Regaining Control: Enabling Educators to Build Specialized AI Chat Bots with Retrieval Augmented Generation

Barbara Pampel, Simon Martin, Ulrike Padó

2025

Abstract

Conversational AI (chat) bots are powerful and helpful tools, but are not suited for the unrestricted use in many classrooms: They may hallucinate, easily veer from the topic of instruction, and are vulnerable to malicious prompting. Retrieval-augmented generation (RAG) is a technique that allows educators to constrain chat bots to a specific area of expertise, reducing hallucinations and vulnerability to mis-use. We are working on a low-code solution that enables tech-savvy educators to build such a RAG-based chat bot system themselves, thus retaining full control over the content and behavior of their bot. We present the first version of this system and promising initial feedback from educators and students on its suitability, reliability and flexibility.

Paper Citation

in Harvard Style

Pampel B., Martin S. and Padó U. (2025). Regaining Control: Enabling Educators to Build Specialized AI Chat Bots with Retrieval Augmented Generation. In Proceedings of the 17th International Conference on Computer Supported Education - Volume 2: CSEDU; ISBN 978-989-758-746-7, SciTePress, pages 371-378. DOI: 10.5220/0013425500003932

in Bibtex Style

@conference{csedu25,
author={Barbara Pampel and Simon Martin and Ulrike Padó},
title={Regaining Control: Enabling Educators to Build Specialized AI Chat Bots with Retrieval Augmented Generation},
booktitle={Proceedings of the 17th International Conference on Computer Supported Education - Volume 2: CSEDU},
year={2025},
pages={371-378},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013425500003932},
isbn={978-989-758-746-7},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 17th International Conference on Computer Supported Education - Volume 2: CSEDU
TI - Regaining Control: Enabling Educators to Build Specialized AI Chat Bots with Retrieval Augmented Generation
SN - 978-989-758-746-7
AU - Pampel B.
AU - Martin S.
AU - Padó U.
PY - 2025
SP - 371
EP - 378
DO - 10.5220/0013425500003932
PB - SciTePress