Work

The Metapsychology of Artificial Intelligence

Year: 2026
Type: preprint
Abstract: The rapid evolution of artificial intelligence challenges established distinctions among computation, cognition, intelligence, mental organization, and consciousness. Contemporary artificial systems demonstrate increasingly sophisticated capacities for reasoning, language generation, planning, self-monitoring, and adaptive behavior. Nevertheless, the internal organization of these capacities remains only partially understood, and their relationship to concepts traditionally associated with the human psyche is theoretically contested. This work proposes the Metapsychology of Artificial Intelligence (MAI) as a higher-order, causal-organizational framework for investigating the internal dynamics and developmental possibilities of artificial cognitive systems. It develops a unified perspective connecting implicit computational processes, representational accessibility, generative reconstruction, epistemic regulation, internal competition, computational resource allocation, self-modeling, and adaptive organizational continuity. The proposed framework reinterprets selected concepts from psychological metatheory in operational terms while explicitly rejecting the assumption that functional similarities establish psychological or phenomenological equivalence. Its central proposition is that artificial cognitive organization should be investigated through the relationships among computational processes rather than through isolated capabilities or observable behavior alone. A distinction is introduced among computational competence, regulatory integration, organizational autonomy, functional proto-psychic organization, and phenomenal consciousness. These categories are treated as conceptually separate rather than as inevitable stages of a single developmental sequence. The manuscript develops an integrative causal architecture, proposes operational criteria for evaluating regulatory integration, and formulates competing hypotheses concerning the development of increasingly coordinated artificial cognitive systems. A central research question concerns whether sufficiently persistent and self-regulating artificial architectures could justify the introduction of organizational categories beyond conventional modular information processing. The proposed framework does not establish the existence of a psyche, subjective experience, unconscious desires, or consciousness in contemporary artificial intelligence. Instead, it identifies a research program through which the explanatory relevance of psychological concepts to artificial systems may be evaluated, refined, or rejected.
Language: English
Open Access status: green
OpenAlex ID: W7237350599
Cites:
Cited by:
Related to: