Algorithmic and technical-mechanism concepts are marginal in students' AI networks: machine learning was produced by a single participant across 436 respondents¶

Claim · All claims
Evidence · 1 study (2 entries) · associational r1–r2 · q2
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person · recheck by 2031: baselines and conditions change

Subclaims¶

q2 i? The Reasoning category yielded only f = 24 responses across 436 participants (Algorithm f = 10, Logic f = 8, Automation f = 5, Machine Learning f = 1), while anthropomorphic Mind (f = 47) dominates the Intelligence category. → İsmail Şan and Hanife Gülhan Orhan Karsak 2026

Evidence¶

İsmail Şan and Hanife Gülhan Orhan Karsak 2026¶

İsmail Şan and Hanife Gülhan Orhan Karsak. (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. https://doi.org/10.1186/s41239-026-00621-w

q2 · i? · associational · r1

Frequency analysis of the Mental & Technical Capacity theme in the WAT corpus (N = 436) shows the Reasoning category produced just f = 24 responses, and "machine learning - was produced by a single participant across 436 respondents". Intelligence (f = 55) consists almost entirely of Mind (f = 47); no effect size is printed.

"The code with the highest surface validity as an AI literacy concept - machine learning - was produced by a single participant across 436 respondents."

İsmail Şan and Hanife Gülhan Orhan Karsak 2026 (2)¶

İsmail Şan and Hanife Gülhan Orhan Karsak. (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. https://doi.org/10.1186/s41239-026-00621-w

q2 · i? · associational · r2

The authors' interpretation of the frequency pattern: anthropomorphic attribution coexists with absence of procedural or algorithmic concepts, consistent with the black-box orientation described in the literature review.

"Students are willing to attribute mind-like qualities to AI but are not activating the procedural or algorithmic concepts that would support critical understanding of how AI produces its outputs."

Discussion¶

Related Claims¶

• Across 133 K-12 ML activities, a technical and a societal perspective on the same algorithm co-occur in roughly 3% of cases — related
• Of 89 coded supervised learning activities, 69 confine learners to the User role, 17 reach Creator, and 8 reach Machine — related
• Turkish undergraduates' AI associations are anchored by utilitarian concepts, with Convenience and Speed alone accounting for 18.5% of all responses — related
• AI governance vocabulary (bias, transparency, privacy, accountability, regulation, fairness) is a structural zero in students' AI associative networks — related