AI VS HUMAN: ACADEMIC ESSAY AUTHENTICITY CHALLENGE

ABSTRACT

The rise of artificial intelligence (AI) has transformed many industries, including the academic sector, where AI-generated essays pose a threat to academic integrity. This project, titled "AI VS HUMAN: ACADEMIC ESSAY AUTHENTICITY CHALLENGE," focuses on developing a system to differentiate between AI-generated and human-written academic essays. The core objective is to design a machine learning-based solution that accurately classifies text as either AI-generated or human-authored. The system leverages multiple machine learning algorithms, including Convolutional Neural Networks (CNN), Transformer models, and Random Forest Classifiers, to analyze and predict text origins using English text data. Additionally, the frontend is designed to accept user input in both English and Arabic. For Arabic input, the text is first translated into English before classification, enabling seamless detection of AI-generated content across multiple languages. The model is trained on a diverse dataset consisting of academic essays, including those written by both AI and humans. Through extensive training and evaluation, the model’s performance is assessed based on accuracy and precision in distinguishing between human-written and AI-generated texts. This tool can assist educators, researchers, and academic institutions in identifying non-original content and promoting academic honesty. The project ultimately aims to combat the growing challenge of AI in academia by providing an efficient and reliable solution for authenticity detection.

Keywords: AI detection, human-written, AI-generated, academic integrity, essay authenticity, machine learning, CNN, Transformer, Random Forest Classifier, text classification.