What Computers Can’t Do
The philosopher of science Hubert Dreyfus’ influential 1972 book, What Computers Can’t Do: A Critique of Artificial Intelligence, argues that logic-based AI cannot replicate human intuition and contextual understanding.
Dreyfus first articulated this position in his combative 1965 review of Alan Newell and Herbert Simon’s AI research for the RAND corporation, the national research thinktank offering analysis to the US military.
Drawing on phenomenology, Dreyfus suggests that human intelligence is fundamentally different from computer intelligence and that, without embodied knowledge, computers will be incapable of intellectual tasks that require intuition and experience.
Expanding this argument in What Computers Can’t Do, and its sequel What Computers Still Can’t Do, Dreyfus predicts that logic-based AI will fail to achieve robust flexibility and intuitive sense-making because of its faulty account of ‘intelligence as a passive receiving of context-free facts into a structure of already stored data’ (Dreyfus, 1992: 34).
Human learning, from his perspective, entails broad embodied “know-how” shaped by contextual particularities, unfolding moods, and sensory-motor skills. A key problem with symbolic processing AI’s model of learning is that ‘one cannot substitute an extractable web of beliefs for the whole cloth of our concrete everyday experiences’ (Dreyfus, 1992: 54).
Published in 1985, Dreyfus’s follow-up book, Mind Over Machine: The Power of Human Intuition and Expertise in the Era of the Computer, co-authored with his brother Stuart Dreyfus, argues that AI expert systems require anticipation of all possible contingencies in order to code ‘rules for response’ and, as such, the litheness and immanent insight of expert human intuition is ‘forfeited’ (Dreyfus and Dreyfus, 1985: 31).