How to select the right AI tool for the job? An Autonomy-values framework for educational AI with a focus on academic advising

This study presents an Autonomy-values framework for analysing, designing, and selecting AI academic advising systems. Through focus groups with advisors using Nominal Group Technique, we identified contextually specific values that were then generalised using Legitimation Code Theory's Autonomy lens. The resulting framework maps advising values across four positions: Student-Centred, Advisor-Centred, Institution-Centred, and Social-Centred. By combining values-sensitive design with LCT Autonomy, this work advances knowledge on human-centred approaches to AI implementation in academic advising and broader educational contexts.