Hubert Dreyfus

Hubert Lederer Dreyfus (1929 – 2005) was an American philosopher and professor at UC Berkeley who became one of the most important critics of artificial intelligence and cognitive science. His seminal work, 'What Computers Can't Do' (1972), challenged the dominant assumptions of early AI research, arguing that human intelligence fundamentally depends on embodied experience, intuition, and contextual understanding that machines cannot replicate. Drawing on phenomenology and the philosophy of Maurice Merleau-Ponty, Dreyfus demonstrated that human expertise develops through stages of skill acquisition that cannot be reduced to explicit rule-following—insights later validated by cognitive science research. His five-stage model of skill acquisition (novice, beginner, competent, proficient, expert) became influential across fields including education, nursing, and military training. Though initially criticized by AI enthusiasts, his skepticism has been substantially vindicated as researchers recognize the limitations of purely computational approaches to intelligence. Dreyfus's work established that human understanding involves embodied, social, and emotional dimensions that resist formalization. He taught at Berkeley for decades, influencing generations of philosophers, computer scientists, and cognitive researchers. His legacy represents a crucial corrective to reductionist approaches in cognitive science and artificial intelligence, affirming the irreducible complexity of human consciousness and expertise.