Closing the Distance: Explainable AI for Diagnosing Neurotypicality in Disability Assessors
This paper proposes a complex machine learning framework to diagnose high-risk neurotypicality in clinicians, bureaucrats, data scientists and 'innovation leaders' involved in algorithmic disability assessment. We hypothesise that the urge to algorithmically contain disabled people functions as a psychological distancing technologyâallowing professionals to comply with policy without risking actual relationship. Our TabPFN-NeuroMix model outputs a Neurotypicality Proximity Avoidance Index (NPAI) with explainability via SHAP, repurposed as Shapley Ableism Partitioning.