What type of study is this?

This is a Quantitative Study study.

October 9, 2025Open Access

Findings of the BEA 2025 Shared Task on Pedagogical Ability Assessment of AI-powered Tutors

View Full Paper

Ask AI

Bookmark

Authors

Shared task evaluates pedagogical ability of AI tutors using large language models, indicating areas for improvement.

Performance evaluation of AI tutors showed significant variability in pedagogical abilities, highlighting areas for development.
The best F1 scores were 71.81 for mistake identification and 58.34 for providing guidance across three-class problems.
Fifty international teams participated, utilizing diverse approaches to evaluate mistake remediation in educational contexts.
Resources and data from this task will support ongoing research into improving AI tutoring systems.

Kochmar et al. (2025) studied this question.

View Full Paper

Ask AI

Bookmark