Enhancing AI Text Detection: Embracing Bi-LSTM and Transformer Block Fusion

This research paper introduced an innovative approach to enhance AI text detection by combining Bidirectional Long Short-Term Memory (Bi-LSTM) with Transformer block architectures. The goal was to accurately differentiate between AIgenerated and human-written text. The proposed model utilizes an input layer, embedding layer, Bidirectional LSTM, custom Transformer block, convolutional layer, global max pooling, dense layers, and a sigmoid output layer for binary classification. By leveraging the strengths of both architectures, the model aims to address challenges related to authenticity and misuse of AIgenerated text. Evaluation metrics for instance accuracy, precision, recall, and F1-score are used to measure model performance. Overall, this research contributes to advancing AI text detection capabilities, fostering trust, integrity, and safety in digital communication. The study not only contributes to the academic understanding of text detection but also provides practical tools for distinguishing between AI-generated and human-written text. By leveraging a sophisticated architecture combining Bi-LSTM and Transformer Block fusion, the proposed method achieves significant advancements in text classification tasks. Through rigorous evaluation and experimentation, this research offers insights into the effectiveness of the proposed approach in accurately identifying AI -generated text, thereby addressing a critical need in the field of text detection. This study lays the groundwork for future research and development in AI-driven text detection systems, offering a scalable solution for detecting AIgenerated content across various digital platforms.