Acceptance of Teacher-Authored AI Chatbots in Education: Comparing Student and Teacher Perspectives
Type
conference paper
Date Issued
2026-03-17
Author(s)
Abstract
Theoretical background:
With the growing availability of generative AI, interest in its didactic potential for classroom teaching has increased (Dwivedi et al., 2023). Acceptance is a key precondition for the sustainable integration of AI chatbots – both from the perspective of students and teachers (Acosta-Enriquez et al., 2024; Dwivedi et al., 2019; Polyportis & Pahos, 2025; Venkatesh et al., 2003). While students mainly emphasize usefulness and motivational aspects, teachers tend to focus more strongly on questions of pedagogical integration and related challenges. Empirical comparisons of these perspectives remain scarce – particularly with regard to teacher-authored chatbots (Yoo & Kim, 2025), i.e., AI systems whose dialogues and scenarios are specifically designed by teachers for their classrooms.
Research questions:
Against this background, the present study examines the acceptance of teacher-authored chatbots from both student and teacher perspectives. The study addresses the following research questions:
1. How do students evaluate the acceptance and perceived usefulness of teacher-authored chatbots?
2. What attitudes and experiences do teachers report regarding the pedagogical integration and use of such chatbots?
3. To what extent do the perspectives of students and teachers differ?
Methods:
To answer these questions, we investigated the acceptance of teacher-authored chatbots implemented in the learning app Brian (Brian, 2025) at the upper secondary level. The study is based on a pre-post survey of N = 1455 students from 20 different schools across 10 cantons (high school and vocational education) conducted between February and June 2025, complemented by an online survey of N = 27 teachers. Students and teachers used the app for at least six weeks, during which we conducted two surveys: a pre-test at the beginning and a post-test at the end of the trial. We assessed attitudes toward usefulness, motivation, performance improvement, practicality, and integration, using scales adapted from UTAUT models (Acosta-Enriquez et al., 2024; Dwivedi et al., 2019; Venkatesh et al., 2003). In addition, open-ended questions captured individual experiences. Data analysis included latent mean comparisons as well as non-parametric tests (Mann-Whitney U, Wilcoxon) to examine pre-post changes and subgroup differences.
Results:
From the students’ perspective, acceptance of Brian was moderate and independent of gender or school type. The highest agreement was found for its use in school-based learning processes (M = 3.53, SD = 1.13, Max = 5, N = 938) and as a support tool for learning activities (M = 3.34, SD = 1.13). However, initial enthusiasm declined after several weeks of use (the latent mean declined by 0.577) – possibly reflecting a novelty effect (Deng et al., 2025; Fryer et al., 2017). Overall, Brian seemed to have only a small effect on grades (V = 40801, p = .014, r = .10, N = 564). A differentiated analysis suggested that lower-performing students appeared to benefit more (V = 1422.5, p < .001, r = .78), while higher-performing students showed little additional improvement.
Teachers emphasized pedagogical opportunities such as differentiated feedback, flexible support, scaffolding, and the promotion of self-regulated learning. At the same time, they reported challenges related to designing appropriate tasks, the quality and depth of chatbot responses, time investment, and trust in the technology.
While both groups expressed generally positive attitudes, our results would seem to suggest that sustained acceptance requires purposeful pedagogical integration and support to stabilize expectations and highlight long-term benefits. This study contributes to the empirical understanding of teacher-authored chatbots and highlights the need to fully exploit their potential. For educational research, this points to the importance of developing integrative models that address both students’ motivational needs and teachers’ professional requirements.
With the growing availability of generative AI, interest in its didactic potential for classroom teaching has increased (Dwivedi et al., 2023). Acceptance is a key precondition for the sustainable integration of AI chatbots – both from the perspective of students and teachers (Acosta-Enriquez et al., 2024; Dwivedi et al., 2019; Polyportis & Pahos, 2025; Venkatesh et al., 2003). While students mainly emphasize usefulness and motivational aspects, teachers tend to focus more strongly on questions of pedagogical integration and related challenges. Empirical comparisons of these perspectives remain scarce – particularly with regard to teacher-authored chatbots (Yoo & Kim, 2025), i.e., AI systems whose dialogues and scenarios are specifically designed by teachers for their classrooms.
Research questions:
Against this background, the present study examines the acceptance of teacher-authored chatbots from both student and teacher perspectives. The study addresses the following research questions:
1. How do students evaluate the acceptance and perceived usefulness of teacher-authored chatbots?
2. What attitudes and experiences do teachers report regarding the pedagogical integration and use of such chatbots?
3. To what extent do the perspectives of students and teachers differ?
Methods:
To answer these questions, we investigated the acceptance of teacher-authored chatbots implemented in the learning app Brian (Brian, 2025) at the upper secondary level. The study is based on a pre-post survey of N = 1455 students from 20 different schools across 10 cantons (high school and vocational education) conducted between February and June 2025, complemented by an online survey of N = 27 teachers. Students and teachers used the app for at least six weeks, during which we conducted two surveys: a pre-test at the beginning and a post-test at the end of the trial. We assessed attitudes toward usefulness, motivation, performance improvement, practicality, and integration, using scales adapted from UTAUT models (Acosta-Enriquez et al., 2024; Dwivedi et al., 2019; Venkatesh et al., 2003). In addition, open-ended questions captured individual experiences. Data analysis included latent mean comparisons as well as non-parametric tests (Mann-Whitney U, Wilcoxon) to examine pre-post changes and subgroup differences.
Results:
From the students’ perspective, acceptance of Brian was moderate and independent of gender or school type. The highest agreement was found for its use in school-based learning processes (M = 3.53, SD = 1.13, Max = 5, N = 938) and as a support tool for learning activities (M = 3.34, SD = 1.13). However, initial enthusiasm declined after several weeks of use (the latent mean declined by 0.577) – possibly reflecting a novelty effect (Deng et al., 2025; Fryer et al., 2017). Overall, Brian seemed to have only a small effect on grades (V = 40801, p = .014, r = .10, N = 564). A differentiated analysis suggested that lower-performing students appeared to benefit more (V = 1422.5, p < .001, r = .78), while higher-performing students showed little additional improvement.
Teachers emphasized pedagogical opportunities such as differentiated feedback, flexible support, scaffolding, and the promotion of self-regulated learning. At the same time, they reported challenges related to designing appropriate tasks, the quality and depth of chatbot responses, time investment, and trust in the technology.
While both groups expressed generally positive attitudes, our results would seem to suggest that sustained acceptance requires purposeful pedagogical integration and support to stabilize expectations and highlight long-term benefits. This study contributes to the empirical understanding of teacher-authored chatbots and highlights the need to fully exploit their potential. For educational research, this points to the importance of developing integrative models that address both students’ motivational needs and teachers’ professional requirements.
Language
English (United States)
Keywords
AI chatbots
technology acceptance
teacher-authored chatbots
AI-supported learning
artificial intelligence
HSG Classification
contribution to scientific community
Refereed
No
Event Title
13. GEBF-Kongress
Event Location
Technische Universität München
Event Date
16.-18. März 2026
Subject(s)
Division(s)
Contact Email Address
mandana.york@unisg.ch