Paper
Expert Systems: Commercializing Artificial Intelligence
Published Jan 1, 2022 · David C. Brock, B. Grad
IEEE Annals of the History of Computing
3
Citations
0
Influential Citations
Abstract
As this issue goes to press, news and discussion surrounding the commercial use of artificial intelligence (AI) by large technology firms—particularly machine learning and especially in the United States and China—are daily occurrences. Thoroughly commercial AI is placed at the center of discussions about geopolitics, military affairs, finance, labor, health, and art. This Special Issue of the Annals began with the observation that today’s commercial AI, and breathless discussions about its future implications, has a history. Starting in the 1970s, a very different form of AI from today’s machine learning developed and gained great attention: expert systems. Expert systems used logical reasoning to draw conclusions based on large sets of “If, Then” rules, engineered to represent the knowledge of domain experts. The 1980s witnessed a wave of commercialization of expert systems, by startups, by private sector firms, and by the U.S. military and its contractors. Throughout the decade, and accompanying this commercialization wave, scientists, engineers, and commentators put forward sweeping views of the importance of commercial expert systems for industry, the military, and culture. Despite the importance of the commercialization of expert systems to the history of AI, and its clear resonances with our present moment, this history was underdocumented and underexamined. (For more on this perspective, see David C. Brock’s, “Learning from artificial intelligence’s previous awakenings: The history of expert systems,” AI Magazine, vol. 39, no. 3, pp. 3–15, Fall 2018.) In May 2018, the Software Industry Special Interest Group (SI SIG) and the Software History Center of the Computer History Museum (CHM) conducted a two-day meeting for 20 of the pioneers who had formed expert systems companies or led the research and development work in expert systems during the 1970–2000 period. In addition, several computer historians and other representatives fromCHMand SI SIG participated. In conjunction with each meeting, several oral histories were also conducted to further enrich the sources for historians and researchers. Thismeetingwasmoderated by David C. Brock and Burt Grad, the guest coeditors of this Special Issue of theAnnals This continued a 20-year effort by SI SIG in collaboration with CHM and the Charles Babbage Institute to collect the first-hand recollections of software industry pioneers regarding the companies they had formed and grown in the 20th century. It was the 13th of these pioneer meetings, each of which has been video recorded and the transcripts of the workshops have been edited and posted on the CHM website. This issue includes two articles from historians, Hallam Stevens and Shreeharsh Kelkar. Stevens details the early commercialization of expert systems in molecular biology and biotechnology, while Kelkar examines the history behind software systems for instructing and tutoring students, so-called “edtech.” Paul Harmon, who edited a leading expert systems industry newsletter, provides an overview of the industry, structuring how different companies addressed the potential market opportunities. The next three articles are by company founders who tell the story of their own background in AI and how they were able to start expert systems companies, what markets they addressed, and how their companies fared over the next 10–20 years. They also describe their own experiences afterward. These authors are Doug Lenat (MCC/CyCorp), Peter Hart (Syntelligence), and Monte Zweben (Red Pepper/Blue Martini). Reid Smith covers the wide range of expert systems and other AI technologies that were used by Schlumberger to enhance their work in fossil fuel exploration. Following these articles is an anecdote by Fritz Kunze who talks about forming Franz, Inc., which was a long-lived company, following a quite different path from the companies formed by Stanford University Ph.D.s. To complete the issue, there is an edited oral history of Peter Friedland (interviewed by David Grier) who helped form and run IntelliCorp (which later spawned Teknowledge), and a biography of Edward Feigenbaum by Pamela McCorduck, who has sadly died since writing this biography. Ed Feigenbaum is 1058-618
Full text analysis coming soon...