The Rise of Artificial Intelligence

The formal field of artificial intelligence is generally dated to a summer workshop held at Dartmouth College in 1956, organized by John McCarthy, Marvin Minsky, Nathaniel Rochester, and Claude Shannon, at which the term "artificial intelligence" itself was coined by McCarthy. Early AI research through the 1950s and 1960s centered on what became known as symbolic AI, or "good old-fashioned AI," an approach based on representing knowledge through explicit rules, logical statements, and symbolic structures that a program could manipulate to reason toward conclusions, exemplified by systems such as Allen Newell and Herbert Simon's Logic Theorist, from 1956, and General Problem Solver, from 1957. This paradigm produced notable early successes, including programs that could prove mathematical theorems, play checkers, and solve algebra word problems, leading several prominent researchers to make confidently optimistic predictions in the 1960s about how quickly machines might achieve broad, human-level intelligence. Symbolic approaches later evolved into expert systems during the 1970s and 1980s, programs such as MYCIN, developed at Stanford to diagnose bacterial infections, that encoded the specialized knowledge of human experts into large sets of rules for narrow professional domains, achieving real commercial use in fields such as medical diagnosis and chemical analysis, even as their reliance on manually hand-coded knowledge limited how well they could generalize beyond their original narrow domains.