Survey on AI-Generated Plagiarism Detection: The Impact of Large Language Models on Academic Integrity
Shushanta Pudasaini, Luis Miralles-Pechuán, David Lillis and Marisa Llorens Salvador
Journal of Academic Ethics, Nov. 2024.
Abstract
A survey conducted in 2023 surveyed 3,017 high school and college students. It found that almost one-third of them confessed to using ChatGPT for assistance with their homework. The rise of Large Language Models (LLMs) such as ChatGPT and Gemini has led to a surge in academic misconduct. Students can now complete their assignments and exams just by asking an LLM for solutions to the given problem, without putting in the effort required for learning. And, what is more worrying, educators do not have the proper tools to detect it. The more advanced AI tools become, the more human-like text they generate, and the more difficult they are to detect. Additionally, some educators find it difficult to adapt their teaching and assessment methods to avoid plagiarism. This paper is focused on how LLMs and AI-Generated Content (AIGC) have affected education. It first shows the relationship between LLMs and academic dishonesty. Then, it reviews state-of-the-art solutions for preventing academic plagiarism in detail, including a survey of the main datasets, algorithms, tools, and evasion strategies for plagiarism detection. Lastly, it identifies gaps in existing solutions and presents potential long-term solutions based on AI tools and educational approaches to address plagiarism in an ever-changing world.