Building responsible AI chatbot platforms in higher education: an evidence-based framework from design to implementation
Generative AI presents opportunities and challenges for higher education stakeholders. While most campuses are encouraging the use of generative AI, frameworks for responsible integration and evidence-based implementation are still emerging. This Curriculum, Instruction, and Pedagogy article offers...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2025-07-01
|
| Series: | Frontiers in Education |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/feduc.2025.1604934/full |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Generative AI presents opportunities and challenges for higher education stakeholders. While most campuses are encouraging the use of generative AI, frameworks for responsible integration and evidence-based implementation are still emerging. This Curriculum, Instruction, and Pedagogy article offers a use case of UT Austin’s approach to this dilemma through an innovative generative AI teaching and learning chatbot platform called UT Sage. Based on the demonstrated benefits of chatbot technologies in education, we developed UT Sage as a generative AI platform that is both student- and faculty-facing. The platform has two distinct features, one a tutorbot interface for students and the other, an instructional design agent or builder bot designed to coach faculty to create custom tutors using the science of learning. We believe UT Sage offers a first-of-its-kind generative AI tool that supports responsible use and drives active, student-centered learning and evidence-based instructional design at scale. Our findings include an overview of early lessons learned and future implications derived from the development and pilot testing of a campus-wide tutorbot platform at a major research university. We provide a comprehensive report on a single pedagogical innovation rather than an empirical study on generative AI. Our findings are limited by the constraints of autoethnographic approaches (all authors were involved in the project) and user-testing research. The practical implications of this work include two frameworks, derived from autoethnographic analysis, that we used to guide the responsible and pedagogically efficacious implementation of generative AI tutorbots in higher education. |
|---|---|
| ISSN: | 2504-284X |