My PhD project sits at the intersection of psychology, education science, and machine learning.
I am using large language models and other machine learning techniques to grade, and feedback student work; specifically, my aim is to develop new tools for teachers and students and validate them experimentally and in the classroom.

Recent studies indicate that large language models can award accurate grades (Golchin et al., 2024; Grévisse, 2024; Ishida et al., 2024; Floden, 2025) and provide useful feedback (Dai et al., 2023; Jia et al., 2024; Meyer et al., 2024) in some cases. However, it remains unclear if these findings generalize to all subjects and courses. Additionally, technical barriers and potential ethical implications make educators hesitant to rely on LLMs in practice.

I want to address this by creating and continuously improving tools that make it easy for educators to utilize LLM grading and feedback in their courses. As my first project I integrated the learning management system Canvas with an API for prompting large language models (code:
https://github.com/lukekorthals/ai-support) and validated it during a programming course. Over the course of my PhD, I want to progress from closed form assessment (e.g., programming assignments), to more open assessment (e.g., short open essay questions), and eventually investigate LLM capabilities to assess scientific writing (e.g., student theses). During my research I will develop and compare different methods and investigate their psychometric properties; thus, my PhD project aligns well with the goals of IOPS.