Dissertation: Examining AI Tool Dependency and Critical Thinking Competency Development in Competency-Based Education: An Applied Study at Western Governors University
Doctor of Business Administration, South College, Knoxville, TN, 2026
Committee Chair: Dr. Damian Dunbar, DBA
The Problem
While teaching prompt engineering to hundreds of students a month at Western Governors University, I kept noticing the same pattern: students would score at full competency on AI-accessible coursework, then struggle on proctored assessments covering that exact same material with no AI access. I call this the competency illusion, and it's a serious problem hiding in plain sight. It threatens the validity of the credential itself, and it's especially troubling in a course explicitly designed to teach AI literacy, since it suggests students can perform well with AI's help without actually building the underlying skill the course is meant to develop.
Why It Matters
This isn't just a classroom curiosity. Competency-based education runs on the premise that passing an assessment means a student can actually do the thing being assessed. If AI-assisted performance is masking gaps in independent ability, that premise breaks down, and it breaks down at scale, across any program that blends AI-accessible coursework with the expectation of independent competency.
The Approach
I grounded the research in cognitive load theory, specifically the distinction between extraneous load (the unhelpful friction good instruction should remove) and germane load (the productive mental struggle that actually builds understanding). My working theory was that AI tools do a great job reducing the first kind of load, but in doing so, they can accidentally eliminate the second kind too, the exact struggle a student needs in order to genuinely learn rather than just produce a correct-looking output.
The study used a mixed-methods design: a quantitative strand correlating AI usage patterns with the size of the competency gap between AI-assisted and AI-restricted assessments, a qualitative strand built on in-depth student interviews and a faculty focus group to understand what's actually happening cognitively when students lean on AI, and an applied strand where an expert panel evaluated the feasibility of proposed solutions.
The Solution
Rather than simply stripping AI access from coursework, which would solve the problem mechanically but work against everything competency-based education is trying to be, I proposed a coordinated set of three interventions that preserve open AI access while closing the gap it creates:
A metacognitive reflection protocol that has students pause before and after assessments to check what they actually know versus what they think they know, a redesign of assessment items that requires students to show their reasoning, identify assumptions, and defend their choices rather than just produce a final answer, and an AI literacy curriculum that teaches students the practical difference between using AI to augment their own thinking and using it to replace that thinking entirely.
The Outcome
The result is a phased, real-world implementation plan, not just a set of recommendations, but a governance structure, a budget range, a timeline, and a way to measure whether it's actually working once it's in place. The goal throughout was to protect what makes AI-integrated education valuable while restoring confidence that a passing grade actually means what it's supposed to mean.