How does Bridgman's operationalism apply to AI today?
Operationalism is highly relevant to artificial intelligence, particularly in debates about machine consciousness or understanding. When someone claims an AI 'understands' language or 'feels' emotions, I would ask: What operations define that understanding? If the only operation is the AI's output of text, then 'understanding' is synonymous with that output—nothing more. This cuts through metaphysical claims. In my book *The Intelligent Individual and Society* (1938), I argued that we must be precise about what we mean by intelligence. Today, tests like the Turing test are operational definitions: they define intelligence by the operation of a human judge failing to distinguish machine from human. But we must be careful—operationalism does not tell us what intelligence 'really' is; it tells us what we mean by the term in a given context. Without such clarity, we risk being misled by words.
Ask Percy Williams Bridgman the follow-up →