Digital innovation in teacher training in Saxony: Five years of learning with chatbots

For five years now – and thus long before the hype surrounding ChatGPT and similar technologies began – universities in Saxony have been using artificial intelligence (AI) to introduce innovative digital approaches to university teaching with great success. The joint project tech4comp, and later tech4compKI, first introduced chatbots as virtual mentors for learning support during the 2019/2020 winter semester.

Scientists at TU Dresden, Leipzig University of Applied Sciences, and Chemnitz University of Technology have since been continuously developing these tools – coordinated by our colleagues at Leipzig University – alongside partners from the German Research Center for Artificial Intelligence (DFKI), Freie Universität Berlin, Martin Luther University Halle-Wittenberg , and RWTH Aachen. The objective of the joint project, which is funded by the Federal Ministry of Education and Research (BMBF), is to develop scalable and efficient, digitally-supported mentoring technologies to aid students of education, teaching, mathematics, and computer science, especially during independent study phases.

When it was introduced five years ago, this use of AI technology in university teaching in general, and in a pedagogical discipline in particular, was unique in Germany. Over 7,500 future teachers have since received an introduction to learning with AI mentors. In addition, as of the 2023/2024 academic year with the introduction of new state teacher examination regulations, TU Dresden has made this a mandatory study component for all students studying to become teachers.

This technology has already demonstrated its worth in Prof. Thomas Köhler's courses at TU Dresden’s Chair of Educational Technology: “The chatbot works as a virtual mentor that students can interact with whenever they see fit. Students can ask the chatbot questions at any time, or send it texts written by themselves to receive automatic feedback. In recent years, we have seen that working with the chatbot helps to individualize, and thus improve, the learning process,” emphasizes Thomas Köhler.

Group coordinator and project manager Prof. Heinz-Werner Wollersheim of Leipzig University highlights just how innovative the project is – especially for students studying to become teachers, who often view innovations in media technology with skepticism: “At least since the launch of ChatGPT, teachers need to take a closer look at the potential and challenges of generative AI. We offer students the opportunity to gain initial experience in this area, for example with a chatbot that taps into the potential of large language models (LLM) and that is specially adapted to our module content, unlike ChatGPT. This is important because students, quite rightly, have high expectations of the quality of AI services.”

The Mentoring Workbench, a web interface and central component of the project, connects the chatbot with other tools such as a search engine and a timetable. It is also seamlessly embedded into OPAL, the joint Learning Management System (LMS) of the universities in Saxony, and the LMS Moodle. This enables students to use the tool within a learning environment they are familiar with. As OPAL is already used by more than 80 percent of all students in Saxony, this underlines the remarkable potential of this new technology.

The project managers aim to continue developing the chatbot in the future. To this end, they are aiming to introduce the technology to all universities in Saxony with a view to fundamentally improving digital education in the long term. At the same time, new projects are emerging that are advancing the development and optimization of AI technologies in university teaching.

Since 2024, the Junior Research Group “Situating AI-based Mentoring,” headed by Dr. Sandra Hummel at TU Dresden, has been investigating how AI-supported mentoring can be customized more precisely and therefore adapted more effectively to students' individual learning needs. The interdisciplinary group places great importance on transferring findings from research directly into AI development. Their goal is to create personalized learning support that adapts to learners' progress and needs in real time.

In these times of teacher shortages and lesson cancellations, Thomas Köhler sees great potential in technology-enhanced learning. “With our partners, we will continue to investigate organizational and technological requirements, and to improve the potential applications of our learning technologies. With this as a basis, we will also be able to make a profit from implementing these mentoring tools outside of the university, for example at schools or for independent learning, especially in the short term.”

More information on the project: https://tu-dresden.de/codip/projekte/projektoverview/tech4comp

About CODIP TU Dresden

The Center for Open Digital Innovation and Participation (CODIP) researches how digital media and tools can be used for teaching, learning and research, and how their use changes the world of work and everyday life. With more than 70 employees, it conducts interdisciplinary work into digital media and tools while developing suitable concepts, educational programs, and digital applications. It focuses on research and development topics both on its own initiative and in alliances with partners from TU Dresden, the entire Federal Republic of Germany and 15 countries worldwide.

Early-Career Research Group “Situated Mentoring”

As part of ScaDS.AI (Center for Data Science, Artificial Intelligence and Big Data with locations in Dresden and Leipzig), the Junior Research Group combines education research with AI development to create systems that are not only technically efficient, but also pedagogically meaningful and effective in university teaching. This marks a new approach in AI development, where the focus is on learning and the technology is tailored to support the individual needs of students in the long term.

Contact:

Prof. Werner Wollersheim
Group Coordinator and Project Manager Leipzig University
Tel.: +49 341 97-31411
Mail: 

Prof. Thomas Köhler
Project Manager TU Dresden
Tel.: +49 351 463-32772
Mail: 

Dr. Sandra Hummel
Head of the Situated Mentoring Early-Career Research Group
ScaDS.AI Dresden/Leipzig
Tel.: +49 351 463-39913
Mail: