The ArcProof Verification Framework™: Evidence-Based, Predictively Valid Skill Verification for an AI-Mediated Labor Market

Abstract

This paper introduces the ArcProof Verification Framework: a four-stage model for defining a skill claim, eliciting performance evidence, scoring that evidence with appropriate safeguards, and examining whether the resulting credential relates to outcomes over time. Five commitments: fairness and accessibility, AI scoring governance, learner data rights, temporal validity, and interoperability apply across the four stages. The framework draws on evidence-centered assessment design and argument-based validity theory. Its proposed contribution is to connect an individual skill claim to performance evidence and then test the relationship between the credential and relevant outcomes. That final relationship has not yet been established. This paper sets out the framework and the research needed to evaluate it.