GenAI Knowledge, Epistemic Orientation, and Intellectual Values Predict Undergraduate Students' Critical GenAI Use

Markus H. Hefter, Benjamin Paaßen & Kirsten Berthold (2026) — AI Educ. (MDPI), 2, 27. Open Access, CC BY 4.0. doi:10.3390/aieduc2030027.

📄 Full text (MDPI, OA)

Summary

A correlational study (N = 67 undergraduate psychology students, Bielefeld University) testing two protective factors against uncritical GenAI overreliance: (1) knowledge about genAI and (2) the disposition to engage in critical thinking — operationalised via Kuhn's framework as epistemic orientation (tendency away from absolutist toward evaluativist beliefs) and intellectual values (viewing intellectual engagement as worthwhile). Both factors are framed as components of AI literacy and both are trainable, motivating intervention recommendations.^[raw/papers/mdpi-2026-critical-genai-use-predictors.md]

Method highlights

• Critical use of genAI measured with a newly developed 10-item scale (α = .78) covering validation behaviours (cross-checking outputs, verifying factual accuracy, comparing with literature).

• Knowledge measured two ways — the study's key methodological contribution: self-reported (18 SNAIL items, α = .90) AND objective/performance-based (12 items from Köhler & Hartig's ChatGPT scale + Hornberger et al.'s AI literacy test), each objective item paired with a certainty rating to compute a misconception score (correctness × certainty; Eitel/Hefter procedure).

• Dispositions: epistemic orientation (5 items, α = .70), intellectual values (14 fictive-person items, α = .90), plus need for cognition (NFC, 4-item short scale, α = .54).^[raw/papers/mdpi-2026-critical-genai-use-predictors.md]

Key findings

H1 — GenAI knowledge predicts critical use: SUPPORTED

• Self-reported knowledge: r = .55, p < .001; objective knowledge: r = .31, p = .011 — both significant.

• The stronger self-report correlation likely reflects common-method bias (both self-report Likert scales) rather than a substantive difference — the objective test cross-validates the finding.^[raw/papers/mdpi-2026-critical-genai-use-predictors.md]

H2 — Critical-thinking disposition predicts critical use: SUPPORTED

• Multiple regression: F(2,64) = 17.19, p < .001, R² = .33 — a third of the variance in critical GenAI use.

• Epistemic orientation β = .36, p < .001; intellectual values β = .40, p < .001; low multicollinearity (VIF = 1.05), Durbin-Watson 2.23.^[raw/papers/mdpi-2026-critical-genai-use-predictors.md]

Exploratory — misconceptions are narrow but real

• Only 9% of participants had misconception scores below zero; just 3 of 12 items showed actual misconceptions (incorrect + high confidence).

• The standout: 97% confidently believed "GenAI performs web searches" (misconception score −3.00) — though the authors note current chatbots technically trigger web searches, so this item may be superseded by evolving tool capabilities.

• Most students correctly and confidently knew GenAI "may provide content that is not based on facts" (score 2.91, highest).^[raw/papers/mdpi-2026-critical-genai-use-predictors.md]

Exploratory — need for cognition (NFC)

• Actual genAI use correlated only with interest (r = .54) and perceived usefulness (r = .58) — not with knowledge or dispositions.

• Perceived usefulness negatively correlated with NFC (r = −.27, p = .027): high-NFC students view GenAI utility more sceptically.

• NFC positively correlated with critical use (r = .25, p = .044): the motivational drive to invest cognitive effort bridges AI literacy and actual critical behaviour.^[raw/papers/mdpi-2026-critical-genai-use-predictors.md]

Implications

• Institutions should move beyond basic tool training: target conceptual misconceptions (e.g. via refutation texts), foster epistemic orientation and intellectual values (video-based interventions shown effective in prior Hefter work), and consider NFC when designing interventions (low-NFC learners may need extra incentives like feedback/interactivity).

• Knowledge about human learning (metacognitive understanding of cognitive offloading costs) is proposed as an additional "meta-knowledge" predictor worth studying.

• Limitations: small psychology-only sample (88% female), self-report critical-use scale, correlational design (no causality), and the risk that static knowledge-test items go stale as tools evolve — future work should use performance-based measures like the GLAT and objective behaviour logs, and experimental/longitudinal designs.^[raw/papers/mdpi-2026-critical-genai-use-predictors.md]

Related Pages

• ai-literacy — Knowledge + dispositions as the two protective factors

• over-reliance — The target risk: uncritical overreliance on GenAI outputs

• chatgpt-critical-creative-thinking-review — Critical-thinking framing of GenAI use

• ai-literacy-assessment-misalignment — Self-reported vs performance-based literacy measurement (mirrored here)

• metacognition — Knowledge about human learning as meta-knowledge; NFC bridge

• student-experience — Undergraduate usage and validation behaviours

• higher-ed — Deployment context

• educational-theory — Kuhn's epistemic development and intellectual values framing

• hallucination-risk — Why cross-checking outputs matters

Citation

APA: Hefter, M. H., Paaßen, B., & Berthold, K. (2026). GenAI knowledge, epistemic orientation, and intellectual values predict undergraduate students' critical GenAI use. AI Educ., 2, 27. https://doi.org/10.3390/aieduc2030027