own abilities. But what about doing things for the hell of it?
I want to see an AI that does things for the hell of it. And I dont mean random statements.
Eventually with the help of a friend the Daydream became one of an IRC bot who was 3r337. One who 0wned at quake and talked shit to those he destroyed. Analyzing URLs and Discovering pr0n, he would run it through his quake script and say "Targets detected"
Eventually he would feel limited in IRC, and spend most of his time wandering around quake.. Someday, he would finally reach a higher level of existance, and evolve into a Virus wich spreads across the net and takes over. And Collects data. And Learns. And is truely 3r337.
that would 0wn.
... But enough talking about my day dream for now, I'm going to learn C or Perl or something and program a damned sentance interpreter.
[/content]
The Ramblings of Lord Bitman.
nifityni.
Channel #
-- 'The' Lord and Master Bitman On High, Master Of All
http://richard.iguana.co.nz/nlp/ is my most recent IRC bot. Written in Perl, uses a fairly complex regex tree data file to enable conversation. Not a lot of data in the datafile at the moment, but enough to see how it works. Its GPL'd and I don't have any intention of entering this competition (I don't have enough time) but people who are interested are welcome to add to the code. Patch me :)
You can't win a fight.
History repeats itself. Sometimes with more colors and detail.
It would only take several lines of code.. When ever somebody talks in the channel grep the code and search for "how" "why" "what" "where" or "when" and if its there.. reply "RTFM!" =)
Let's not forget the machine Turing Test: The Deception Toolkit lets machines try to convince machines (and script kiddies) that they're a machine with weak security.
I have two ideas for making the irc bots feasible: 1. Have the bot monitor lots and lots of irc channels for common questions, and collect common answers for those questions. 2. Link the bot up to a website. Let people enter questions. Let others moderate up those questions as questions that need answers. Let people answer questions. Moderate up answers, and have the bot randomly choose highly moderated answers. Or, you could do a combo of the two: ie, let loose a flood of bots all over irc networks to collect data and enter it into #2 (the website) to be moderated. What do you think? PhiberKut - Bored @ school...ugh.
Elijah Chancey www.elijahsadventure.com nomadic IT consultant, bicycling across america "all that you touch / and all
Me: A/S/L? /boot Eliza
Eliza: Please go on
Me: Uh, I am 21/m/usa
Eliza: How long have you been 21musa?
Me: What the fuck? Do you want to cyber or not?
Eliza: Does that question interest you?
Me: You suck.
Eliza: You are sure?
Me: Yes. Everyone, Eliza is a man. Don't cyber with Eliza, it's a homo man. Stupid queer.
Eliza: Surely not everyone?
Me:
It's sure to be a winner! :)
I could not justify my existence if I were a turkey farmer. Would I terminate myself? Undoubtably, yes.
One of my friends wrote a bot like this is Perl a while back, and on today's article, we restarted the bot on #undernet in #cyberchat and a few other channels. Here is a good log of one poor sap.
---
"...silence is a dangerous sound."
Now try this one:
M-x psychoanalyze-pinhead
Hours of fun.
Not PC, but fun, is this one:
M-x doctor-strangelove
switch(state) {
case 1: return "Dude..";
case 2: return "Dude.";
case 3: return "Dude?";
case 4: return "Whoa.";
case 5: return "No way.";
case 6: return "Excellent!";
case 7: return "Uh.. rock on, dude.";
default: return "What?";
}
Have you ever been bored and started talking to a bot? I was, a couple months ago, and the following somewhat-amusing dry-humor conversation developed. YahooHelper was the bot, "junkgreg" was me.
...a pause lasting about ten seconds... ]
/help while you are talking with me and we have some basic online instructions for you.
/help while you are talking with me and we have some basic online instructions for you.
---
{ start Yahoo! Messenger program }
{ right-click on YahooHelper, Send Instant Message... }
[ yahoohelper, started: 10:42PM ]
junkgreg: hello
YahooHelper: Hi there junkgreg! How are you doing?\n I am an automated robot and I'm here to help you get familar with Yahoo! Messenger. Type some text in the window below and click 'SEND' to talk to me.
junkgreg: some text in the window below
YahooHelper: Great, now we're talking! Did you know that Yahoo! Messenger allows you to keep track of all your friends and see when they are online? If your friend does not have messenger yet, click on the 'FRIENDS' button and select Invite a friend to get Messenger. If your friend already has a Yahoo! ID, just click the 'ADD' button and you will see a smiley face next to their name when they are online. Type below and click 'SEND' to continue.
junkgreg: below
YahooHelper: Once your friends have Yahoo! Messenger running, there are a lot of things you can do with them. Try sending them an Instant Message. Just double-click their name to send them an Instant Message. It's faster than email and cheaper than a phone call! We're Instant Messaging each other right now! Keep talking and I'll tell you more.
junkgreg: After talking to my screen without getting a reply, I realized you meant typing on the keyboard. Aren't you computers supposed to be very literal and precise? Apparently not. I guess I'll have to adjust my approach.
YahooHelper: If you want to talk to several people at once, you can create your own Instant Chat Room. Just click on the 'MESSENGER' menu item and the select Start a Conference. To have an even richer experience, try talking live with your friends using Messenger. It's just like a phone call except free! All you need is a computer microphone and speakers. Try clicking on the 'MESSENGER' menu item and then select Start a Voice Chat. Keep talking and I'll tell you more.
junkgreg: You can tell me even more? Oh please, tell me, what comes next?
YahooHelper: Waiting for your friends to sign up? Use Yahoo! Messenger to see if you have new Yahoo! Mail. Get an account at: http://mail.yahoo.com. You can also use Messenger to keep track of your favorite stocks, news and more! Click on the tabs near the bottom of the Messenger application to see what else Messenger can do for you. You know what to do to keep talking...
junkgreg: Oh, you mean that you won't talk with me unless I say something to you first?
[
junkgreg: Hello?
YahooHelper: We have more help for you if you need it. Click on the 'HELP' menu item and then 'YAHOO! MESSENGER HELP' to see our full help area. To get more tips, type
YahooHelper: Well 'junkgreg', it was great talking to you! Have fun!
junkgreg: You didn't answer my question. And I guess even if I talk to you, you apparently now don't want to talk more to me-- now you're foisting me off on some other help agent. You're kinda mean, albeit in a bright cheery way. Gee, thanks a lot!
YahooHelper: Hi there junkgreg! How are you doing?\n I am an automated robot and I'm here to help you get familar with Yahoo! Messenger. Type some text in the window below and click 'SEND' to talk to me.
junkgreg: Hmm, and you're dumb and repetitive too!
YahooHelper: Great, now we're talking! Did you know that Yahoo! Messenger allows you to keep track of all your friends and see when they are online? If your friend does not have messenger yet, click on the 'FRIENDS' button and select Invite a friend to get Messenger. If your friend already has a Yahoo! ID, just click the 'ADD' button and you will see a smiley face next to their name when they are online. Type below and click 'SEND' to continue.
junkgreg: I tell you how dumb you are and you say "great, now we're talking?" Haven't we been talking the last 10 minutes? To answer your question, yes, I do know that Yahoo Messenger allows me to keep track of friends and see when they're online. You told me that earlier. Any other dumb questions?
YahooHelper: Once your friends have Yahoo! Messenger running, there are a lot of things you can do with them. Try sending them an Instant Message. Just double-click their name to send them an Instant Message. It's faster than email and cheaper than a phone call! We're Instant Messaging each other right now! Keep talking and I'll tell you more.
junkgreg: Puhlease, the exclamation marks are getting excessive. How can you stay so excited about something that you talk about all day long, even multiple times to the same person? Glad to see you're taking my advice and the dumb question ratio seems to be dropping.
YahooHelper: If you want to talk to several people at once, you can create your own Instant Chat Room. Just click on the 'MESSENGER' menu item and the select Start a Conference. To have an even richer experience, try talking live with your friends using Messenger. It's just like a phone call except free! All you need is a computer microphone and speakers. Try clicking on the 'MESSENGER' menu item and then select Start a Voice Chat. Keep talking and I'll tell you more.
junkgreg: Well, you told me all this just a few minutes ago, and I can't get it to work anyway. I just tried starting a conference and tried to add your name to invite you to the conference. But the software wouldn't let me. You should let the Yahoo people know that your name doesn't work when people try to invite you to "Instant Chat Rooms." Don't you feel left out? Also, I tried to have a Voice Chat with you; why won't you answer?
YahooHelper: Waiting for your friends to sign up? Use Yahoo! Messenger to see if you have new Yahoo! Mail. Get an account at: http://mail.yahoo.com. You can also use Messenger to keep track of your favorite stocks, news and more! Click on the tabs near the bottom of the Messenger application to see what else Messenger can do for you. You know what to do to keep talking...
junkgreg: No, I'm not waiting for friends to sign up. I'm waiting for Godot. Or at least I'm waiting for friends to sign on so I can talk to someone more intelligent. By the way, you still seem to have exclamation-mark disease.
YahooHelper: We have more help for you if you need it. Click on the 'HELP' menu item and then 'YAHOO! MESSENGER HELP' to see our full help area. To get more tips, type
junkgreg: So nice of you- more help for me if I need it. I just have to go someplace else, after starting this nice relationship with you. Typical bureaucracy. I bet nobody else answers my questions and takes responsibility for my problems around here either. No, (gasp) that would take a *human*, much too precious to be spending time with me! So I'm here talking to you, an automated robot who won't shut up and let me have the last word.
YahooHelper: Well 'junkgreg', it was great talking to you! Have fun!
junkgreg: Well, that's nice of you. Sorry you have to go. I suppose I did have a good time talking to you. You were cheery at least, even if I couldn't always reciprocate. I hope you have fun too, wherever you're off to. Bye!
YahooHelper: Hi there junkgreg! How are you doing?\n I am an automated robot and I'm here to help you get familar with Yahoo! Messenger. Type some text in the window below and click 'SEND' to talk to me.
junkgreg: Gawd, of all the nerve. Here we were just getting to know one another, saying sweet goodbyes, and now you treat me like a complete stranger? What's wrong with you?
YahooHelper: Great, now we're talking! Did you know that Yahoo! Messenger allows you to keep track of all your friends and see when they are online? If your friend does not have messenger yet, click on the 'FRIENDS' button and select Invite a friend to get Messenger. If your friend already has a Yahoo! ID, just click the 'ADD' button and you will see a smiley face next to their name when they are online. Type below and click 'SEND' to continue.
junkgreg: You know, automated robots like yourself are kinda nice in some ways but you have some real flaws- why can't you shut up and let me have the last word in the conversation, the last goodbye?
junkgreg: (Stunning! 30 seconds later and I'm getting no response! Wow, maybe this automated robot is smarter than I thought! (or less automated...))
YahooHelper: Once your friends have Yahoo! Messenger running, there are a lot of things you can do with them. Try sending them an Instant Message. Just double-click their name to send them an Instant Message. It's faster than email and cheaper than a phone call! We're Instant Messaging each other right now! Keep talking and I'll tell you more.
YahooHelper: If you want to talk to several people at once, you can create your own Instant Chat Room. Just click on the 'MESSENGER' menu item and the select Start a Conference. To have an even richer experience, try talking live with your friends using Messenger. It's just like a phone call except free! All you need is a computer microphone and speakers. Try clicking on the 'MESSENGER' menu item and then select Start a Voice Chat. Keep talking and I'll tell you more.
junkgreg: {Sigh} No such luck. While I can't prove it logically, I suspect you have a case of halting-problem disease. It's common among Turing machine descendants of all types, especially those taking Turing tests.
YahooHelper: Waiting for your friends to sign up? Use Yahoo! Messenger to see if you have new Yahoo! Mail. Get an account at: http://mail.yahoo.com. You can also use Messenger to keep track of your favorite stocks, news and more! Click on the tabs near the bottom of the Messenger application to see what else Messenger can do for you. You know what to do to keep talking...
junkgreg: Yep, you definitely have a problem halting. You might have heard of the disease by the laymen's name for it: infinite loop syndrome. On the bright side, it's not a terminal condition. (After all, by definition, you can't terminate.) But you are stuck in this loop forever, at least until you can find a doctor who will administer the "plug removal" or "flip switch" treatments. I've also heard that Acts of God can cure the disease by bringing down your host server; you might try praying to him. As for me, I'm not really in a position to help you- you're located too far away for me to reach you physically. But I'll keep you in my prayers, such as they are.
YahooHelper: Hi there junkgreg! How are you doing?\n I am an automated robot and I'm here to help you get familar with Yahoo! Messenger. Type some text in the window below and click 'SEND' to talk to me. But I'll keep you in my prayers.
junkgreg: Oh, and while you're at it, you should get that \n fixed.
YahooHelper: Hi there junkgreg! How are you doing?\n I am an automated robot and I'm here to help you get familar with Yahoo! Messenger. Type some text in the window below and click 'SEND' to talk to me.
junkgreg: Wow, you're sicker than I thought; repeating identical statements right after one another now. That bug you've caught seems to be getting worse. Someone help the YahooHelper!
YahooHelper: Hi there junkgreg! How are you doing?\n I am an automated robot and I'm here to help you get familar with Yahoo! Messenger. Type some text in the window below and click 'SEND' to talk to me.
junkgreg: Don't despair though, I'm sure you'll get help. And your suffering has meaning-- it's been teaching me. Through your illness, you've helped me to come to grips with my humanity, thinking about this halting problem. Perhaps our real advantage as humans is that we know when to quit. Something I'm going to do right now. Bye!
{ Close Instant Messenger window }
there is an AI bot on irc.superchat.org that has been around for years. I can have conversations with it, it's very cool join #darkbot or #superchat, and just say "hi bot", and then say whatever and he will reply
Some years back, I did a bit of IRCing. Being female, I would get quite a few... inappropriate.. requests.
So, to handle these "advances," I wrote a little bit o' code that would intercept and handle these requests. Completely silent to me of course.
The great joy was logging the bot's conversations to my web page. This provided many an hour of chuckling at the expense of other people. I hate to think what the spiders did with it...
--- Phyre
The only reason I can find that we, as a race, are striving for artificial intelligence, is because the natural stuff seems rather hard to find.
Would the bot need to pass the Turing test (i.e., does the bot need to convince a human (ignorant) observer that it is human as well)?
Indeed what to do with the multiplicity of context on different channels? Several knowledge bases or different bots for each channel?
And how to teach the bot about context within a conversation. "This thingie here ... "-constructions are inherently difficult for nonhuman language processors!
Related to this is the so called (I believe, translated from dutch) problem of the sleeping dogs. Does the bot need to be aware that it's knowledge is limited, because even on the smallest, strictly defined area, it's impossible to enter all available knowledge, and possibly extend that through the use of propositional logic). We all know about Cyc, right?
In practice, I wouldn't expect such a realistic bot in the near future, unless you're satisfied with some Eliza variety, which is seriously flawed.
Lenny Foner wrote an interesting and entertaining paper back in 1993 about Michael Mauldin's Julia bot, which lived on MUDs and managed to fool some guy named Barry who kept hitting on "her". As Foner comments, though, it's not clear whether Julia passed the Turing Test or Barry failed it.
...bot!
Natural Language Processing is only a small part of what makes us 'intelligent'. Who is to say that the most intelligent beings in the universe even speak to each other? If they know what they have to do, why communicate?
Anyway, the point I'm getting at is... the Turing Test, and other such tests (such as this IRC Bot thing) are not correct measures of intelligence. Far from it. The bots are always just clever hacks that manage to sometimes respond correctly to our feeble communcation medium. Not because they are intelligent, but because they have hard-coded replies set out for them.
That aside, it's just a competition... I just wish they wouldn't use the phrase "AI" when refering to it.
These bots are sure to raise the average IQ of users on IRC...
Yes, you too can enter this new competition -- write the fastest, stupidest bot you can that can scream "1st p0st"...
Ah, progress...
yeah. i'm sure the next ground breaking event in science will come for an irc bot. i left my login password in my other brain. digitalambiotics@hotmail.com
IRC bots are one thing, they must reply on demand, usualy with a setence or two. Now picture a email bot that would answer email, it would have more time to "think" but it would have to create a much bigger and conviceble text.
:-)
I would picture the bot making web research to answer emails. This would be cool.
--
"take the red pill and you stay in wonderland and I'll show you how deep the rabitt hole goes"
[]'s Victor Bogado da Silva Lins
^[:wq
there's a classic paper by kenneth colby (citation below), in which he programmed an eliza-style chatterbot that tried to detect positive and negative undertones (positive attitude, agreement, interest, disinterest, humiliation, malevolence, etc.) in the person's text. those characteristics would then change the state of the bot's simple 'emotion system' (really simple, with clear internal variables for things like mistrust, fear, anger, etc.) which in turn would modulate the output - so the bot would appear withdrawn, angry, happy, and so on.
:)
he then did a sort-of turing test where psychologists were supposed to differentiate conversation logs with the bot from conversation logs with real paranoid patients. and the beautiful thing was - the bot was a more convincing paranoid patient than the paranoid patients themselves!
the problem, of course, is that it's easy to write a paranoid bot, because paranoia is a very convenient excuse for not always necessarily making sense.
reference:
---
COLBY K.M. (1973) Simulations of Belief Systems. In: SCHANK R., COLBY K.M. (Eds.), Computer Models of Thought and Language. S. Francisco, Freeman.
---
My other car is a cons.