EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Generative AI for feedback and collaborative knowledge construction in preservice physics teacher education

Yuchen Jiang, Xiu-Mei Feng, Yuting Liu, Yiting Wang, Li Xie, Lei Bao

Physical Review Physics Education Research · 06.02.2026 · 22 (1)

DOI: 10.1103/hm13-jv98 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Physical Review Physics Education Research, online erschienen am 06.02.2026. Verfasst von Yuchen Jiang und 5 weiteren Autor:innen. Die Arbeit wurde bisher 4-mal zitiert (Crossref) und nennt 61 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Central University Basic Research Fund of China, Central China Normal University.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Generative artificial intelligence (GenAI) offers strong potential to transform teaching practices; however, empirical research on its classroom effectiveness remains limited. This study explores how GenAI can be integrated into preservice physics teacher training by replacing two traditional instructor roles in an instructional design course: providing evaluative feedback and facilitating collaboration. A GenAI-based assistant was developed to assess instructional designs according to predefined criteria and to provide personalized feedback. A total of 110 preservice physics teachers received feedback from one of the three sources: a teaching expert, a teaching assistant, or GenAI to revise their instructional designs. They subsequently collaborated with GenAI for further refinement. Results indicate that GenAI evaluations closely aligned with those of human experts and teaching assistants. In blinded evaluations, preservice physics teachers rated GenAI feedback as comparable to human feedback. Furthermore, collaboration with GenAI resulted in significant improvements in the overall quality of instructional design. This progress was particularly most evident in the structured, content-expandable components. While these findings demonstrate GenAI’s feasibility from a functional perspective, challenges related to preservice teachers’ perceptions also emerged. Although design quality improved through collaboration with GenAI, independent design skills, measured by a final paper-and-pencil test, showed no similar improvement. Preservice physics teachers acknowledged the need to filter and revise GenAI-generated content, but often directly adopted its suggestions due to the convenience of GenAI. Notably, feedback was rated lower once participants realized it was AI-generated, even if the actual source was not, revealing trust-related concerns. This study moves beyond technical evaluations of GenAI, offering insights into the pedagogical challenges and limitations that arise when GenAI assumes the roles that are traditionally performed by human educators.

Förderung

  • Central University Basic Research Fund of China
  • Central China Normal University

Zitieren

APA

Jiang, Y., Feng, X., Liu, Y., Wang, Y., Xie, L., & Bao, L. (2026). Generative AI for feedback and collaborative knowledge construction in preservice physics teacher education. Physical Review Physics Education Research, 22(1). https://doi.org/10.1103/hm13-jv98

Vancouver

Jiang Y, Feng X, Liu Y, Wang Y, Xie L, Bao L. Generative AI for feedback and collaborative knowledge construction in preservice physics teacher education. Physical Review Physics Education Research. 2026;22(1). doi:10.1103/hm13-jv98

BibTeX
@article{Jiang2026,
  author = {Jiang, Yuchen and Feng, Xiu-Mei and Liu, Yuting and Wang, Yiting and Xie, Li and Bao, Lei},
  title = {Generative AI for feedback and collaborative knowledge construction in preservice physics teacher education},
  journal = {Physical Review Physics Education Research},
  year = {2026},
  volume = {22},
  number = {1},
  doi = {10.1103/hm13-jv98}
}
Kostenlos starten →