Mathematics Teacher Interaction Styles with Generative Artificial Intelligence: Personalization in Word- Problem Design
Year: 2026
Type: preprint
Abstract: As generative artificial intelligence (GenAI) plays an increasing role in K-12 mathematics classrooms, understanding how teachers collaborate with these technologies during instructional design is essential. Drawing on frameworks related to personalization theory, Technological Pedagogical Content Knowledge, and adaptive epistemic agency, this study examined how seven seventh-grade mathematics teachers interacted with ChatGPT to personalize mathematics word problems to students’ interests. Analyses of teacher-GenAI interactions identified three interaction styles: AI-Dominated-Passive Prompt Personalization, Teacher-Dominated-Rigid Personalization, and Human-AI partnership-Collaborative Personalization. Students' ratings of the resulting personalized problems varied across teachers and interaction styles, with the strongest positive patterns occurring among teachers characterized by Collaborative Personalization. Collaborative personalization illustrates how teachers may maintain epistemic authority for evaluating mathematical, pedagogical, and contextual appropriateness while selectively granting GenAI generative authority to expand personalization possibilities. These findings highlight the importance of preparing teachers to critically evaluate and collaboratively engage with GenAI during instructional design.
Source: Research Square
Authors: Theodora Beauchamp(Formal analysis, Methodology, Writing – original draft, Writing – review & editing), Candace Walkington(Conceptualization, Funding acquisition, Investigation, Methodology, Supervision, Writing – review & editing), Tiffini Pruitt-Britton(Writing – original draft, Writing – review & editing)
Language: English
Subfield: Education
Field: Social Sciences
Domain: Social Sciences
SDG: Quality education
Open Access status: green
Funder: National Science Foundation
OpenAlex ID: W7235833554
