Hanly@CCM.UManitoba.CA (Ken Hanly) (10/06/88)
The relationship between computers and `class' is perhaps even more evident in the explicit agendas of important players in the AI game than in the hidden agenda of the pigeonhole principle. Consider the following remark by E. Feigenbaum (Prof. of Computer Science at Stanford) "In Japan, the Fifth Generation Project aims to overtake the American lead in this most important of all modern technologies by establishing a `knowledge industry' in which knowledge will be a salable commodity like food and oil." (UNDERSTANDING ARTIFICIAL INTELLIGENCE, Texas Instruments publishing Dallas, l985, p. v) So much for Unger's comment concerning not considering computers as means for material production. Feigenbaum is concerned that the U.S. meet the Japanese challenge. The development of the field of AI has always been closely intertwined with both business and defense interests. The first computers were designed to serve computational and code cracking needs during the second world war. Even the nutty genius of computing Turing served as part of a team used to crack the German code ENIGMA. In l958 the Defence Advanced Research Projects Assoc. DARPA was set up. Among other things DARPA established the ARPA network so that their researchers could communicate with one another. In l983 DARPA announced a 600 million dollar 5 year program explcitly designed to develop and apply AI technology to military ends including autonomous systems -vehicles that would be able to run themselves, battle management systems, and so forth. The other response of the US to counter Japan's moves involved setting up a consortium of 2l firms to form the Microelectronics and Computer Technology COrp.(MCC) This consortium includes many of the big corporate players in AI including Bell, Boeing,Digital, Kodak, RCA, Sperry. 3 M etc.etc. It is perhaps symptomatic that this group is headed by ADMIRAL Bobby Inman stressing the comfortable linking up of the business and military corporate worlds. Even the beginings of AI reflect the cosy connection between business and AI -and academia and business. The four organisers of the Dartmouth Conference consisted of two people from academia and two from industry and the conference itself was funded by a Rockefeller Foundation grant. In the area of computers and AI in particular the academic pursuit of knowledge for its own sake and a critical understanding of society is becoming more and more subverted, and this occurs often not because of oppressive control by business and/or the military rather it is often through the actions of academics themselves. Prof. Roger Schank was (is?) chairman of Yale's dept. of computer science as well as chairman of the board of an AI company Cognitive Systems. He applauds the idea of the entrepreneurial university where the university becomes linked with outside corporate interests. Ken
marsh@linus.UUCP (Ralph Marshall) (10/07/88)
Ken Hanly writes: > The relationship between computers and `class' is perhaps even more > evident in the explicit agendas of important players in the AI game than > in the hidden agenda of the pigeonhole principle. > > Consider the following remark by E. Feigenbaum (Prof. of Computer > Science at Stanford) "In Japan, the Fifth Generation Project aims to > overtake the American lead in this most important of all modern > technologies by establishing a `knowledge industry' in which knowledge > will be a salable commodity like food and oil." (UNDERSTANDING > ARTIFICIAL INTELLIGENCE, Texas Instruments publishing Dallas, l985, p. > v) So much for Unger's comment concerning not considering computers as > means for material production. > > Feigenbaum is concerned that the U.S. meet the Japanese challenge. The > development of the field of AI has always been closely intertwined with > both business and defense interests. The first computers were designed > to serve computational and code cracking needs during the second world > war. Even the nutty genius of computing Turing served as part of a team > used to crack the German code ENIGMA. Having read this forum for several weeks now I am ready to walk into the den of Marxist thinking with a somewhat different approach. In order to make my viepoint (bias ?) apparent, I should state that Yes, Virginia, MITRE is the quintessential military-industrial entity, and I do in fact work in one of their AI groups. I've also taken the course Judith Perrolle used to flesh out the book which was quoted in the article which started off this whole discussion. I think she has a very interesting perspective, especially her self-description as a "realistic socialist". Anyway, on to my main points. The comment above about "computers [are a] means for material production" was evidently intended to be taken as a perjorative, but I fail to see the reason. Computers help perform a variety of tasks, and since each of us needs to earn some sort of living, to the extent that computers make our jobs easier (and allow our employers to continue to make a profit and thus keep us on the payroll) I don't see what is wrong with using them as just another tool for producing goods and services for sale. I realize that not everybody who uses a computer at work sees it as an improvement, but that can be said about almost any tool introduced into a factory setting. If used improperly it can make life miserable for the people using it; that does not mean that the tool itself is at fault. Remember, many handicapped people have productive and satisfying jobs due to a variety of computer products and aids that they could not perform otherwise. The next comment regards the view that AI as a means of making money is a "Bad Thing." There are obviously many views of what AI is trying to do, but the main focus of most investigation is to produce systems that "Do what I mean, not what I said." This is a valuable benefit if it can be achieved because people who have the less glamorous computer jobs will no longer have to memorize arcane commands and codes and do nothing more than keypunch all day. There are clearly tasks that computers can perform faster, and if AI technology can help people do them with less boredom I view that as a good thing. The conclusion I am trying to draw here is that targetting all business uses of comptuers as Bad is an inaccurate charge. There are clearly many things that need to be improved (such as computers that monitor performance all day long and rate you on the number of operations performed correctly per hour), but there are also many interesting applications that can replace rote human activity with automated work. The big problem is how to allow the people freed from the dull work to find something more interesting to do with their newly acquired time and mental energy. Computers cannot be just for intellectual stimulation. Making a computer is an expensive process that requires LOTS of customers before individual units can be available to private individuals. (I once worked on a job overseas which involved installing some straight-forward office automation software such a word-processors and spreadsheets in a small landscaping firm. The person I was training was a Pakistani national who wanted to know how I had learned to do all of this. When I told him that I had practiced on my home IBM-PC he was astounded that a kid could own his own entire computer. The point is that the masses can't have computers for their own enlightenment until they have jobs which provide income so that they can afford them, and so that the country in which they live has the required infrastructure to do interesting things with them. All of these benefits are a direct result of capitalist profit making enriching people.) Well, that's my 2 bits worth. Ralph Marshall