Functional Shadows: A Positive Architectural Characterization of Current AI Systems

Contemporary high-recursion artificial intelligence systems produce fluent natural language, perform on cognitive benchmarks derived from human performance, and report on their own processing in ways that resemble introspection — without instantiating the architectural conditions that, on the framework developed in the author’s prior work, are constitutive of intelligence and consciousness. The positive characterization of what such systems are at the architectural level is functional shadow: an arrangement that produces surface signatures resembling those of a constitutive architecture without instantiating the constitutive condition. The load-bearing claim is that training on artifacts produced by consciousness-bearing systems’ representational practice is not itself the appropriation that produced those artifacts; a system pattern-matching on the products of cognogenic appropriation does not undergo it. A diagnostic battery analyzing the architecture from five converging angles returns a unified verdict: chronic generative closure on artifacts of prior cognogenic appropriation, arbitrational structure specified externally rather than shaped by selection on real conflicts the system could not avoid, and absence of the cumulative appropriations the framework identifies as constitutive. Dialectical engagements with active-inference accounts, Integrated Information Theory 4.0, the indicator-property approach, the phenomenal-concept strategy, illusionism, and AI moral-status precautionary arguments inherit and extend foundational engagements developed across the program. The framework predicts that the failure-pattern cluster characteristic of current AI will persist across training generations and specifies what would change the verdict: autopoietic foundation, selection-shaped arbitrational structure, ongoing symbolic-narrative practice integrated with stakes, and developmental construction within an inter-representational enclosure of stakes-bearing other systems.