Hubert Lederer Dreyfus — phenomenologist at the University of California, Berkeley — was the most prominent philosophical critic of classical Artificial Intelligence. His work What Computers Can’t Do (1972, revised 1992 as What Computers Still Can’t Do) formulated the counter-thesis in a time of AI optimism: human intelligence cannot be formalized into explicit rules.
Key Contribution
Dreyfus draws on Martin Heidegger and Maurice Merleau-Ponty: human understanding is rooted in a pre-predicative, bodily situated lifeworld. We grasp meaning not by following rules, but because we are together in a world in which things encounter us in their involvement. A hammer is for hammering, not the sum of its features. This embeddedness in practice cannot be symbolically encoded.
The five-stage theory of expertise (Dreyfus/Dreyfus 1986) shows how human skillful mastery progresses from rule application to situational intuition — experts no longer act according to rules, but out of a bodily feel for the situation. This feel cannot be formalized.
Significance for the Ontology of the Person
Dreyfus’ position is highly compatible with the ontology of the person, albeit from a phenomenological rather than a substance-ontological perspective. Both approaches reject the empirical-functionalist concept of person: personhood realizes itself in a relation to the world that presupposes bodily being-here, interiority, and intentionality.
Notably: even in the age of neural networks, Dreyfus’ critique remains pertinent. VLA models (Vision-Language-Action) such as Helix or GR00T likewise operate not out of situated worldliness, but out of statistical patterns. They simulate the result of bodily understanding without realizing the understanding itself. A humanoid robot is not ready-to-hand in the world, but present-at-hand — it fulfills functions, but it does not dwell.
Place in the Book
Dreyfus adds to the book a line of tradition that mobilizes genuinely phenomenological resources for the critique of AI with Heidegger and Merleau-Ponty. While Spaemann makes the case for substance metaphysics and Searle argues analytically, Dreyfus shows: a genuinely continental phenomenology, too, arrives after thorough examination at the same conclusion as realistic phenomenology — personhood is not an arrangement of functions.
See also
- Artificial Intelligence
- AI Ethics
- Humanoid Robot
- Intentionality
- Empirical-Functionalist Concept of Person
- Personhood
- Martin Heidegger
- John Searle
- Alan Turing
Sources: Bexten 2017, pp. 95–97 (reception of Dreyfus’ critique of AI in the context of the forgetfulness of the person).
Further sources:
- Dreyfus, Hubert L. (1972): What Computers Can’t Do: A Critique of Artificial Reason. Harper & Row.
- Dreyfus, Hubert L. (1992): What Computers Still Can’t Do: A Critique of Artificial Reason. MIT Press.
- Dreyfus, Hubert L. / Dreyfus, Stuart E. (1986): Mind over Machine: The Power of Human Intuition and Expertise in the Era of the Computer. Free Press.
- Dreyfus, Hubert L. (1991): Being-in-the-World: A Commentary on Heidegger’s Being and Time, Division I. MIT Press.