Distinguishing Human-Generated and Machine-Generated Text Using Deep Learning

Abstract

The advent of machine learning and natural language processing has led to significant advancements in text generation, which differentiate human-written content from machine-generated content. Differentiating between these two sources has become crucial in various domains, including journalism, academia, cybersecurity, and content moderation. This research delves into developing a deep learning model that accurately distinguishes between human and machine-generated texts and addresses the challenges posed by the proliferation of automated text generation systems. A dataset comprising 100,000 labelled instances has been used to train/test and predict. The model achieves a commendable accuracy to outperform the state-of-art.