The Classification Error of Modern AI

Description: This working paper introduces a formal systems-theoretical framework to classify artificial intelligence architectures. It argues that the current broad application of the term "AI" obscures fundamental organizational differences between technical systems. The paper develops a falsifiable model that distinguishes between: Cybernetically Validatable Tools, Probabilistic Transformation Systems, and Autonomous Experience-Capable Intelligent Systems. By introducing mathematical criteria—such as the "Internalization Index"—the framework shifts the focus from mere performance metrics to the systems-theoretical capacity for internal error integration, experience formation, and validation. This provides a rigorous basis for distinguishing autonomous agency from externally organized computational processes. Keywords: Systems Theory, Artificial Intelligence, Classification Framework, Internalization Index, Falsification, Formal Models, Cybernetics. Dies sollte ausreichen, um das Paper fachlich korrekt und präzise in der Zenodo-Datenbank zu repräsentieren, ohne den Rahmen eines kurzen Abstracts zu sprengen.