Abstract:
In an era marked by the ascendancy of artificial intelligence (AI) and natural language processing (NLP), the emergence of advanced language models, such as those produced by ChatGPT, has revolutionized human-machine interactions. However, alongside the proliferation of AI-generated text comes the critical challenge of distinguishing between machine-generated and human-authored content, prompting concerns across diverse domains. This research endeavors to address this challenge by proposing a novel AI-generated text detection system specifically focused on identifying ChatGPT-generated introductions from Wikipedia articles. Employing a structured systems analysis and design methodology (SSADM) and a three-tier architecture, the study systematically investigates the problem background, offers innovative solutions, and designs a robust architecture. Through a comprehensive methodology encompassing literature review, requirement elicitation, design, implementation, and evaluation, the research aims to bridge the accuracy gap in existing detection approaches, thus contributing to the advancement of AI-generated text detection systems and safeguarding stakeholders against potential misuse and abuse of machine-generated content.
