Preservice educators rate generative AI highest for logistical support (time saving, clarity, organizing thought) and lowest for epistemic trust, accuracy and consistency¶

Claim · All claims
Evidence · 1 study · design r2 · q2
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q2 i? Among 131 preservice educators, trust in the LLM received the lowest mean rating (M = 3.08) while time saving received the highest (M = 3.95) on 5-point Likert items. → George Efthimiou 2026

Evidence¶

George Efthimiou 2026¶

George Efthimiou, Katerina Plakitsi. (2026). Preservice Early Childhood Educators' Engagement with Generative AI in Science Activity Design: The Case of Magnetism. https://www.mdpi.com/journal/educationsciences doi:10.35542/osf.io/m5erf_v1

q2 · i? · design · r2

Descriptive analysis of post-interaction Likert reflection items from 131 preservice educators after a 90-minute AI-supported magnetism activity design workshop. The lowest-rated items were trust, accuracy and consistency; the highest-rated were logistical, with time saving at "M = 3.95, SD = 0.87".

"Conversely, trust in the LLM yielded the lowest mean score in the dataset ( M = 3.08, SD = 0.82), alongside perceived accuracy (M = 3.18, SD = 0.83) and consistency (M = 3.26, SD = 0.78)."

Discussion¶

Related Claims¶

• Students rated ChatGPT highly for completing activities more quickly, among the highest-scoring survey items in both courses — related
• Epistemic trust in AI-generated content did not differ significantly across the five AI platforms — related
• Reviewed studies consistently raise concerns about AI accuracy, overreliance, academic integrity, and reduced critical engagement in Arab EL2 contexts — related
• Students rated the flipped-learning component highest and the GenAI module lower, with GenAI rated stronger for ideation than for time saving or direct output use — reports the opposite
• Practice with feedback is more efficient than lecture, saving about 24% of instructional time in Study 1 and yielding 2–3 times higher learning rates in Study 2 — related
• Assignment support was the highest-rated Class Companion function (83% Effective/Very Effective, M = 3.38), while assessing knowledge and skills was the lowest (71%) — related
• LLM-generated retrieval practice was practical to deploy and well received, but high correctness rates (about 70% of items at p=1.0) restricted score variance — related
• Educators did not perceive the iterative CTML review process as slowing them down, rating production efficiency at 4.26/5 — related
• Survey items on critical evaluation and validation of ChatGPT outputs (Q11–Q13) received the lowest scores in both courses despite explicit validation instructions — a narrower finding that bears on this claim