Design and Implementation Plagiarism Checker Application with DetectGPT using Scheduler Algorithm

Abstract

The increasing use of Generative AI in the field of writing in the academic world creates problems in the field of plagiarism. this problem has prompted an urgent need for plagiarism detection tools. However, not all universities are able to implement a system that is able to detect sentences or paragraphs resulting from generative AI. Recent research seeks to overcome this obstacle using AI itself, specifically through the DetectGPT model. However, the implementation of this technology has limitations, such as requiring large computing resources and quite a long examination time. Using a GPU can speed up the process, but not all users can implement it. Solutions to minimize processing time include using more than one worker, however, scheduling is essential to maximize plagiarism detection efficiency. This research proposes the implementation of the above system with the help of the scheduler. The results obtained in this system prototype are an average waiting time for checking of 457,778 seconds to complete 10 tasks with the help of 3 workers running at the same time.