#09 - Course Description Chatbot for Teachers at TU Graz

Improving your own course descriptions using generative AI

In this article, we would like to present the course description chatbot at Graz University of Technology, explain why a good course description is useful and how you can use AI to easily improve your own course descriptions with a chatbot based on structured feedback.

The application interface is available in German for the time being.

Necessity for detailed course descriptions

The course description in TUGRAZonline is the most important publicly available information about your course, which students and other interested parties use to find information.

The obligation to publicly disclose objectives, content, methods, assessment criteria and assessment standards is enshrined in the Universities Act (Section 76 (2) UG).

Students need this information in order to learn about the requirements and the amount of work to be expected, but also, for example, to plan stays abroad and have their credits transferred.

The course description chatbot for teachers at TU Graz

The course description chatbot provides teachers at TU Graz with an application that allows them to check and easily improve their own course descriptions for TUGRAZonline during the course registration period.

You only have access to your own courses in the course description chatbot. After signing in, you can choose one of your courses.

Recommendations for good course descriptions can be found at TU4U.

In the first step, you will receive AI-generated feedback that uses a traffic light system. It is based on the existing course description texts and the recommendations for the respective section.

LV description chatbot: a description is shown in the left-hand column. In the right-hand column, there is feedback with a traffic light system. On the left is a menu that also shows the traffic light colours.

Fig. 1: Example for AI-based feedback using the traffic light system. The application is only available in German for now.

In the second step, you can start a chat for each section and optimise the description with the help of AI. A few selected sample prompts are available to give you ideas for revision.

For example, you can create a draft for the respective section containing placeholders, which you can then fill with your own information from the course.

Chat window in the course description chatbot. Below the chat window, there is an input field and suggested example prompts.

Fig. 2: Chatbot for AI-based editing of the course description. The application is only available in German for now.

You can repeatedly request feedback on all versions created with the help of AI, and in the end save your preferred version. Saved versions can also be translated with the help of AI.

With one click, the texts are copied to the clipboard and manually transferred to TUGRAZonline.

Saved responses in three columns: Column 1 shows the original text, column 2 shows the text from the chatbot, and column 3 shows the translation.

Fig. 3: Saved texts and translations. The application is only available in German for now.

If you use the AI-based chatbot for revisions, please do not transmit any personal data and observe copyright regulations and the basic principles of AI at Graz University of Technology.

Further information on preparing a course with the help of AI can be found in the Didactics section in the article #09 Use of AI-based tools in teaching (Part 1)

Prompting strategy for structured feedback

The course description chatbot generates feedback based on a predefined system prompt for the respective section. We want to make transparent how the course description chatbot works.

All predefined prompts that the course description chatbot uses for feedback can therefore be viewed in the ‘System prompts’ menu.

You are also welcome to reuse these prompts in your own contexts or other AI applications.