EUVIMEDEUVIMED entdecken
Crossref

Agentic Generative AI as a Dialogic Scaffold for Problem-Posing Pedagogy

Agnes M. Taveros, Joedel Penaranda, Marianne Agnes Mendoza

International Journal of Learning, Teaching and Educational Research · 30.07.2026 · 25 (7) · 1063-1084

DOI: 10.26803/ijlter.25.7.48 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Learning, Teaching and Educational Research, online erschienen am 30.07.2026. Verfasst von Agnes M. Taveros und 2 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Dialogic problem posing requires learners to generate, justify, critique, and revise curriculum-aligned problems, yet these routines are difficult to sustain when an instructor must support several groups simultaneously. This study examined whether an answer-constrained agentic generative artificial intelligence (AI) workspace could function as a dialogic scaffold without displacing student agency. An explanatory sequential mixed-methods quasi-experimental design involved 96 students enrolled in four intact undergraduate course sections during an eight-week intervention. Two sections used an instructor-managed GPT-class AI workspace composed of planning, critique, and dialogue-coaching agents, while two comparison sections completed the same routines through instructor-guided printed prompts. Quantitative data included pretest and posttest problem-posing quality scores, peer dialogic engagement ratings from transcribed collaborative work, and substantive revision-cycle counts. Qualitative data came from focus groups, draft histories, final artefacts, and an instructor interview. Controlling for pretest performance and prior course achievement, the AI group outperformed the comparison group in problem-posing quality (partial eta squared =.13) and dialogic engagement (partial eta squared =.17) and completed more substantive revisions (d = 1.81). Qualitative findings indicated that the workspace broadened planning, surfaced coherence problems earlier, and prompted more equitable peer participation. These findings suggest that generative AI can support inquiry-oriented learning when it is bounded, answer-constrained, and positioned as a scaffold for critique and dialogue rather than a source of finished answers.

Zitieren

APA

Taveros, A. M., Penaranda, J., & Mendoza, M. A. (2026). Agentic Generative AI as a Dialogic Scaffold for Problem-Posing Pedagogy. International Journal of Learning, Teaching and Educational Research, 25(7), 1063-1084. https://doi.org/10.26803/ijlter.25.7.48

Vancouver

Taveros AM, Penaranda J, Mendoza MA. Agentic Generative AI as a Dialogic Scaffold for Problem-Posing Pedagogy. International Journal of Learning, Teaching and Educational Research. 2026;25(7):1063-1084. doi:10.26803/ijlter.25.7.48

BibTeX
@article{Taveros2026,
  author = {Taveros, Agnes M. and Penaranda, Joedel and Mendoza, Marianne Agnes},
  title = {Agentic Generative AI as a Dialogic Scaffold for Problem-Posing Pedagogy},
  journal = {International Journal of Learning, Teaching and Educational Research},
  year = {2026},
  volume = {25},
  number = {7},
  pages = {1063--1084},
  doi = {10.26803/ijlter.25.7.48}
}
Kostenlos starten →