An Investigative Analysis on Generation of AI Text Using Deep Learning Models for Large Language Models
Saranyya Chimata, Ankamma Rao Bollimuntha, DevaKiran Devagiri, Sowmyasri Puligadda, Venkatrama Phani Kumar S, Venkata Krishna Kishore Kolli · 2024
In this study, we propose a method for predicting whether a given text is generated by AI or human authors using Long Short-Term Memory (LSTM) networks. The proposed methodology involves preprocessing the text data, including tokenization and padding, followed by training an LSTM-based binary classification model. We utilize a dataset containing essays labeled as either AI-generated or human-generated for model training and evaluation. Our experiments involve splitting the dataset into training and testing sets, training the LSTM model on the training data, and evaluating its performance on the testing data. We assess the model's performance using various evaluation metrics, including accuracy, precision, recall, and F1 score. The results demonstrate the effectiveness of the LSTM based approach in accurately predicting the origin of the text, with promising performance metrics indicating its potential for real-world applications in identifying AI-generated content.