Abstract

This article investigates the fundamental transformation of contemporary scientific practice under the aegis of Artificial Intelligence (AI), using the philosophies of Karl Popper and Thomas Kuhn as coordinating axes. The central problem revolves around the distinction between falsifiable theories and non-falsifiable systems (“black box” structures), questioning whether AI represents an extension of “normal science” or an absolute paradigmatic rupture. Through Popper’s critical rationalism, the crisis of demarcation is examined in a context where large-scale statistical induction replaces logical deduction and human conjecture. Parallel to this, through Kuhn’s lens, the article analyzes the incommensurability between science based on universal laws and science based on high-dimensional correlations. The text incorporates the views of classic commentators such as Imre Lakatos and Paul Feyerabend, cross-referencing them with contemporary philosophers of science like Nancy Cartwright, Nick Bostrom, Susan Schneider, and Ray Kurzweil. It discusses the nature of evidence, the redefinition of empirical validation, and the influence of social and historical contexts on the validation of knowledge. It concludes that AI is not merely an instrument but a new disciplinary matrix that demands a revision of the ethical and methodological foundations of the philosophy of science.