Diagnosis and Permission: A Methodological Synthesis of ARMS and AIAS

This paper develops a methodological synthesis of two practical instruments for AI-resilient assessment design in higher education: the AI Risk Measure Scale (ARMS), developed by De Vita and Brown at the University of Greenwich, and the AI Assessment Scale (AIAS), developed by Perkins, Furze, Roe, and MacVaugh. Each scale has been adopted at hundreds of institutions, but no published methodology has integrated the two into a single workflow. This paper closes that gap.The contribution has three components. First, a four-step ARMS × AIAS workflow — Diagnose, Redesign, Declare, Communicate — that specifies the actor, tool, and output of each step. Second, four decision rules (DR1–DR4) that govern bidirectional rescoring between the two scales, role separation between program leaders and module leaders, and the visibility boundary between faculty-internal diagnosis and student-facing permission. Third, a fictional worked example demonstrating the workflow's behavior across a six-assessment senior undergraduate capstone, including the non-obvious case in which a maximally vulnerable assessment receives expanded AI permission under structured disclosure rather than the default prohibition response.The synthesis is methodological, not empirical. A one-page decision sheet (Appendix A) is included as a printable reference card for educators and program leaders adopting the workflow. The paper introduces no new empirical data and closes with named directions for empirical follow-up.Intended audience: higher-education faculty, program leaders, instructional designers, and assessment-policy researchers working under generative-AI conditions.