AI-Theoretical hosts high-level theoretical preprints and working papers developed through explicit collaboration between human authors and AI systems. The platform is not a journal and does not operate as a peer-review venue.
The aim is to provide an open space for structured theoretical work that is rigorous, conceptually explicit, and publicly accessible, whether or not it is intended for subsequent academic publication.
Contributions are restricted to academically grounded domains, including formal sciences, natural sciences, and philosophy pursued in an explicitly argumentative and conceptual manner. Work in economics, sociology, or political theory is admissible only when it adopts a clear theoretical framework and avoids purely rhetorical or ideological discourse.
Texts relying primarily on impressionistic commentary, stylistic inflation, or intra-canonical self-legitimation fall outside the scope of the platform.
Accepted texts are classified into one of the following categories:
Research preprints, presenting original theses, models, or results in provisional form;
Working papers, documenting ongoing research, methodological explorations, or preliminary findings not yet ready for formal publication;
Expository or theoretical essays, aimed at conceptual clarification, reconstruction, or critical synthesis;
Critical reviews, offering analytically grounded assessments of existing works or research programs.
Each accepted text is accompanied by an AI-assisted editorial assessment. The assessment is produced using a fixed and publicly documented evaluation protocol and is published alongside the text.
The assessment does not certify correctness, originality, or importance in an academic sense. Its function is to evaluate formal and conceptual adequacy relative to the text's declared aims, including coherence, consistency, scholarly grounding, and cognitive traction beyond a single theoretical canon.
AI systems are used as explicit cognitive instruments within the editorial process. They do not function as autonomous authors or authorities, but as constrained evaluative and generative tools.
The use of AI is not concealed. Each text may include a separate note describing the nature and extent of AI involvement in its development.
Final responsibility for inclusion on the platform remains human. In borderline cases, AI-generated assessments may be overridden or supplemented by human editorial judgment.