Metaphysical Assumptions as the Foundation for Artificial Intelligence

Abstract

Artificial Intelligence is often presented as a purely technical and mathematical enterprise. Yet beneath algorithms, data, and artificial neural networks lie deep metaphysical commitments about mind, matter and causation. This paper argues that AI systems are not metaphysically neutral. They presuppose specific answers to classical philosophical questions: For instance, what is intelligence? Can it be separated from embodiment? Is the mind reducible to computation? By tracing the historical and conceptual links between metaphysics and AI, the paper argues that Cartesian dualism, functionalism, and materialism shape AI’s design and goals. Understanding these hidden assumptions is crucial for evaluating AI’s capabilities, limits, and social impacts. The paper concludes that responsible AI development requires explicit engagement with metaphysics and not neutral nor avoidance of it.