India’s rapid growth in digital payments has been accompanied by an alarming surge in financial fraud, pushing regulators and financial institutions to increasingly rely on artificial intelligence (AI) to secure the ecosystem. With over 18,000 crore digital transactions recorded in 2024-25 and UPI volumes soaring by 137 per cent to ₹200 trillion in 2023-24, the shift to cashless payments has been transformative but fraught with risks.
Between April 2024 and January 2025 alone, the country reported 24 lakh digital fraud cases, leading to losses of ₹4,245 crore, a 67 percent spike from the previous year. High-value frauds exceeding ₹1 lakh have also multiplied, with 29,082 such cases resulting in losses of around ₹175 crore.
Structural Gaps and Sophisticated Fraud Tactics
Analysts note that the rise in fraud is being driven by both user-side vulnerabilities and systemic lapses. A large segment of users remains unaware of fake payment links, phishing attempts, and fraudulent apps. At the same time, compliance gaps in merchant verification, onboarding norms, and enforcement of regulatory standards are creating exploitable weak points.
Cybercriminals are leveraging advanced tools such as AI-generated content, deepfakes, and manipulative digital platforms to mislead unsuspecting users. These evolving tactics have made it increasingly difficult to rely solely on traditional fraud detection mechanisms.
AI-Powered Solutions Gain Ground
To address these challenges, regulators and institutions are deploying AI and machine learning (ML) systems for proactive monitoring. The Reserve Bank of India has rolled out MuleHunter.AI, a model designed to identify and deactivate mule accounts used in large-scale scams. The National Payments Corporation of India (NPCI), meanwhile, is piloting a federated AI framework with leading banks to strengthen real-time fraud detection across the network.
Global models are also being adopted. Mastercard’s decision intelligence platform, which analyses over 16,000 crore transactions annually, assigns instant risk scores to block unauthorised payments. AI-driven systems are now increasingly equipped to detect anomalies, analyse behavioural patterns, and automate incident responses through security orchestration, automation, and response (SOAR) mechanisms.
Experts caution, however, that AI is not without its challenges. False positives, adversarial manipulation of open AI models, and data privacy concerns remain pressing hurdles. Going forward, experts suggest integrating AI into zero-trust architectures and fostering multi-stakeholder collaboration between regulators, banks, and technology providers to ensure resilience against evolving digital threats.