In the early 1980s the commercial face of AI was the expert system: a program that encoded the rules a human specialist would use - to diagnose an illness, configure a computer, or prospect for minerals - and applied them to new cases. The discipline of building these systems had a name, “knowledge engineering,” and a leading advocate in Stanford’s Edward Feigenbaum, often called the father of expert systems.
Feigenbaum did not just build the systems; he promoted the vision. In 1983 he and Pamela McCorduck published “The Fifth Generation: Artificial Intelligence and Japan’s Computer Challenge to the World,” arguing that knowledge-based systems were the next era of computing and that the United States risked losing it to a major Japanese research program. The book had real influence on business and technology policy, and it helped fuel a wave of expert-systems startups, dedicated hardware vendors, and corporate AI groups.
The boom did not last. Expert systems were expensive to build, brittle outside their narrow domains, and hard to maintain as the encoded rules grew. They ran on costly specialized machines just as cheap general-purpose workstations were arriving. By the late 1980s the market for that specialized AI hardware and tooling had collapsed, taking much of the funding enthusiasm with it - a downturn often grouped with the broader “AI winter” of the period.
Expert systems were not useless; narrow rule-based systems still do real work today. But “knowledge engineering as the future of computing” was a hype cycle, and the bust that followed left a generation of executives wary of the phrase “artificial intelligence” - wariness it took decades and a different technology to overcome.
Note: Feigenbaum’s own 1983 book and his role as the era’s leading promoter anchor the boom; the late-1980s collapse is a matter of broad historical record rather than a single primary document.