**Title: Moving Past Tool Adoption: A Five-Stage Developmental Path for AI Literacy in Higher Education

Abstract

AI literacy is widely acknowledged as an essential competency for university graduates. However, student interaction with AI technologies tends to fall into two distinct, problematic categories: complete avoidance, stemming from fear, mistrust, ethical reservations, or lack of access; or uncritical dependency, which generates fluent content while concealing significant gaps in understanding. While current AI literacy frameworks offer useful definitions of competencies, they often lack the specificity needed to diagnose a learner’s starting point or map their progression toward responsible, critical engagement.

To address this gap, this study introduces a five-stage AI Literacy Continuum designed for higher education. The stages are: 1) Not Yet Engaged, 2) Uncritical Use, 3) Informed Use, 4) Critical Evaluation, and 5) Improvement. This model serves as a practical diagnostic and instructional roadmap, complementing existing dimensional frameworks and aligning with international standards such as those from UNESCO and the OECD.

The paper details a design-based implementation case study conducted at North Carolina State University. Between Fall 2024 and Spring 2026, credit-bearing courses and intensive hands-on workshops involved over 330 participants. Due to the absence of a validated pre/post assessment instrument or a control group, the results are presented as observational and practice-based. The data suggests that participants generally shifted from non-engagement or uncritical use toward informed engagement. Furthermore, sustained, discipline-specific experiences provided stronger evidence of development in critical evaluation and improvement-oriented practices. The discussion covers curricular pathways, equity issues, and assessment methods, arguing that AI literacy must be viewed not merely as the adoption of tools, but as a developmental capacity to understand, evaluate, and responsibly apply AI systems within both disciplinary and societal contexts.

Source: arXiv Generated at: 2026-06-02 00:00:00 UTC