Between 1946 and 1953 the Josiah Macy, Jr. Foundation sponsored a series of ten conferences, held under the heading “Circular Causal and Feedback Mechanisms in Biological and Social Systems,” that became the founding meetings of cybernetics. They brought together an unusually broad group, including the mathematician Norbert Wiener, the neuroscientist Warren McCulloch, the mathematician John von Neumann, the anthropologist Margaret Mead, and many others from biology, psychology, engineering, and the social sciences, to hammer out a shared vocabulary for how systems control themselves.
The conferences were where ideas like feedback, information, and self-regulation were first debated as general principles applicable equally to machines, brains, and societies, rather than as topics belonging to one discipline. The participants worked out how to talk about goal-directed behavior in mechanical and biological systems using the same terms, and these concepts went on to shape neuroscience, control engineering, the social sciences, and the early thinking that fed into artificial intelligence. The complete transcripts were later edited by Claus Pias and published as “Cybernetics: The Macy Conferences 1946-1953,” which serves as the primary source here.
The meetings matter to the prehistory of artificial intelligence because they created the interdisciplinary culture in which questions about mind, machine, and information could be asked together. McCulloch and Pitts’s model of the neuron, Wiener’s cybernetics, and von Neumann’s work on automata all circulated through this community before AI was named in 1956.
For a general reader, the Macy Conferences are a striking example of how a shared vocabulary, invented across disciplines, can open up entirely new fields, in this case the study of control and communication that underlies modern AI.