Early ML & symbolic AI
What were the AI winters and why did rules-based AI stall?
Twice the field promised more than it delivered, and funding collapsed. The deeper reason rules-based AI ran out of road is that hand-written rules cannot cover the real world.
An AI winter, as Wikipedia defines it, “is a period of reduced funding and interest in AI research.”[1] The pattern was a hype cycle: big promises, disappointment, then funding cuts. There were “two major ‘winters’ approximately 1974-1980 and 1987-2000.”[1]
The first was triggered partly in Britain. In 1973 Sir James Lighthill was asked by Parliament to assess UK AI research; his report “criticized the utter failure of AI to achieve its ‘grandiose objectives’” and “led to the complete dismantling of AI research in the UK,” surviving in only a few universities.[1] The second winter hit the expert-system business: “in 1987 … the market for specialized LISP-based AI hardware collapsed.”[1]
The deeper reason rules-based AI stalled is built into the approach. Even successful systems like XCON “proved too expensive to maintain. They were difficult to update, they could not learn, they were ‘brittle’.”[1] Brittle means they broke on anything outside their written rules; they had no common sense to fall back on. Building them was slow and costly too, because of the “knowledge acquisition problem”: getting enough time from the scarce human experts whose knowledge had to be turned into rules by hand.[2] A system that cannot learn must be told everything, and no one can write down every rule the world needs. That limit is exactly what the learning approach was built to get around.
References
- AI winter — Wikipedia
- Expert system — Wikipedia