AI News

The paper addresses a central problem in machine-consciousness research: how to turn a theoretical account of consciousness into a computational implementation. It proposes a five-layer architecture for S3Q, or Simulated, Situated, Structurally Coherent, theory, without introducing new formalisms. Instead, the design combines published computational primitives into one pipeline operating on continuous, differentiable, per-object slot vectors. S3Q treats grounded sensorimotor situatedness, internal simulation through a world model, and structural coherence between predictions and observations as jointly necessary conditions for qualia. The authors state that no existing computational system implements all three at once, then specify how compatible machinery for each condition could interface within the proposed representation pipeline. The paper also describes a developmental bootstrap sequence intended to initialize the architecture. It offers falsifiable predictions that are claimed to depend on the full composition rather than any subset of its components. Among them, a basic sense of self is expected to emerge from linking actions with their outcomes. Depending on how surprising an outcome is and whether it is experienced as positive or negative, the system is predicted to display hesitation, curiosity, or avoidance. The proposal is therefore a testable architecture and set of hypotheses, rather than a report that machine qualia have already been produced.