India’s digital payments landscape is booming like never before, but the rise has come with a steep cost. With over 18,000 crore transactions recorded in 2024-25, the Unified Payments Interface (UPI) alone surged by 137% to ₹200 trillion in FY 2023-24. Yet, this exponential growth has been accompanied by a surge in financial fraud.
Between April 2024 and January 2025, the country reported 24 lakh digital fraud cases, leading to losses of ₹4,245 crore, a 67% jump from the previous year. High-value cyber frauds, involving amounts above ₹1 lakh, also spiked, with 29,082 incidents causing ₹175 crore in losses.
Data Protection and DPDP Act Readiness: Hundreds of Senior Leaders Sign Up for CDPO Program
Why Fraud is Rising Despite Stronger Security
The rapid migration to mobile-based and UPI platforms has outpaced user awareness and digital literacy, leaving millions vulnerable to fake links, phishing attempts, and fraudulent apps. Criminals, too, have grown more sophisticated, leveraging AI-generated content, deepfakes, and advanced scams.
But fraud is not just a result of gullible users or smarter criminals. Experts point to systemic compliance lapses: weak enforcement of onboarding rules, poor merchant verification, and inconsistent application of regulatory protocols. These blind spots create fertile ground for large-scale fraud.
AI Steps In: From MuleHunter to Federated Models
To counter this, regulators and institutions are deploying AI-driven compliance systems.
- The Reserve Bank of India (RBI) has launched MuleHunter.AI to detect and eliminate mule accounts, frequently used in phishing scams and fraudulent transfers.
- The National Payments Corporation of India (NPCI) has started a federated AI pilot with banks to strengthen fraud detection and risk modelling.
- Mastercard’s intelligence platform already analyses 16,000 crore global transactions annually, generating real-time risk scores to block unauthorised activity.
These AI systems rely on anomaly detection, behavioural analysis, automated incident response (SOAR), and adaptive endpoint protection. Natural language processing helps spot phishing emails, while deep learning identifies persistent cyber threats.
Challenges: Privacy, False Positives & Adversarial AI
While promising, AI adoption isn’t without risks. Training AI requires massive datasets, raising privacy concerns, while limited data may lead to false positives, flagging legitimate users as suspicious. More alarmingly, adversarial AI, where attackers manipulate open AI systems & poses a future challenge.
Experts argue that banks must adopt AI-driven zero-trust architecture, where no user or system is automatically trusted. Instead, trust must be continuously verified. Alongside a multi-stakeholder approach, spanning regulators, financial institutions, and tech providers, it is crucial to ensure digital innovation is not undermined by vulnerabilities.
As India strengthens its AI-enabled defences, the battle is no longer just against fraudsters, but also against the very tools they exploit.