In the mid-1970s, the Lighthill report (UK, 1973) and cuts at U.S. agencies hit AI funding. The gap between limited demos and promises of machine translation or general robots had become untenable.
Scalability problems—combinatorial explosion in search, brittle hand-coded knowledge—became obvious. Vision and natural-language projects advanced more slowly than advertised.
The “winter” did not stop all research, but it did change incentives: more focus on bounded applications and less on near-term AGI. The promise–disappointment pattern would repeat in the 1980s–90s.
