Publication Date
Fall 2024
Degree Type
Master's Project
Degree Name
Master of Science in Computer Science (MSCS)
Department
Computer Science
First Advisor
Faranak Abri
Second Advisor
Nada Attar
Third Advisor
Pranathi Kunadi
Keywords
AI Generated Text Detection, Machine Learning, Deep Learning, LLM’s, Embeddings, LSTM, BiLSTM, BERT, RoBERTa
Abstract
The detection of AI-generated text by the application of advanced machine learning techniques not only presents a promising approach toward distinguishing human-written content from machine-generated text, but also identifies the source model used for the generation of the text. This helps address the growing concerns about authenticity and accountability in digital communication. The difference between human-generated and AI-generated text lies in the core of several applications, from news media to academic integrity.It also helps in ensuring the transparency and trust in content-driven environments. However, existing models fall short in accurately detecting AI-generated text and identifying the specific AI source due to the complex nature of AI-generated text. To address this, it is essential to leverage advanced machine learning models and embedding techniques that can capture subtle linguistic and contextual patterns of the AI generated text. In this study, text classification was performed to develop classification models that distinguish AI-generated content from human-written text and further identify the specific AI model used, offering a multilayered approach to detection. The results reveal significant improvements in the detection accuracy and source identification, supporting applications in content moderation.
Recommended Citation
Tatavarthi, Anjana Priyatham, "AI Generated Text Detection & Source Identification" (2024). Master's Projects. 1454.
https://scholarworks.sjsu.edu/etd_projects/1454
