This accepted manuscript is structurally governed by THE META-INDEX (Zenodo DOI: 10.5281/zenodo.18169167) This paper examines the integration of artificial intelligence into human cognitive processes through two conceptual orientations: Instrumental Augmentation and Cognitive Delegation. The central argument is that AI should be understood as a high-capacity cognitive tool that supports reasoning, rather than as an authority that supplants it. The paper introduces the Calculator Paradox as its primary theoretical contribution: a tool amplifies cognitive capacity as long as the user continues to reason independently; when the tool begins to replace deliberative responsibility, the risk of cognitive dependency and epistemic loss of control emerges. The Calculator Paradox is distinguished from the existing automation bias literature by its focus on a deeper epistemic mechanism: not merely over-reliance on outputs, but the progressive atrophy of the capacity to evaluate those outputs — a self-undermining dynamic that automation bias, as classically defined, does not capture. Drawing on cognitive psychology and the philosophy of mind, the paper examines how AI systems intensify this dynamic by producing not merely outputs but rhetorically structured justifications that carry the surface form of reasoned judgment. The paper develops the concepts of cognitive hygiene and oversight ethics, introduces epistemic sovereignty as a normative framework, and proposes cortical civilisation as a broader social orientation for governing human-AI relations. Four empirically testable hypotheses are identified for future research. Philosophy of AI : Submission ID'yi (12199)