AI-Generated and Human-Written Text Detection Using DistilBERT
Joshua Mawudem Gakpetor, Martin Doe, Michael Yeboah-Sarpong Damoah, Dominic Dalyngton Damoah, John Kingsley Arthur, Michael Tetteh Asare · 2024
This research proposes a system for detecting AI-generated text, utilizing the DistilBERT model, a streamlined variant of the larger BERT architecture. The system is designed to address the growing challenge of differentiating between AI-generated and human-written text: classifying an essay or a paragraph as human-written or AI-generated. Text classification is the process of classifying documents into predefined categories based on their content The methodology encompasses comprehensive data preprocessing, tokenization and encoding, and strategic hyperparameter tuning. DistilBERT is fine-tuned on a curated dataset encompassing diverse text types to enhance its detection accuracy.