Ontological Commitments in Formal Models of Artificial General Intelligence

The paper examines the ontological commitments embedded in two canonical frameworks of Artificial General Intelligence: AIXI and the Gödel Machine. Its central aim is to identify which entities, structures, and assumptions each framework must presuppose for its core claims to hold. To do so, it adopts Quine’s criterion of ontological commitment and proposes an ontological signature composed of four dimensions: the presupposed theory of agency, the structure of the domain over which the model quantifies, the ontological status of internal states, and the formal criterion of intelligence.