Expert plan preferences were split, with some experts preferring the lower-scoring general-purpose plan for its social-emotional emphasis¶

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? Seven experts preferred C2, four preferred C1, and one expressed no preference, showing expert judgment incorporates affective and contextual dimensions beyond structural alignment. → Karaismailoglu 2026

Evidence¶

Karaismailoglu 2026¶

Karaismailoglu, F., Surmeli, H., & Yildirim, M. (2026). Suitability of Artificial Intelligence Supported Lesson Plans from the Perspective of Science Education Experts. Journal of Science Education and Technology. https://doi.org/10.1007/s10956-026-10342-4

q2 · i? · design · r2

Analysis of individual expert total scores and stated preferences (Table 4) among eleven experts evaluating both plans. The printed split is 7 for C2, 4 for C1, 1 none; one expert favored C1 for "more emphasis on social-emotional skills."

"the preference split in Table 4 (7 experts preferring C2, 4 preferring C1, and 1 expressing no preference) indicates that although C2 received higher average scores in Table 2, expert opinions were not unanimous."

Discussion¶

Related Claims¶

• The general-purpose platform scored higher in early EDBL stages while the education-focused platform scored higher in feedback-intensive iterative stages — related
• The education-focused AI platform (C2) received numerically higher expert ratings than the general-purpose platform (C1) across all eight overall quality criteria — related
• Most experts judged the AI-generated plans conditionally applicable, endorsing use only after correction rather than direct classroom use
• Automated metrics and LLM judges can misalign with expert human preferences in non-verifiable educational evaluation — a broader claim this one bears on