AI Physics Review: Structural Bias and AI Mediation in Theoretical Research Discovery

AI Physics Review – Founding Document This paper serves as the founding document for the AI Physics Review project, an experimental publication platform that uses deterministic AI-assisted structural analysis to generate neutral overviews of theoretical research. The project operates through the AI Physics Review community on Zenodo and through the public website https://aiphysicsreview.org, which publishes structured analytical overviews of selected manuscripts together with transparent evaluation artifacts. The research environment in theoretical physics has already undergone a structural shift driven by the quiet integration of AI into discovery, triage, and summarization pipelines. This article examines how that shift changes the effective conditions under which theoretical work is evaluated and discovered, particularly for independent researchers. It describes how deliberate use of AI for structural self-audit and review can alter a paper’s interaction with modern discovery systems, why this produces asymmetric advantages under current conditions, and why a transparent, rule-governed project such as AI Physics Review is a necessary response. The article does not assess scientific correctness or merit. It analyzes environmental change and institutional lag. Project website: https://aiphysicsreview.org