Profiles differ in self-perceived academic skills: Engaged and Literate users report the highest problem-solving and cognitive endurance gains, while Uncritical Reliant users report the highest independent thinking¶

Claim · All claims
Evidence · 3 studies · 3 associational r2 · q2 · i2 medium
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q2 i2 Clusters differed in problem solving and task execution, F(2, 820) = 35.00, η2 = 0.08, with Cluster 1 highest (M = 3.57). → Amzalag M 2026 q2 i2 Clusters differed in cognitive endurance, F(2, 820) = 21.40, η2 = 0.05, with Cluster 1 highest (M = 3.36). → Amzalag M 2026 (2) q2 i2 A reversal emerged for independent thinking, F(2, 810) = 16.00, η2 = 0.04, with Cluster 2 highest rather than Cluster 1. → Amzalag M 2026 (3)

Evidence¶

Amzalag M 2026¶

Amzalag M, Zviel-Girshin R and Beimel D (2026) The illusion of cognitive independence in the age of generative AI: unpacking GenAI literacy and learner profiles. Front. Educ. 11:1861548. https://doi.org/10.3389/feduc.2026.1861548

q2 · i2 · associational · r2

One-way ANOVA across the three learner profiles on self-perceived academic skills. The article reports Cluster 1 "reported significantly higher perceived enhancement in problem-solving skills" with η2 = 0.08; outcomes are self-reported perceptions.

"For problem solving and task execution [F(2, 820) = 35.00, p < .001, η2 = 0.08], Cluster 1 (M = 3.57) reported significantly higher perceived enhancement in problem-solving skills compared to Cluster 2 (M = 3.26) and Cluster 3 (M = 3.29)."

Amzalag M 2026 (2)¶

Amzalag M, Zviel-Girshin R and Beimel D (2026) The illusion of cognitive independence in the age of generative AI: unpacking GenAI literacy and learner profiles. Front. Educ. 11:1861548. https://doi.org/10.3389/feduc.2026.1861548

q2 · i2 · associational · r2

One-way ANOVA on perceived cognitive endurance across clusters, with Cluster 1 significantly outperforming Clusters 2 and 3 (η2 = 0.05). The measure is self-perceived skill change, not an objective performance test.

"Cluster 1 (M = 3.36) demonstrated the highest levels of cognitive endurance [F(2, 820) = 21.40, p < .001, η2= 0.05], significantly outperforming Clusters 2 and 3."

Amzalag M 2026 (3)¶

Amzalag M, Zviel-Girshin R and Beimel D (2026) The illusion of cognitive independence in the age of generative AI: unpacking GenAI literacy and learner profiles. Front. Educ. 11:1861548. https://doi.org/10.3389/feduc.2026.1861548

q2 · i2 · associational · r2

One-way ANOVA on perceived independent thinking across clusters; the article describes a striking reversal in which the Uncritical Reliant profile, not the Engaged and Literate profile, scored highest (η2 = 0.04). The supplied text is truncated at this point, so the Cluster 2 mean is not available.

"However, a striking reversal was observed for independent thinking [F(2, 810) = 16.00, p < .001, η2 = 0.04]: Cluster 2"

Discussion¶

Learner Variables¶

• Problem Solving — predictor: learners who differ on it differ in outcomes

Related Claims¶

• Cluster analysis identifies three stable learner profiles: Engaged and Literate, Uncritical Reliant, and Skeptical Minimal users — related
• Profiles differ in immediate usage behaviors: Engaged and Literate users verify most, while Skeptical Minimal users report the highest shortcut reliance — related
• Latent profile analysis of pre-service teachers' CT skills identifies three profiles (Novice, Developing, Proficient), revealing heterogeneous, non-linear skill acquisition — related
• High-achieving students less often endorse AI-supported autonomy, effective learning, and active engagement, and more often agree that reliance on AI hinders critical thinking and independent problem solving — related
• The six clusters differ substantially in academic performance: Final Grade ANOVA η² = 0.770, with C0 exceeding C1 by 40.17 exam-score points — related
• Higher GenAI usage clusters report greater perceived difficulty, lower self-assessed understanding, and different trust in AI-generated code — related
• Higher academic self-efficacy profiles report greater overall academic AI use, with a large effect (η² = .36) — related
• Confused guessers had significantly lower pre- and post-test skill scores than most other clusters, but their science learning gains were not significantly different from those of Clusters 1, 2, and 4. — related
• Problem-solving-prepared students often accepted the AI solution uncritically or critiqued it based on their own misconceptions — related