Knowing Its Name, Not Its Nature: Word Association Mapping of Student AI Cognition and Evidence-Based Micro-Credential Design in Turkish Higher Education

Synthesis: İsmail Şan and Hakan Orhan Karsak (2026) used the psycholinguistic Word Association Test (WAT) to map the cognitive representations of AI held by 436 Turkish university undergraduates, producing a needs assessment for designing AI Literacy micro-credential programs grounded in learners' actual knowledge structures. Across a seven-stage analytical protocol (1376 coded responses, inter-rater reliability κ=0.87), the authors found students' associative networks were dominated by utilitarian concepts — convenience, speed, and technology — while algorithmic transparency, ethical governance, and technical mechanisms were structurally absent. Network analysis revealed strong polarisation between a positive-utility cluster and a negative-risk cluster (τ=−0.819), indicating a pervasive "black-box" orientation toward AI. The findings ground an Ethical AI and Workforce Readiness micro-credential framework that deliberately bridges students' experiential, instrumental knowledge of AI and the ethical and AI Governance frameworks needed for responsible use.

Key Findings

• Cognitive fragmentation, not ignorance: Students' AI associations clustered around instrumental utility (convenience f=145, speed f=110, together ~18.5% of all responses), while technical-register and governance concepts were not merely underrepresented but structurally isolated from the dominant utilitarian cluster by a strong negative association (τ=−0.819). The authors argue this reflects two processed-as-mutually-exclusive cognitive frames, not a simple deficit.
• A "black-box" orientation toward AI: Algorithmic transparency, data privacy, model uncertainty, and accountability were absent from students' associative networks, confirming a pervasive but purely functional understanding of how AI systems work.
• A replicable diagnostic methodology: The study contributes the first large-scale WAT-based cognitive mapping of AI knowledge structures in Turkish higher education, demonstrating a data-driven needs-assessment method (canonical coding, Kendall's τ, multidimensional scaling, threshold-network construction) for evidence-based credential design.
• Design implication — ethics modules need bridges: Delivering an ethics module in isolation risks encountering a cognitive architecture with no existing schema to receive it; curricula must explicitly construct connections between learners' experiential tool knowledge and ethical/governance frameworks.
• Curricular priority shift: Rather than reproducing what students already know, credential programs should sequence content to close the specific structural gap between utilitarian and ethical dimensions of AI literacy.

What this means for practice

• Curriculum designers. Open AI literacy content with the experiential situations students already associate with AI — convenience, speed, efficiency at work — then build outward to how that convenience is produced, who pays the costs, and what governance makes it fair, rather than starting from abstract definitions of machine learning.
• Do not run ethics as an isolated module: the associative data show utilitarian and risk/governance concepts as structurally segregated frames (τ = −0.819), so the credential must construct the bridges between tool experience and transparency, privacy, and governance.
• Educators. Administer an associative diagnostic such as the word association test at course entry to surface the actual structure of learners' prior knowledge instead of assuming a content deficit.
• Require a structured ethical impact assessment of an AI application in the student's own field — algorithmic bias and fairness, data privacy and consent, accountability and audit mechanisms, and the regulatory landscape — with a written recommendation to a named decision-maker.

Limitations

• The 436 undergraduates were unevenly distributed: 80.3% from social sciences and humanities, 12.2% from natural and applied sciences, and 7.6% from health sciences, with 211 students (48.4%) from Kırklareli University and only nine (2.1%) from Hatay Mustafa Kemal University.
• The study collected no age, gender, or prior AI-tool-experience data, so the authors report those variables remain unexamined; the WAT also captures spontaneous recall at a single time point, and only longitudinal pre/post WAT designs could attribute cognitive change to a credential.
• The multidimensional scaling solution returned a stress value of 0.396, and the authors describe spontaneous recall as a measure with inherent limits that require methodological triangulation.
• The proposed micro-credential framework has not been tested: the authors state its effectiveness is yet to be empirically validated and call for employer surveys to establish whether the portfolio criteria match professional standards.

Connected Concepts

• AI Literacy
• Higher Education
• Generative AI
• Ethics
• AI Governance
• Curriculum Design
• Assessment
• Student Experience

Connected Articles

• AI Literacy Interventions in Education: A Meta-Analysis of Effects and Moderators — AI Literacy Interventions in Education: A Meta-Analysis
• Enhancing AI Literacy Course Satisfaction Through Empowerment in AI Problem-Solving and Ethical Awareness: Development and Validation of an AI Project-Based Learning Scale — Enhancing AI Literacy Course Satisfaction Through Empowerment in AI Problem-Solving
• What Does the Credential Still Certify? Cognitive Stewardship for AI-Mediated Education — What Does the Credential Still Certify? Cognitive Stewardship
• Trust and Reliance on AI in Education: AI Literacy and Need for Cognition as Moderators — Trust and Reliance on AI in Education

Citation

Şan, İ., & Orhan Karsak, H. (2026). Knowing its name, not its nature: Word association mapping of student AI cognition and evidence-based micro-credential design in Turkish higher education. International Journal of Educational Technology in Higher Education, 23, 46.