Abstract: Plagiarism is a significant issue in computer science education, particularly in large courses where manual inspection is impractical. Students often use creative obfuscation methods to conceal their plagiarism. While manual obfuscation is tedious and requires some expertise, generative AI has made it easier than ever to obfuscate plagiarized solutions. Moreover, traditional detection methods, primarily designed for code, often fall short when applied to modeling artifacts. In this talk, we will explore the current landscape of software plagiarism detection, focusing on the intricacies of identifying plagiarism for modeling assignments. We will discuss both automated and manual obfuscation techniques, focusing on automation via generative AI. Despite these challenges, there are positive aspects of AI in education, and we will conclude by highlighting these aspects.

Outline:

1. Plagiarism in CS Education
2. Modelling Plagiarism
3. Automated Obfuscation and Generative AI
4. Using AI in Education

Links:

• JPlag Demo

• JPlag on GitHub

• JPlag on Helmholtz RSD

I am a researcher at Karlsruhe Institute of Technology (KIT). My research is focused on software plagiarism and collusion detection, in particular obfuscation attacks on software plagiarism detectors. I am the project lead of JPlag, which is used by many universities around the world.