Authors:
Amitabha Maheshwari, Praveen Aronkar

Addresses:
Department of Management, Shriram Institute of Information Technology, Banmore, Madhya Pradesh, India. Department of Management, Prestige Institute of Management and Research, Gwalior, Madhya Pradesh, India.

Abstract:

This research paper is in response to the use of Artificial Intelligence (AI) to detect financial fraud using text analysis on Generative Pre-trained Transformers (GPT). Spammers have continued to utilise advanced language models to generate copied content; consequently, more traditional anti-fraud methods are correspondingly less effective. This article proposes a novel approach that combines Natural Language Processing (NLP) and machine learning techniques to detect deception patterns in GPT-generated content. This is achieved by generating a new dataset of authentic and artificially created financial reports, including emails, reports, and social media posts. The training dataset is tested and validated using a collection of AI models, which includes a fine-tuned version of GPT-3.5, a Long Short-Term Memory (LSTM) network, and a Transformer-based classifier. Python is the primary tool used in this paper, with TensorFlow and PyTorch packages employed for model development, and scikit-learn utilised for performance analysis. The outcome demonstrates that the developed AI system can identify phishing text with extremely high accuracy, providing financial institutions with a reasonable opportunity to enhance their ability to combat fraud in the digital era. The research highlights the future of artificial intelligence in combating new forms of fraud and emphasises the need for ongoing innovation in this area.

Keywords: Artificial Intelligence; Financial Fraud; Generative Pre-trained Transformer; Machine Learning; Natural Language Processing; Long Short-Term Memory; Transformer-Based Classifier.

Received on: 07/01/2025, Revised on: 02/04/2025, Accepted on: 25/06/2025, Published on: 09/09/2025

DOI: 10.69888/FTSTPL.2025.000450

FMDB Transactions on Sustainable Technoprise Letters, 2025 Vol. 3 No. 3, Pages: 178-186

• Views : 44
• Downloads : 6

Download PDF