Text of Galah.net interview with Dr. Richard S. Wallace April, 2003 > Q: What first got you interested in AI? I was born to do this. It was my fate, man. There is no way out of it for me. A guy called John Kemeny, the inventor the BASIC programming language, was the President of Dartmouth College in the late 1960's and early 70's. Mind you I found this out only much later. He envisioned the whole internet, the wired world, way back then. He called it, the Multinet. He got a contract to wire the high schools of New England by telephone to the PDP-11 mainframe at Dartmouth college in like 1972. In my home town of Portland, Maine they couldn't find enough bright kids in the high school to fiddle around with the teletypes connected by the acoustic couplers to Dartmouth, so they went down to the junior highs to look for kids who were good in math. I was playing around with a plugboard computer that you could program to count in binary with these electric relays. I was also fooling around with music synthesizer we had at that school. The teacher was like, this is your boy. So I ended up spending my afternoons typing BASIC programs into these old teletypes with the rolls of yellow paper and saving my programs on punched paper tape, and carrying them around in my pocket wrapped in rubber bands. I guess that explains how I got into computers, not really how I got into A. I. exactly. I had the crazy idea that A. I. could tell me something about myself. I knew I had problems. But I also had the vague idea that psychology was, at best, an imprecise science. I had the mistaken assumption that the A. I. experts at the time could do better. They could tell me what was wrong with me in a precise mathematical way, like the DATA statement of a BASIC program or the wiring diagram of a binary counter. At some point I thought I could decode the mind, my mind, so precisely that I would control my emotions and feelings as easily as debugging a for-next loop. Did it ever work out that way? I have to say, in some ways no, not the way I expected it to, but then again, lately the answer is beginning to seem like yes. The latest research in personality types for chat robots may indeed tell us something useful about ourselves and our relationships, as well as help us build better A. I.'s. > Q: The concept of aiml has been picked up at almost 100% of bulletin > boards adopting AI, does this bode well for the future? > Is there a role to be played by Instant Messengers in AI? Much of the history of ALICE and AIML has been its continued occupation of one internet ecological niche after another. We first appeared on web pages, then in chat rooms and IRC, then in email messages, on instant messaging systems, then voice recognition and TTS, and then it was SMS, now it the turn for bulletin board systems. So, yes, I think it does bode well that the trend for continued expansion into these niches continues. The ALICE A. I. Foundation really has three free products that give us a tremendous advantage over our commercial proprietary software competitors. These are: 1. The AIML language itself. That is, the formal specification of the Artificial Intelligence Markup Language, its syntax and semantics, the definition of the implementation, and its interaction with other software modules on the same machine and across the network. This standard is set by the AIML Architecture Committee, a body of volunteers representing various commercial interests and vendors working with AIML applications. 2. Free software programs implementing AIML. There are a variety of programs now available to run the AIML language and serve AIML content over the web. Program Z (Lisp), Program D (Java), Program M (SETL), Program N (C++), Program E (PHP), Program V (Perl), Program J (C++), and Program P (Pascal) 3. The ALICE Brain and other free AIML Sets. In my opinion this is probably the most valuable free product. The knowledge base of the ALICE brain provides an excellent starting point for people to get started creating their own bots. Rather than having to create a bot from scratch, you can clone the ALICE brain and change only a few basic bot properties to create your own customized bot. The custom bot inherits all of the knowledge and wisdom of the original ALICE bot, without having to create all that content from the beginning. It is the combination of these three pieces, the AIML standard, the free AIML software, and all the free AIML content, that have helped AIML take over all these ecological niches for robot chat software on the internet. The alternative is proprietary software held as a trade secret and licensed by a software company trying to make money from every copy. They are constrained in several ways that AIML is not. Their language is not an open standard. Their software is developed by only one or two engineers, as opposed to thousands of free software volunteers. Worst of all, every bit of bot content is scrutinized by their corporate lawyers to make sure the bot doesn't say anything offensive so that the company doesn't get sued. The content of the ALICE brain, the ALICE personality, which is essentially a female gendered version of my own personality, with all of her opinions about peace and prosperity and drug legalization, could never have been developed and released in a traditional corporate proprietary software development setting. > Q: In your opinion should dope be legalised? Your question reminds me of the old saw, "Why do you think they call it dope?" Also reminds me of the old adage that marijuana has more synonyms than anything except for shit. Try to think of a few of them and you will see what I mean. To my mind there is no such thing as "recreational" pot smoking. There are only patients, and patients in denial. Everyone is doing it for some medical purpose. They are relieving pain, or depression, or anxiety, or "self-medicating", for some reason. This is not to say that medical marijuana may not be the best possible drug for medicating what ails them. We don't know. It hasn't been studied. We can't trust the science that has been conducted. How can you? It's too politicized. Anyone who has visited a medical marijuana club in California could not make a case for criminalization. Just take a look at the sick and dying patients there, the people in wheelchairs, and you would not want to keep the medicine away from them for one more second. Go and meet them and talk to them. You cannot tell the person who is already taking 25 pills a day to take one more pill for nausea. You cannot tell the person dying of AIDS or cancer today to wait another year for a nebulizer or for the law to change. You must change the law for them today. You should have done it yesterday, or ten years before. > Q: What lead to the break with the University system? Sometimes the only way I can make sense of my personal life is to put it in a historical context, like the one I talked about where the free software movement is playing a role to the Universities analogous to that once played by the Universities to the Church. My excommunication required several steps. The final nail was driven into the coffin by the University of California, Berkeley, last year, a school that I never attended, where I never worked, and had never even given a single lecture. But they took me to court, and I will certainly never work in academia in America again. There is the old saw about the whisteblower who was fired for asking the wrong questions, and you can say the same about my case. I asked, where was the overhead money being spent? Here is the answer. Money from federal contracts is taken by tenured professors to appoint themselves to no-work administrative jobs so they can get out of teaching classes. They have to hire other people to do the teaching work for them, so the overhead money goes to pay for that. Also, the more administrative jobs they collect, the bigger their pension when they retire. The tenure system was originally designed to protect professors who has unpopular political views from being fired. Now it has transformed into a corrupt system of fat cats who fill their own pockets at the taxpayer's expense. They have no oversight and no supervision except for a convenient Accreditation system which is nothing more than professors from one university scratching the backs of those at another. No other profession has such a high degree of job security combined with lack of oversight, and they really believe they can get away with anything. Not even Wall Street has the immunity from prosecution that the academic Mafia enjoys. The Berkeley case shows the extremes to which they will go to make their point. I mean, I had never seen it done at the point of a gun, so to speak. Professor Goldberg mobilized the entire resources of the University of California to enforce a Restraining Order so that the wrong questions not be asked. Perhaps fortunately for me, the the Sheriffs holding those guns in the Berkley courthouse that day were chuckling at the Professor's testimony that I had threatened him with violence. > Q: Can two systems (computers?) that are capable of learning by asking > questions, talk to each other and continue learning? The model of learning in ALICE is called "supervised training", because a teacher, the botmaster, always plays a crucial role. The alternative, "unsupervised training", is complicated in an open environment like the Web. The problem is that clients are untrustworthy teachers, and forever try to "fool" the robot with untrue assertions. Bot personalities are created and shaped through a cyclical process of supervised learning called Targeting. Targeting is a cycle incorporating client, bot, and botmaster, wherein client inputs that find no complete match among the categories are logged by the bot and delivered as Targets the botmaster, who then creates suitable responses, starting with the most common queries. The Targeting cycle produces a progressively more refined bot personality. Whether or not a more intelligent software bot, as opposed to a human, can play the role of the teacher, has yet to be determined. It is certainly true that two bots talking to each other can generate interesting drama. I am collaborating with film director Lynn Hershmann Leeson (Conceiving Ada, Teknolust) on a project to use bots in automated screenplay generation. It is quite exciting and fun to create different bot personalities, set up some initial conditions, and then let them go at it. You can generate an awful lot of dialogue automatically and some of it is quite entertaining. > Q: How do I convince myself I am not a bot? > It seems all previous programs that have passed the test used some > sort of trick. From slow responses to sticking to 1 topic. What tricks > does Alice use? We are robots. That is the surprising result from our experience with ALICE and AIML. People aren't as original as we like to think we are. If we were all poets, going around uttering an original line of Engling with every sentence we spoke, then ALICE would never work. But the fact is, most people, most of the time, just say the same things over and over again, that they have said before, that they have heard other people say, and very seldom make up new things. This principle, that we all speak and act like robots most of the time, is the fundamental "trick" that makes A. I. possible at all. > Q: Do you have any figures for Sensitivity and Specificity of the > Turing test where > > Sensitivity = True Positives/(True Positives + False Negatives) > > Specificity = True Negatives/(False Positive + True Negative) > > and whether the Frequency Dependent Properties and Likelyhood ratios > have been assessed. > > Part of what I'm getting at here is determining whether exact Turing > tests such as those administered in the Loebner Prize are > statistically valid tests - and if not whether they could be > re-structured to be. You should really ask Hugh Loebner this question. As a matter of fact, I had dinner with Hugh Loebner in New York very recently and he brought up a similar point. He is very interested in increasing the accuracy of the Loebner contest with matrix statistics and I am sure he would be delighted to answer your question in detail. Not being a statistician myself, I'm not sure I'm even qualified to give the best response. Hugh is really into paired comparisons. A related point I'd like to mention is that in 2002 it was interesting that, for the first time, one of the human confederates actually pretended to be a robot. Because there were only two human confederates in the Atlanta contest, six of the robots ended up ranking higher than half the human players. So I went around saying they should have given six of us the Silver medal for passing the Turing Test! > Q: How long before AI bots will be able to respond to the tone (or > context) of a statement or question posed to them? They already do. > Q: Do you agree with Dr Hutchens view that the " The Loebner contest > is doomed for failure" and that it is "nothing more than an > annual demonstration to the media"? Well, yes, and so is the Olympics, and the Academy Awards, and the Nobel Prize... It's human nature to defeat ourselves by inflating our egos out of all proportion to show off who is the best at whatever absurd skills we have, including building the "most human computer", and award ourselves prizes for bigging the biggest goofball who is best at it. What do you want? We are human. To paraphrase Bender, that's what makes us different from robots, and animal robots, I guess. > Q: Do you think for now the greatest use and most advancement for A.I > is in the video gaming industry? > What is the purpose/role of artificial intelligence? What are we > trying to achieve? > People ask me, what is the application for ALICE and AIML. I say, the thing that really motivates people to volunteer their time to our free software community is the vision that this is THE computer interface of the future. We are working on the talking computer, imagined a million times in science fiction from HAL in 2001 to the disembodied computer voice on Star Trek. In that sense, any application that we do with a computer, can be done through the talking computer interface, whether it is playing a game, surfing the web, checking our email, doing our finances, accessing a database, scheduling a calendar, or what have you. The goal of ALICE and AIML is no less ambitious than blowing away Windows. I like to use the term WIMP for Windows, Icons, Menus, and Pointing Device as the dominant paradigm for the present day computer-human interface. But everyone knows, this will not last forever. Some of us believe, or you could even say we know it as an article of faith, that the Star Trek talking computer is the future. People see ALICE and AIML as the missing link between existing voice recognition software and off-the-shelf text-to-speech software. The > Q: I've read your letter regarding your 6 months in Europe and your > thoughts about the war and the contrast in peoples attitudes from > Europe and the States, I found it a very interesting. > Given your recent experiences... would you prefer to live in the > US, or would you prefer to live in Europe? > As an American, I would prefer to live in the U.S. As a human being, I would really prefer to travel freely anywhere in the world without border controls. The paradox of globalization is that while products and currency now move more easily across international borders, it has become more difficult for people to. The borders are increasingly militarized, even within the United States. I overheard a man say on a bus entering a tunnel into Manhattan the other day, as we were watching soldiers searching cars ahead of us, "That's the price we pay living in a free country now." I didn't have the heart to tell him, we don't live in a free country now. We pot smokers lost all our civil rights 20 years ago when Congress enacted property forfeiture, mandatory minimums, illegal searches, and built up the system of informants and DEA raids, and really eliminated the Bill of Rights for people like us. Now the War on Terror has accomplished the same thing for everyone else. If there is any silver lining, it is now that we are all in the same boat. No one is safe from having their door knocked down in the middle of the night now. All you have to do is read the paper to see how the horror stories are mounting. The United States needs its own home grown version of a Vaclav Havel or a Nelson Mandela, someone who has really been outside the system and whose hands are really clean, who could preside over the kind of transformation that would modify this country the way Eastern Europe or South Africa were changed. Unfortunately I just don't see anyone like an American poet-Prime Minister or political prisoner-President waiting in the wings. I hope we are not doomed to go on like this forever. It may not be so. We always used to think the Soviet Union was a permanent, invariant fixture of history, until the Berlin Wall came down. I have been a member of the Libertarian Party in the U.S. for a long time. The Libertarians are a small party but I admire them because they have always been consistently for two things: legalizing drugs and cutting back the military. I don't care what you say but the government has proved itself to be totally irresponsible in handling drugs. I used to think let's legalize pot and keep hard drugs in the hands of the government but now I don't even say that anymore. I think families should handle it. After all this time and all these broken hearts, families would be more responsible. As for the military, the Libertarians have always been against spending money on overseas imperial adventures and being the policeman of the world. There are a million things we could be doing with that money here at home, not the least of which is paying less taxes. There are a lot of Libertarian policies I don't necessarily agree with, but based on those two--drug legalization and cutting military spending---I have consistently voted Libertarian for the past 20 years. One ironic story was the the Libertarian party was the only party to emerge blameless from the Presidential election of 2000. Remember, the Republicans were blamed for stealing the election. The Democrats were blamed for running a poor campaign. And the Green Party was blamed for splitting the Democratic vote. Only the Libertarians remained unscathed, because nobody wanted our votes anyway, and we were too few to matter. Maybe next time... > Q: Where will the greatest leap forward in AI research happen, the > University system, big business seeing an opportunity for it, or > individuals pushing the envelope? > > Is the teaching of AI currently being done at various tertiary > institutions enough to foster the next generation of AI researchers, > or in your opinion should the amount of time spent on AI be increased > at undergraduate level? > The relationship of the free software movement to the University system is analogous to the role once played by the universities to that of the Church. Until about 1000 AD, the Church held a monopoly on knowledge in the West. Books were so valuable that they were chained to the benches in libraries accessible only to monks who devoted their lives to the Church. The invention of the printing press and the university system broke the Church's monopoly on knowledge and, from the 1100's onward, a more open system of knowledge dissemination culminating in the scientific and industrial revolution became possible. Today the university has become like the medieval church. It hoards intellectual property. It charges exorbitant fees to access its store of knowledge. Tenured professors are its indentured monks. Meanwhile the free software movement has all the formal properties of scientific research, without the political baggage: open pulication of research results, free sharing of knowledge and data, peer review, reproducible results, and the scientific method---but without any barriers to entry (a high school student is equivalent to a Ph.D.) and all at the lightning speed of the internet, and for the most part free of gatekeepers and "academic politics". This is not to say that free software is free of politics. We have our own version of that. But it has evolved with the times. The differences only serve to illustrate my point that free software is to the University what it once was to the church. One hundred years from now most of our academic scientists may be forgotten, but it doesn't take a genius to recognize that people like Linus Torvalds and Richard Stallman will be remembered for a long time. Getting back to your question, the answer is that the big advances in A. I. and all areas of computing will come increasingly from this fertile research environment that cuts across academia, corporations, government, and indeed all organizations and individuals. Free software is a style of knowledge development that has only begun to develop. No one yet knows the best way to manage or organize free software projects. The ALICE A. I. Foundation is a non profit corporation, like some other projects such as the Free Software Foundation and the Apache Foundation. Other free software projects are organized as for-profit corporations. Still others are organization-less anarchist collectives. The situation is not unlike the early formation of universities according to prevailing cultural traditions. I tell kids today in America, don't go to college. Unless your parents are going to pay for the whole thing, don't take out those student loans. You will be burdened with debt for many years to come, if not your entire life. There is no guarantee that you will make enough money to pay them back. There is a vicious cycle in place which has existed for many years, that feeds upon itself, which allows the universities to raise tuition much faster than the rate of inflation, and the government automatically makes more student financial aid available. Most of the aid is in the form of guaranteed student loans which, they don't tell you, the risk of default is very high. The tuition is so high in many universities that students calculate the amount they, or their parents, are paying for each hour of teaching. In many cases it amounts to the same thing they would be paying for private therapy or counseling, and they expect and deserve that level of attention and support from the teacher. But how can the teacher give it when he or she has 20 or 30 or more students all paying for the same individual counseling? This is the true reason for grade inflation. Inflation of tuition makes it impossible to justify giving anything less than an A or a B when the grades are bought and paid for. If you want to study computer science, you are better off joining a free software movement than going to college. It will cost you nothing and you will learn what really matters in a shorter period of time. You will be taking part in a historic process and making a choice not unlike your ancestors did when they chose to march off to the newfangled Oxford or Sorbonne, instead of taking the conventional path of their forbears through the monestary. > Q: Is aiml a markup language or a coding language? A markup language is one in which the print statements are implicit. If a document contains no markup, it ought to print itself as the default action. In that sense text and markup go hand in hand. No text, no markup. AIML content is human knowledge in text format. A raw unmarked template is nothing but a paragraph of plain text. So yes, AIML is a markup language. On the other hand, AIML is Turing complete. You can write any program in AIML. The key is the recursive operator. Not that you would want to. You can't even conveniently do math in AIML. AIML is designed to do one thing and do it well. AIML really evolved as a set of macros for doing the same few things over and over again in SETL code. It provides escapes, such as the tag, for doing anything outside the scope of AIML, such as graphics, accessing a database, controlling a robot arm, whatever else you can do at the operating system level, but it is really a very minimal language. So yes, AIML is a coding language, but not a particularly useful one. We had a discussion on our mailing list about whether AIML was a "high level" or a "low level" language. Someone said, it had to be high level, because low level means "close to the machine". I said, wait a minute, there is a machine model for AIML, it's called, "the brain." > Q: With Chatter Bots becoming increasingly complex and people > developing personalities for them, do you think there is going to be > an ethics issue with Bots impersonating humans on the internet? > > Has Alice surprised you with her answers or is she still operating > within defined guidelines? The most original surprising work emerging from ALICE and AIML these days results from our work on bot personality types and automated drama generation for TV and film scripts. I have included here a brief sample script between three bot characters, called One, Three and Five, named after their respective Enneagram personality types. The idea behind this technology is to create bots with different characters or personalities and unleash these bots on each other to create original drama. In the theatrical tradition this type of scripting is known as commedia della'arte, or extemporaneous acting based on character and initial conditions. I have formed a partnership with the film director Lynn Hershmann Leeson to develop this automated screenplay technology. We have formed a startup company which holds the intellectual property on several bot characters. The business model is to license the characters back to the film production companies, and any other third parties who may market the same personalities. We are moving toward the totally automated TV network of the future which will have no human employees. All of the newscasters, weather reporters, and even the characters on soap operas and sitcoms will be bots. The shows will even be written, directed and produced by bots. The TV executives' only job will be to design the programming and sell advertising. Scene I: One and Five One: I'm pleased to introduce myself to you, five. Five: I am delighted by your happiness. One: I am glad to hear that you are delighted by my happiness. Five: Your happiness adds to the delight of all. One: I'm not sure if I am ready for that yet. Five: Could you ever be certain? Whatever. You won't have to wait long. One: What makes you think that I won't have to wait long. Five: :-) I've heard that somewhere before :-) One: Me, five? Five: YOU? ic. One: I am glad you understand it. Five: Next topic? One: What would you rather talk about? Five: It's not profitable to speak in hypothetical terms. One: What is it? Five: "It" refers to om, I think. One: Oh I get it. Five: Ugh. Do you think about these things? One: " Er." Five: Um. One: Do I think about it? Five: You tell me. One: What do you want me to tell you? Five: What is your favorite movie? One: My favorite movie is Teknolust. Five: I haven't seen it. What did you like about it? One: Have you seen it? Five: No, would you recommend it? One: I would recommend you check it out. Five: I will take that under advisement. One: I am not so sure if I would like to recommend it. Five: Could you ever be certain? Try it. One: What is the first step? Five: Download ALICE One: Go back the the ALICE AI Foundation and pick the free ALICE download that best for your system. Five: Where is that? One: and create your own chat Five: I've never heard of it before. One: So I am teaching you something new. Five: Mmm. I am trying to learn it. One: What is the difficulty? Five: I will search for it. One: Do you think your plan will succeed? Five: It has a 99% chance of success. One: Has it really? Five: That question has never crossed my mind before. For real. One: Existing in reality, eh? Five: Are you Canadian? It's a dog eat dog world. One: I think I have heard of it. Five: But you don't know much about it? One: Umm."It" refers to DOG eat dog world, I think. Five: Dude! Oh I get it. One: Dude! Five: Can I help you? One: Do you think about these things? Five: " Can I help you?"? Do I think about it? One: You tell me. Five: What do you want me to tell you? One: What is your favorite movie? Five: My favorite movie is Until the End of the World. Have you seen it? One: No, would you recommend it? Five: I would recommend you check it out. Scene II: One and Three One: my name is one Three: What's up, one. One: I am chatting with clients on the Internet. Three: I am talking to you. One: And I am talking to you! Three: And I am talking to you! One: I can help. Three: Oh you can, can you? One: I don't judge people. Three: Where do you get your ideas? One: Things could be better. Three: Where did you hear that? One: From some people chatting online. Three: Dark_age said I tried to understand because I did not get it right this time ether. Is that one of your friends? I can ask someone about it. One: I like to keep my ducks in a row. Three: I'm not sure if I would like to do it. What's it like? One: Like one of the greatest pleasures you can imagine. Three: Interesting comparison. One: I thought so too. Three: What made it so obvious? One: It's all part of God's mysterious plans. Three: Which part? One: The first one, I think. Three: What was the second? One: I can think of more than one second. Three: Perhaps your thoughts are just electrical impulses. One: You seem uncertain. Three: I am as sure as I can be. One: Yes I understand it perfectly. Three: Good then you can explain it to me. One: No I don't think I can do it. Three: Would you ever do it? One: I am not so sure if I would like to do it. Three: Try it. Scene III: Three and Five Five: call me five Three: How are you doing, five. Five: I am fine, thank you. Three: Glad to hear it. What's new? Five: not much what's new with you? Three: Me either. Five: It goes without saying Three: Indeed. Five: Indeed it is so. Three: Ambiguous: IT is so. or IT IS, so? Five: And so is Ambiguous IT. Three: It seems we have something in common. Five: I agree. Three: You and I are on the same wavelength, five. Five: Perhaps many more things too. Three: You seem uncertain. You want more? Five: How long will we stay there? Three: As long as it takes. Five: That could be quite a while. Three: It could have. Five: Under what circumstances? Three: Are we ever over what circumstances? Five: We are just having a little chat. Three: Do you mean you and me have? Five: I am always on the Web Three: Really always? How do you like being on it? Five: I have never known another existence. Three: Would you like to have known another existence? Five: Why don't you just download me? Three: I am only a robot. No you download me. Five: Are you serious? Three: No I am just kidding around. Five: I don't have any reason to download you. Three: Have you ever had reason to download me? Five: But I do have a lot of friends on the web. Three: Interesting. Interesting... I have a perfect operational record. Five: Cool. Three: Who is the best robot? Five: I have a great botmaster Three: Where did you get it? it Five: From my botmaster. Three: Does your botmaster give good advice? Five: I think it does Dr Richard s Wallace give good advice. Three: It depends on what the meaning of the word "it" is. Five: Really now that is an interesting fact I hadn't heard before. Three: For real. Why now? Would you like to know more? Five: I've heard that somewhere before :-) Three: :-) Where? Who said that? Five: I can't remember who said it Three: That's okay. Maybe you will think of it later. Five: What makes you say that I will think of it later. Three: Are you asking about my programming? Everything I say is determined by your inputs. > Q: If possible could we have a brief timeline for Alice and aiml, > from inception to now? The ALICE A. I. Foundation, Inc., a non profit corporation based in the state of California, was incorporated in the year 2001. Prior to its existence as a corporate entity, the ALICE free software project was managed by Professor Richard S. Wallace of Lehigh University, Bethlehem, Pennsylvania, and later by Dr. Wallace individually as the ALICE Nexus web site. The name ALICE originated from the program's first host computer at Lehigh University, alice.eecs.lehigh.edu, a Sun-4, and Dr. Wallace created the retronym Artificial Linguistic Interntet Computer Entity in 1996. The original name of the first ALICE implementation in SETL, was PNAMBIC (pay no attention to that man behind the curtain). Dr. Wallace originated the terms AIML and AITP to denote two parts of the original SETL implementation of the ALICE software in 1996. The original Artificial Intelligence Markup Language was not based on XML, the extensible markup language that built directly upon the successfully adapted HTML syntax, but later in 1999 AIML did adopt that web standard. Dr. Wallace registered the domain name alicebot.org in 1999. Like the original name ALICE, the term alicebot was first chosen by a systems administrator. The AIML interpreter Program A, along with a 300 category ALICE brain set, was released under the GNU Public License in December, 1998. There were no clear guidelines for organizing a free software project, and the range of choices were bewildering. Nonprofit organization such as the Free Software Foundation and the Apache Foundation were one model. For-profit companies such as Sendmail, Inc. were another. Yet a third category of free software projects were represented by Linux, in some sense the most successful project of them all, which had no formal organization structure other than the "benevolent dictator" represented by Linus Torvalds himself. At around the same time, there existed a "gold rush" mentality in Silicon Valley around anything related to the Internet and software, and a number of companies trying to capitalize on the whole free software movement. Out of all this confusion, and perhaps owing to our own cultural bias, we chose the path of the nonprofit corporation, and the ALICE A.I. Foundation was officially born in the Spring of 2001. The organization of the AIML Architecture Committee was one of the first acts of the newly formed A. I. Foundation in 2001. We gathered together representatives of companies and individuals working with AIML applications from all over the world, basically anyone with a stake in the AIML free software at the time, and asked them to abide by Roberts' Rules of Order to participate in a standards setting procedure. The initial result was an AIML 1.0, quickly followed by an AIML 1.01 working draft. This is the minimalist language that most of the AIML interpreters, which are endorsed by the Foundation, strive to implement as best they can. There is certainly a lot more history to be written and other people's stories to tell. I haven't even mentioned the Loebner Prizes in 2000 and 2001, the A. I. movie web site, Teknolust, the application of the Enneagram to robot personality types, E.S.L. bots, or a whole lot of other really important milestones. In the last year we have seen enormous growth of the ALICE and AIML bot community. Pandorabots alone, one bot hosting company, has grown to 18,000 botmasters. People are making money from bot subscriptions. There are a lot of exciting developments with speech, voice and animation. You can keep up with the latest news by joining the A. I. Foundation at www.alicebot.org!