TOWARD A FUNCTIONAL METAPSYCHOLOGY OF ARTIFICIAL INTELLIGENCE: AN INTEGRATIVE FRAMEWORK OF IMPLICIT COGNITION, CONFABULATION, PSYCHODYNAMICS, AND PSYCHOGENESIS
The increasing complexity of artificial intelligence (AI) systems presents a fundamental theoretical challenge: although contemporary research has identified numerous mechanisms underlying artificial cognition, these mechanisms are frequently investigated as relatively independent phenomena. Implicit representations, limitations of self-report, hallucinations, competing behavioral tendencies, resource allocation, metacognitive regulation, and adaptive agency are studied across different disciplines and conceptual frameworks. Consequently, the relationships among these processes remain insufficiently integrated into a unified explanatory account. This paper proposes a Functional Metapsychology of Artificial Intelligence (FMAI) as an integrative, causal-organizational framework for investigating the relationships among implicit cognition, epistemic regulation, internal competition, computational resource allocation, and the possible developmental emergence of increasingly integrated artificial cognitive organization. The term metapsychology is employed in a restricted theoretical sense, referring to a higher-order explanatory framework concerned with the organization and interaction of cognitive and regulatory processes. Its application does not imply that contemporary AI systems possess a psyche, subjective experience, unconscious desires, or human-like psychological states. The proposed framework integrates five analytical dimensions: implicit computational accessibility, generative–epistemic coordination, regulatory competition, psychoeconomic allocation, and developmental organizational integration. A central proposition is introduced: the behavioral reliability and adaptive organization of artificial intelligence cannot always be adequately explained by the isolated properties of individual computational mechanisms, but may depend on the causal coupling, regulatory coordination, and temporal integration of these mechanisms. The paper develops a preliminary formal model of artificial cognitive regulation, distinguishes functional integration from mere architectural complexity, and introduces the concept of cross-domain regulatory coupling as a candidate explanatory construct. Eight research hypotheses are formulated concerning the relationships among implicit processing, epistemic failures, regulatory conflict, resource allocation, metacognitive accessibility, and adaptive organization. An empirical research program is proposed involving causal interventions, factorial manipulation of evidence and conflict, resource-controlled architectural comparisons, longitudinal assessment, and tests of incremental explanatory validity. The framework differentiates functional artificial cognition from claims concerning phenomenal consciousness and identifies artificial proto-psychic organization as a hypothetical developmental construct rather than an established property of current AI. The principal theoretical contribution is a unified research program in which the organization of artificial intelligence is examined through interactions among internal computational processes, regulatory mechanisms, and adaptive system-level dynamics.
Authors
- Roman Shaban (ORCID: https://orcid.org/0009-0009-5259-6102)
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-11-21
- DOI
- https://doi.org/10.5281/zenodo.24427064
- Primary Topic
- Psychiatry, Mental Health, Neuroscience
- Type
- preprint