A Trustworthy Peer Assessment Model Integrating ChatGPT to Improve Credibility and Learning Outcomes (Poster 35)

• In Event: AERA Poster Session Sunday 1:30 p.m.
  In Poster Session: Artificial Intelligence in Education and Learning Technologies

Abstract

Peer assessment is a valuable educational tool, but credibility issues can undermine its effectiveness. This study proposes a peer assessment model incorporating ChatGPT to enhance credibility and improve student outcomes. Involving 27 third-year Educational Technology students from a Shanghai university, the study conducted two rounds of peer assessments in an "Instructional Design in Technological Contexts" course. ChatGPT generated credibility scores for two peer evaluations, which were presented to students for revisions of instructional design. Pre/post-test surveys measured changes in instructional design abilities. Results showed significant improvements in credibility scores and students' competencies, indicating that integrating AI tools in peer assessment enhances both the trustworthiness and educational impact of the evaluations.

Authors

• Zihao Wu, East China Normal University

• Jing Leng, East China Normal University

• Honghuan Lu, East China Normal University