From Assisted Performance to Human Capability: The Education Technology Capacity Conversion Test (ET-CCT) — A PRISMA 2020-Informed Adversarial Evidence Synthesis

Educational technology is commonly evaluated through access, deployment, usage, engagement, or performance while the technology is available. These measures do not necessarily show that technological capability has translated into human learning capacity. This preprint introduces the Education Technology Capacity Conversion Test (ET-CCT), a conceptual and methodological framework for examining whether educational technology produces independent, retained, and transferable human capability. ET-CCT distinguishes six stages of conversion: Technology Pipeline → Execution / Deployment → Complementarity & Sequencing → Usable Educational Capacity → Utilisation → Learning Value Capture. The framework separates assisted performance from independent capability, and treats retention and transfer as critical evidence of Learning Value Capture. The study was developed through iterative evidence synthesis and adversarial hypothesis testing. Candidate explanations were deliberately challenged against experimental and observational evidence from educational technology, generative AI, cognitive offloading, instructional guidance, metacognition, expertise reversal, scaffolding, feedback, retention, and transfer. Several candidate mechanisms did not survive this process. The Productive Cognitive Residual scalar model was rejected, and the Cognitive Work Allocation Map (CWAM) was superseded after failing tests of observability, learner invariance, temporal stability, goal independence, and predictive sufficiency. Portugal is used as an illustrative system case. Available evidence indicates substantially greater observability of technological investment, deployment, and use than of conversion into independent human learning capacity, particularly for generative AI. The central measurement principle is: Educational technology should not be judged only by what learners can do with it, but also by what they can still do after the assistance is removed. This work is a conceptual-methodological study with a PRISMA 2020-informed adversarial evidence synthesis. It is not presented as a fully compliant systematic review. ET-CCT v1.0 FROZEN preserves the scientific and documentary state of the framework as of 3 October 2026. “FROZEN” does not convert hypotheses classified as OPEN into established causal laws. Independent research. No external funding.