An AI-Enhanced Analytic Rubric for Assessing Higher-Order Thinking Skills in Pakistani Higher Education: A Framework Development and Validation Design

Abstract

Artificial intelligence (AI), particularly generative AI, is changing how higher education institutions design learning activities, produce feedback, and judge student work. The change creates an assessment problem: institutions need credible evidence of higher-order thinking (HOTS), while conventional assessment formats can make it difficult to distinguish students’ own reasoning from AI-assisted production. This paper develops the conceptual architecture and validation design of an Artificial Intelligence-Enhanced Analytic Rubric (AI-EAR) for assessing HOTS in Pakistani higher education. The proposed rubric is grounded in the revised Bloom taxonomy, constructive alignment, formative assessment, and human-centred AI principles. It defines five assessable dimensions—analysis, evaluation, creation, metacognitive reflection, and ethical-collaborative reasoning—and assigns AI a bounded decision-support role rather than autonomous scoring. The study proposes four sequential phases: literature and framework synthesis; expert consultation and content validation; pilot administration across purposively selected public and private higher education institutions; and psychometric evaluation through internal consistency, inter-rater reliability, and exploratory factor analysis. Three research questions guide the development and validation process. The paper does not claim empirical validity coefficients because full data collection has not yet been completed. Instead, it presents a transparent, testable validation pathway and specifies the evidence required before the instrument can be recommended for wider institutional use. The framework is designed to support assessment that is criterion-referenced, transparent, ethically bounded, and sensitive to the Pakistani higher-education context.