AI Ed Wiki logoAI Ed Wiki
Use with AI

📄 Research Article

Talking mathematics with AI: Understanding teachers' motivation for utilizing chatbots

● high · 2026-08-16 · math-education, chatgpt, chatbot, teacher-role, motivation, technology-acceptance

Synthesis: Weinhandl and colleagues (2026) examine what motivates Austrian secondary Math Education teachers to use AI chatbots for teaching, within the Unified Theory of Acceptance and Use of Technology (UTAUT) framework. A digital questionnaire study yielded usable data from 448 of the 625 teachers who started (retaining those who answered ≥75% of items overall and ≥2/3 per construct), measured on five-point Likert scales via a modified, German-translated UTAUT instrument (validated by forward–backward translation, expert and teacher review, piloting with 25 teachers, confirmatory factor analysis, and McDonald's ω). Structural equation modelling in lavaan 0.6.17, comparing seven candidate models by chi-squared difference test, selected the original UTAUT predictors plus Perceived Risk (PR). Model fit was acceptable (RMSEA = 0.071, SRMR = 0.067, CFI = 0.913, TLI = 0.892). Direct effects showed Performance Expectancy (PE), Effort Expectancy (EE), and Social Influence (SI) all significantly predicted Behavioral Intention (β = 0.473, 0.287, 0.398; all p < 0.001), while Behavioral Intention (β = 0.687) and Facilitating Conditions (β = 0.297) predicted actual Use Behavior; Perceived Risk was non-significant for both. Age and teaching experience did not moderate the model, but gender partially moderated the BI→UB path. The authors conclude that demonstrable instructional gains (PE) matter most, and that pragmatic factors outweigh perceived risk for this group.

Key Findings

UTAUT drivers confirmed. Performance Expectancy, Effort Expectancy, and Social Influence all significantly and positively shaped Behavioral Intention to use chatbots, with PE the strongest predictor (β = 0.473, p < 0.001), ahead of SI (β = 0.398) and EE (β = 0.287). Behavioral Intention (β = 0.687) and Facilitating Conditions (β = 0.297) predicted actual Use Behavior.

Perceived Risk non-significant. Contrary to earlier literature, PR had no significant effect on Behavioral Intention (β = 0.059, p = .235) or Use Behavior (β = −0.076, p = .136), suggesting evolving teacher perceptions as chatbot familiarity increases.

Mediation via intention. Indirect effects of PE (β = 0.325), EE (β = 0.197), and SI (β = 0.273) on Use Behavior via Behavioral Intention were significant (total indirect β = 0.835, p < 0.001), underscoring intention formation as the bridge from perceptions to classroom adoption.

Gender as partial moderator. Age and teaching experience did not moderate model relationships, but gender significantly moderated the BI→UB path (β = −0.304, p = .012), indicating intention translates into practice differently for male and female teachers.

Domain-specific acceptance. The study extends UTAUT technology-acceptance research to AI chatbots specifically in Math Education, where conversational AI for problem-solving is novel — a context previously dominated by technical or single-case studies, with little quantitative European evidence.

Large validated sample. Data came from 448 teachers across Austria (258 female, 163 male, 2 nonbinary; most aged 51–60, 27%), with robust measurement (CFA fit, McDonald's ω, no multicollinearity, VIF < 2).

Implication for Teacher Education. Because pragmatic drivers — perceived instructional benefit, ease of use, and peer/social norms — dominate adoption, Teacher Education and support programs should foreground evidence of instructional impact and low-friction onboarding rather than risk-mitigation messaging.

Connected Concepts

Connected Articles

Citation

Weinhandl, R., Andic, B., Wijaya, T. T., Bleckenwegner, V., Riegler, V., Baldinger, S., Mayrhofer, J., & Helm, C. (2026). Talking mathematics with AI: Understanding teachers' motivation for utilizing chatbots. Computers and Education Open, 100359. https://doi.org/10.1016/j.caeo.2026.100359