India has witnessed dramatic growth in international online communication via avenues like WhatsApp and video calling services and its effect to interpersonal communication has been widely influenced by it. But this evolution has brought its own new vectors for cybercrime, in the form honeytrap scams and digital arrest frauds. These offenses prey on human weaknesses—particularly fear, romance and shame—to dupe people and steal either money or sensitive information. A systematic study of these scams is provided in this paper including behavioral aspects, case reports, and social engineering tactics exploited by these scams. Last, we present an AI and ML derived detection framework that can pinpoint high-risk conversations and associated user profiles by performing text analytics, image classification, and behavior modeling. Through public data, simulated datasets and ML classifiers, we show that early detection is possible, exposing bottlenecks that may be addressing by authorities and social media platforms. The paper ends with suggestions for enhancing digital literacy, regulation enforcement, and platform responsibility.