With the increasing prevalence of fraudulent phone calls, AI-powered scam detection systems are now crucial weapons in the fight against phone frauds. Distinct machine learning and deep learning techniques have been investigated in recent studies. Techniques to improve scam detection accuracy, including Large Language Models (LLMs), Support Vector Machines (SVM), Recurrent Neural Networks (RNN), and ensemble learning methods. However, these systems encounter difficulties like biased datasets, low recall, false positives, and evolving scam tactics. This survey provides an analytical review of existing literature and identifies the most effective methods for scam detection based on comparative analysis. The study emphasizes the significance of real-time AI adaptation, resource optimization, explainable AI, seamless integration, and continuous feedback loops to increase scam detection systems’ efficacy and dependability.
Paper
Full text
Breaking the Scam Cycle: AI’s Role in Fraudulent Call Detection
Semantic Scholar · 2025
Abstract
With the increasing prevalence of fraudulent phone calls, AI-powered scam detection systems are now crucial weapons in the fight against phone frauds. Distinct machine learning and deep learning techniques have been investigated in recent studies. Techniques to improve scam detection accuracy, including Large Language Models (LLMs), Support Vector Machines (SVM), Recurrent Neural Networks (RNN), and ensemble learning methods. However, these systems encounter difficulties like biased datasets, low recall, false positives, and evolving scam tactics. This survey provides an analytical review of existing literature and identifies the most effective methods for scam detection based on comparative analysis. The study emphasizes the significance of real-time AI adaptation, resource optimization, explainable AI, seamless integration, and continuous feedback loops to increase scam detection systems’ efficacy and dependability.