Professor Steffen Herbold
'In AI engineering, one of my interests is the quality of AI models in times of autonomous driving and ChatGPT. While I was studying for my master's degree in applied computer science, back when AI and machine learning had yet to become major topics of interest, I attended a seminar that combined software engineering with artificial intelligence. Ever since, my research has been revolving around one question: How can I develop AI and machine learning from a software perspective and use them for software development?
While progress in autonomous driving is somewhat slowing, there are indications that language models will have disruptive effects before the decade ends. It may have taken long for the systems to produce more or less acceptable text quality, but the quality we now have works fairly well for lots of application scenarios.
An example: In a project carried out in collaboration with computer linguistics Professor Annette Hautli-Janisz and computer science educator Ute Heuer, we asked teachers to rate two essays generated by ChatGPT using the official evaluation matrix of the Standing Conference of the Ministers of Education and Cultural Affairs. This project demonstrated rather conclusively that pupils are unable to keep up with the language model when writing "for-and-against" essays – neither in their language skills nor in the line of arguments they provide. This proves that we can't just turn a blind eye on technology. I find it terribly exciting to explore how things will pan out going forward.'
More on his research
In his research, Professor Herbold explores the development of applications using AI methods. His focus is on the development of models that can be used for data analysis with machine learning, especially the software-based implementation for productive use of AI methods like those used to predict harvest yields on the basis of satellite data.
