A Cognitive Decision Framework for the AI-Augmented Scholar: Mapping the Suitability Spectrum of Academic Engagement

The integration of generative Artificial Intelligence (AI) into the scientific enterprise marks a profound shift in the cognitive landscape of knowledge production. However, the absence of a formalised framework for engagement threatens the epistemic foundations of research and pedagogy. Through a grounded qualitative investigation of 24 academics across the Natural, Formal, Social, and Applied sciences, we empirically modelled the decision-making processes of the AI-augmented scholar. Here we present a Cognitive Decision Framework and the resulting Suitability Spectrum, identifying ten distinct decision pathways that govern scholarly output. We demonstrate that workflow suitability is not a function of technological fluency, but is determined by a hierarchical interaction between Subject-matter Expertise (SME) and Critical Thinking (CT). Our analysis reveals a pervasive "Expert-Complacency Trap," wherein deep domain knowledge is neutralised by passive AI reliance, resulting in scholarly degradation. We delineate the Human-AI Hybrid state, identifying the Human-Drafted/AI-Refined model as the benchmark for scientific integrity. By formalising the variables that maintain human agency, this framework provides the essential cognitive architecture to safeguard expertise and ensure that the scholar remains the ultimate arbiter of validity in an increasingly automated age.