Reflective Adaptive Leadership in AI-Augmented Decision Environments

Cognitive Sovereignty as a Theoretical Extension

Abstract

Artificial intelligence (AI) is increasingly embedded in professional decision-making, creating a leadership challenge that extends beyond technical implementation to the preservation of reflective human judgement. This study extends Reflective Adaptive Leadership (RAL) by introducing Cognitive Sovereignty (CS) as an AI-specific construct describing a professional's capacity to consciously evaluate, question, and appropriately accept or override AI-generated recommendations while retaining decision ownership. The study aims to examine whether Cognitive Sovereignty moderates the relationship between AI Dependence and Decision Quality in AI-augmented decision environments. An exploratory quantitative design was employed using a 64-item questionnaire administered to 321 professionals across six occupational sectors. The instrument measured AI Dependence, Cognitive Sovereignty, Decision Quality, and Organisational Resilience. Multiple regression analysis was conducted to examine the proposed moderation relationship. The main-effects model was statistically significant, whereas the hypothesised moderation effect was not supported. Reliability and sampling adequacy analyses indicated that the instrument requires further psychometric refinement before confirmatory validation. Despite these measurement limitations, the findings support the conceptual relevance of Reflective Adaptive Leadership as a theoretical framework for understanding reflective human judgement in AI-assisted decision-making. The study contributes to leadership and AI governance literature by proposing Cognitive Sovereignty as a novel theoretical extension of RAL and provides a transparent roadmap for future instrument development and empirical validation.