|
| Andrew
Leonard has written a popular social history of bots. Bots, the
Origin of New Species is not an scholarly book, but it is an academic
and a well-researched one, with an eight-page bibliography, a glossary,
and an index. The book will not appeal to a wide audience, but, if you
have an interest in MOOs. MUDs, IRC, Usenet, search engines and how these
work, then Bots the origin of a new species by Andrew Leonard will
be a fascinating read. The book traces the history of bots and describes
many of the interesting people involved in the development of bots as well
as major events in bot history. Leonard has talked to the right people,
including many bot writers, when researching the material for his
book. Alex Cohen, an early bot writer who worked for McKinley Corporation
when they were developing Magellan, an Internet directory, read the manuscript
for the book, and helped Leonard steer the book toward its completion.
Bots are a complex subject, and Leonard has made them understandable and
interesting. Leonard goes beyond providing an interesting and understandable
history of bots. He describes bots as being a new species, part code
and part human, with many human characteristics: “human names, the ability
to crack a bad joke, zany habits”. (Leonard, 1999. p.15). Bots have
been anthropomorphized. They exhibit personality and perform a service,
and they do this in an autonomous way without direct human supervision.
Bots have a purpose in life, and they have a habitat of their own, cyberspace.
Different bots reside in different spaces of cyberspace. “A bot designed to work in a TinyMUD may not work in a MOO, because of the differences in programming languages used to construct TinyMUDS and MOOs.” (Cohen, 1997, p. 33). The Web welcomes all bots. Chatterbots like Eliza and Julia have taken up residence on Web pages. IRCbots, frozen in suspended animation, are accessible in Web-based archives of code. Gamebots lurk, waiting to do battle, in online gaming networks interwoven with the Web. The Web, irresistibly connecting and binding together all the communities of cyberspace – MUDs and bulletin boards and chat rooms – is the reconvergent Pangaea, the mother continent where all bots someday will roam. (Leonard, 1997, p.152).Bots are everywhere in cyberspace, and it is difficult to enter cyberspace without encountering a bot; however, unless you know what a bot is and how a bot works, you may never realize that you have had an encounter with a bot. Bots are software programs, sofware versions of mechanical robots. They are strings of code written by anyone from teenagers to computer scientists. They receive their directions through the algorithmic rules that guide them –if this happens, do this; if that happens do that. Bots perform a service; they entertain; they annoy; they work; they play. Leonard maintains they are the “first indigenous species of cyberspace.”(Leonard, 1997, p.11). He distinguishes between intelligent agents and bots. Intelligent agents are software programs written to help humans cope with information overload. Although the two overlap, intelligent agents do not have the personality characteristics of a bot, and bots do not have to work as hard as intelligent agents. There may have been more of a distinction between the two when this book was written, but BotSpot, the authority for information about bots on the Internet notes that the term "bot" and "agent" have become interchangeable. (“What is”, 2001) Leonard explains the various types of bots and their origins. He developed a family tree, which traces the various branches of the family. Leonard traces the beginnings of bots to the MIT daemon created by Fernando Corbato, a founding member of the MIT Laboratory for Computer Science. This bot was developed to back up the computer files on the 7095 . “Corbato’s daemon belongs at the top of the great of the great Tree of Bots. Call it the ur-bot, the primeval form to which all present and future bots owe ancestry.” (Leonard, 1997, p.29). Chatterbots occupy a separate branch on the bot family tree. Chatterbots – bots that talk- are special. To rigorous bot anthropologists, chatterbots are the only true bots, the only software programs that dare imitate the personality and mannerisms of real humans and that aspire to don the mantle of an intelligent being. (Leonard, 1997. p.39).Eliza, created by Joseph Weizenbaum in 1996, was the first bot that could carry on a conversation. Eliza was named after Eliza Dolittle from George Bernard Shaw’s Pygmalion. Cohen refers to Eliza as the first "bot erectus", the first software program to impersonate a human successfully. Eliza consists of two parts, a parser that analyses information typed into the program, and a script that selects appropriate responses to type back. Working on patterns of recognition Eliza substituted prepared phrases for keywords posed in questions. Eliza was a hit with the public. The public loved her, and people chatted with Eliza,confiding their secrets to the program, often believing her to be a human. The public's love affair with Eliza did not surprise Leonard, “Any culture that can go mad over pet rocks can certainly fall in love with a computer program.” (Leonard, 1997, p.46). The most famous chatterbot was Julia. Many people have fallen in love with Julia, and Julia has been the subject of research papers. “Julia represents a giant step forward for botkind.” (Leonard, 1997, p.53). Between 1991-93, Julia competed in the Loebner Prize Competition, where she placed third each time. In the Loebner Prize Competition, different computer programs are given the Turing Test. The Turing test is named after the English mathematician, Alan Turing, who helped crack the Nazi enigma code in World War II. In the Turing Test, a human judge, another human and a computer program are in separate rooms. The judge poses questions to the two, trying to determine by their responses which is the bot and which is the human. No bot has passed the Turing test. The Web, however, provides a more even playing field for chatterbots. In the Turing test, the judge has the advantage of knowing that one of the contestants is a bot. On the web, people are expecting whoever is conversing with them to be human. They do not expect to talking to a chatterbot. “Today’s bots live on the Net, an
inherently friendly environment for supposedly intelligent software programs.
The Net makes it easy to be a bot.” (Leonard, 1997, p.60).
Kenneth
Colby was a collaborator with Joseph Weizenbaum in the creation of
Eliza; however, they disagreed over the uses of programs like Eliza. Colby
envisioned Eliza-like programs being used to play a role in the health
system and be used to treat humans. Weizenbaum could not imagine that a
computer and a human could form bonds. “The prospect of Eliza-like
automation entrusted with mental health of a nation still appalls Weizenbaum
today. .. Who ultimately is responsible when they go awry?” (Leonard.
1997, p. 73). Colby, a professor of cognitive science at UCLA, and
his son, Peter, were still busy marketing Overcoming
Depression, a software program to help patients overcome their problems
when the book was written. Peter continues with this work and marketing
the program.
Leonard explains the dual nature of bots. Like humans, there are some bots that perform good deeds and bots that perform bad deeds, and like humans, bots might be good at one time and not so good at another. Scooter, the web robot, behind AltaVista, created by Louis Monier, is an example of a bot that works hard and provides excellent service. It is a good bot, and does what it was created to do. Cohen calls Scooter “the leanest, hungriest bot in cyberspace.” (Leonard, 1997. p. 168). Five thousand lines of tightly packed C code, Scooter is best of breed, smarter than Michael Mauldin’s Lycos spider, faster than Brian Pinkerton’s WebCrawler, more robust than Alex Cohen’s Wobot. Scooter is wily – it knows how to skirt dangerous black holes and avoid nasty compressed files. And Scooter is strong… But most of all, Scooter is a speed demon, able, in the early summer of 1996, to traverse the entire depth and breath of the millions of documents on the Web in little more than a week.” (Leonard, 1997, pp. 168 – 169).Scooter was the reason for AltaVista’s success as a top-rated search engine. Only recently has this position been challenged by some of the second-generation search engines, such as Google, that have been developed after this book was written. (Elgin and Kerstetter, 2001). Not all bots are so helpful. “We will have bots that say one thing and mean another, that disguise themselves to get past checkpoints, that manipulate other bots and humans for their own advantage.” (Cohen, 1997. p.189.) Spambots are annoying and troublesome bots. Spambots are software programs that scan web pages for email addresses, and then initiate messages to these addresses. Spambots have resulted in the writing of cancelbots that locate spam messages that come from a particular address and cancel them. Bad bots give rise to good bots to nullify the effects of bad bots. Bots beget bots. Evil bots will accompany good bots – so it has been from the beginning and so it will be until the end. Bots may have originally evolved as playthings, as experiments in artificial intelligence, or as solutions to problems arising in complex computer-created environment. But bots are far from limited to playing only those roles. As the interface between human and computer, as anthropomorphic reflections of our own deeper character traits, as plowshares that can be hammered into swords, bots will exemplify all our hopes and fears. (Leonard, 1997, p.185).Many bots live a dual existence. Having been created for a helpful purpose on the Net, some of these helpful bots can be used for purposes for which they were not intended. Perhaps, this is a reflection of the complexity of the interaction of human and machine. The ability to write computer programs gives a person the power to write bots, and many Internet users want to test their bot-writing ability and test their power in bot creation when they unleash their bot creations on the Net, giving little thought to the consequences of their creations. Humans plus bot equals organic cyborg. Humans plus bots plus Net equals an unimaginably complex, multicentered collective of interlinked cyborgs – an organism greater than the sum of its parts. (Leonard, 1997, p. 230).The duality of bots is exemplified by the shopping bot. The shopping bot, that is used to assist online shoppers find the best prices on the Net and assist with online shopping, can just as easily be used to deliver a record of what the user has purchased, when and from whom. Thus, the friendly shopping bot can become a surveillance bot, invading privacy and creating electronic records that identify the habits and characteristics of the Net user. Surveillance bots can track human movement through cyberspace, keeping a digital record of where an Internet user. Leonard identifies this is one of the most troubling aspects of the bots’ future: their ability to create electronic profiles of our cyberspace existence. Bot-created profiles won’t be confined to lists of email addresses. As bots become smarter, and the tentacles of the Net intertwine themselves further and further into the minutiae of our existence, bots will be constructing profiles of human behavior that dwarf what is currently possible from reviewing a credit trail or a barcode-generated record of supermarket purchases. (Leonard, 1997, p.188).Bots do not always do what is intended, and behind each bot is the human that created it, ready to respond with a fix. If a bot causes chaos, another bot is created to eliminate the chaos. Leonard refers to this knee jerk reaction of creating new technology fixes to remedy the problems created by technology, the technodialectic. Every technology fix generates a new problem, which generates a new fix, which generates a new problem, and so on. It is an evolutionary process in which the better bots survive and proliferate, and ineffective bots are eliminated. Leonard compares this process of bot survival to Darwin’s theory of natural selection, where only the fittest survive. It is not necessary that bots become smarter; they only need to survive. Evolution just happens – it isn’t headed anywhere, except onward. The theory of natural selection explains only how organisms through time to change to local environments. It does not mandate that those organisms become smarter or stronger or faster or lovelier. It just demands that they survive. (Leonard, 1997, p. 236).Leonard believes that sometimes it appears that the technodialectic indicates progress on the Net, with the appearance of such things as Smart newsreaders, which prevent multiple appearances of the same message from several news groups, and the Robot Exclusion Protocol, aimed at controlling bots from invading servers or parts of servers that display the robot.txt file. However, he realizes that we will probably never see the end of difficulties caused by bad bots. “The technodialectic may mute chaos, but it will never resolve it.” (Leonard, 1997, p. 236). One computer will do as it is told. But a million computers linked together, responding to the needs and desires and obsessions of a million people, will not. A network is inherently unstable. The Web will be ever restless. It is at no one’s beck and call. And no matter how many times software protocols are patched or how many loopholes are plugged, no engineer, or team of engineers, will ever finish the job. The Net may be self-balancing, but it never in balance. (Leonard, 1997. p.234).
|
Main Page | The Author| The Book | The Review | Relevance |Glossary | References | |