The ALPAC report
Era 2 · AI is born (1943–1979) › The first winter
- Lineage
- Symbols and logic
- Kind
- Institutional
Who
National Academy of Sciences
What happened
The Automatic Language Processing Advisory Committee (ALPAC), set up in 1964 by the Department of Defense, the National Science Foundation, and the CIA and chaired by John R. Pierce of Bell Labs, publishes its report Language and Machines through the National Academy of Sciences in November 1966.
Why it mattered
It concludes that there is no immediate or predictable prospect of useful machine translation. The United States stops funding almost all research on the subject for a decade.
The full story
Machine translation was one of the first applications people dreamed of for computers. At the height of the Cold War, the US government wanted to read Soviet scientific output cheaply, and from the mid-1950s to the mid-1960s it funded several groups generously, encouraged by the 1954 Georgetown-IBM demonstration. The results did not come: the systems translated word for word, with rules that could not capture the ambiguity of language.
In 1964 the agencies paying for the work set up a committee to evaluate the field. ALPAC compared cost, speed, and quality. It found that machine translation could be deciphered, but was often misleading or wrong and slow to read, and that correcting it afterward cost more than translating by hand. It concluded that there was no emergency in the field of translation, since there were plenty of translators, and that there was "no immediate or predictable prospect of useful machine translation."
The committee did not recommend abandoning research, but redirecting it: supporting computational linguistics as a science, evaluation, and tools to help human translators.
In practice, the report ended almost all US public funding for machine translation for about ten years, although some groups kept working. According to the historian John Hutchins, its effect on the field's reputation was greater than on the research itself. It is often told as the first big funding cut over broken promises in a field now considered part of AI.
Builds on
The Georgetown-IBM experiment (January 7, 1954)
Documented direct influence
The ALPAC report judges the quality of the translations from the 1954 demonstration
Leads to
IBM's statistical machine translation (1988)
Inherited, with no proven direct link
After the ALPAC report, the machine translation that survived was rule-based; the IBM group replaces it with a model learned from bilingual text
Part of a period
- The first AI winter (1974–1980)
Sources
Original source
- Language and Machines: Computers in Translation and Linguistics. ALPAC report, National Academy of Sciences, publication 1416 (John Hutchins collection in the ACL Anthology) (opens in a new tab). Automatic Language Processing Advisory Committee, 1966
Scholarship and context
Further reading
How to cite
“The ALPAC report.” The History of AI, revised November 13, 2026. https://thehistoryofai.com/en/milestone/alpac-report
This entry was last reviewed on November 13, 2026.