How RBI is using AI to tackle financial fraud
What's the story
The Reserve Bank of India (RBI) is taking major steps to use artificial intelligence (AI) tools for improving its operations.
The focus is mainly on fighting financial fraud and ensuring customer safety.
The move was highlighted in the RBI's latest annual report, released on May 29, where "AI" appeared nearly 20 times across different sections.
Diverse applications
AI initiatives span multiple areas
The annual report revealed that AI is being used in complaint redressal systems, fraud detection models, supervisory frameworks, and international policy submissions.
It is also being integrated into internal workflows through RBI's own generative AI tool called Chat Interface with Retrieval Augmented Generation (ChiRAG).
The report described the potential of emerging technologies like generative AI as "rapidly gaining traction in the central banking landscape."
Tool development
ChiRAG's evolution and external engagement
Initially designed as a tool for information extraction and synthesis, ChiRAG is now being developed into "a sophisticated orchestration layer."
This will seamlessly coordinate with diverse types of information and data associated with RBI's wide array of functions.
Externally, RBI is working closely with its subsidiary, Reserve Bank Information Technology Pvt. Ltd (ReBIT), to integrate AI into its complaint management system in phases.
Fraud prevention
AI's role in fraud detection and prevention
Digital payment frauds are a major concern for the RBI. To combat this, the bank has formed a committee to look into setting up a Digital Payments Intelligence Platform (DPIP).
The platform will use advanced technologies to protect customers from digital payment fraud.
The Reserve Bank Innovation Hub (RBIH), another RBI subsidiary, is building a prototype of DPIP in consultation with five to 10 banks.
AI innovation
RBIH's tool for identifying mule accounts
RBIH has also created a supervised machine learning (ML) tool, MuleHunter.ai, to deal with mule bank accounts.
The model uses advanced AI/ML techniques to learn patterns of mule account activity from data and achieve higher accuracy than traditional systems.
It is being tested and deployed in a few large public sector banks for near-real-time identification of such accounts.
Supervision enhancement
RBI's AI-driven supervisory models and ethical framework
The RBI has also set up an Advanced Supervisory Analytics Group (ASAG) to drive the use of AI/ML techniques in its supervisory process.
The group has already developed several analytics models, including microdata analytics, governance assessment, social media monitoring, fraud vulnerability index, borrowers' vulnerability model, and asset quality prediction model.
In December 2024, the RBI formed an external expert committee to develop a Framework for Responsible and Ethical Enablement of AI (FREE-AI).