The History of AI101 · Module II · Lesson 06 of 21
Article · 12 min

The first AI winter

The 1974 crash, in technical detail.

Summary

The 1974 collapse. The Lighthill Report (UK) and DARPA's disillusionment (US) cut funding after symbolic AI failed to deliver on machine translation and combinatorial planning. The winter was not a failure of ambition — it was a failure of the paradigm's ability to scale past toy problems.

Objectives
  • 01Name the specific technical failures that triggered the 1974 cut.
  • 02State the Lighthill Report's central objection.
  • 03Distinguish funding collapse from research collapse.
The Lesson

The technical causes

Combinatorial explosion in search, brittleness of hand-coded rules, and the collapse of machine translation (Georgetown-IBM's 1954 demo did not scale). Perceptrons had been declared limited (Minsky & Papert, 1969).

The institutional cause

Lighthill's 1973 report to the UK Science Research Council concluded that AI had not delivered on its promises. DARPA reallocated funding. The community shrank; the ambition did not die.

Key Ideas
  • Winters are funding events with technical causes.
  • Combinatorial explosion is the recurring enemy of symbolic methods.
References
  • Lighthill, 'Artificial Intelligence: A General Survey' (1973)