The AI effect describes a pattern where any task an AI system successfully completes ceases to be considered true intelligence. This phenomenon shifts the definition of intelligence to exclude newly mastered capabilities, creating a moving target for researchers.
Edward Geist credits John McCarthy with coining the term AI effect to describe this specific cycle of achievement and dismissal. Bertram Raphael made an early statement in 1970 that established the core observation behind the concept.
Software developers often refuse to use the term artificial intelligence because practical achievements are quickly assimilated into other domains while remaining invisible as AI. Funding pressures during the second AI winter forced many scientists to adopt alternative labels like robotics to secure financial support.
In 1997 IBM's chess-playing computer Deep Blue defeated Garry Kasparov but critics immediately complained that the machine used only brute force methods without true intelligence. John McCarthy expressed disappointment at this outcome by describing the device as lacking deep understanding despite its spectacular success.
Michael Kearns suggested that people subconsciously try to preserve a special role for themselves in the universe by discounting artificial intelligence. Herbert A. Simon noted that the very idea of AI arouses real fear and hostility in some human breasts, leading to defensive reactions that protect human identity.