This content is only partially available in English.
This content is only partially available in English.

Teaching Award for AI Simulations

Pascal Kordt

Nicole Bruhn, CC BY-SA 4.0

In the 2025/26 funding round for the “Innovative Teaching” Award, a project by Pascal Kordt was among those recognized! Congratulations to all the recipients.

Prof. Dr. Pascal Kordt was awarded the “Innovative Teaching 2025/26” prize for his project “AI-Based Simulation of Interaction with a Client in Database Science.” The project aims to replace traditional exercises with a simulation of everyday professional life. We extend our warmest congratulations to Pascal!

Imagine this: A student is tasked with developing a database for a company. But instead of a clear set of instructions, there are only fragments of information scattered among several “colleagues,” each of whom knows only part of the requirements. And: Everyone has their own unique way of communicating. There’s the grumpy colleague with a wealth of knowledge—terse, impatient, and prone to withholding important details unless you ask very specific questions. Or the colleague who’d rather talk about her last vacation and doesn’t know any technical terms, but will need to use the result later. And then there’s the boss who never has time and suddenly asks, “Will it be ready tomorrow?”

These “colleagues” aren’t real people, but are simulated by Large Language Models (LLMs) that acquire their personalities through specific instructions (system prompts). They’re designed to react realistically, unpredictably, and sometimes even frustratingly—just like in real-life work situations.

The goal: Students not only learn how to model databases—they aren’t given a ready-made recipe to follow step by step—but are also challenged to derive requirements from the communication, think critically, and ask the right questions.

Development of the software for the project is scheduled to begin in early 2026. In addition to the Department of Engineering, the Department of Economics is also participating in the project as a partner, represented by Prof. Dr. Martin Huschens. The interactions will be recorded anonymously to facilitate a subsequent scientific study.

The teaching award is endowed with a total of €60,000, which will be divided among the six winning projects.