RBI Guidelines on Artificial Intelligence & Machine Learning in Banking & Credit Risk
Statutory Overview
The Reserve Bank of India has issued comprehensive regulatory directions governing automated algorithmic credit scoring, loan disbursement, fraud detection, and customer verification models deployed by scheduled commercial banks and NBFCs.
Key Compliance Requirements for AI Companies
Explainability Mandate: AI models used in credit scoring or loan rejection must provide auditable, human-interpretable reasons.
Bias Auditing: Regulated entities must conduct bi-annual algorithmic bias audits to ensure models do not discriminate on demographic factors.
Human-in-the-Loop: Critical lending decisions exceeding defined thresholds cannot be fully automated without credit officer review.
Data Localization: All financial training data and inference logs must reside exclusively within RBI-approved Indian data centers.
Legislative Timeline & Milestones
FinTech Innovation Hub Report
RBI Innovation Hub submits recommendations on responsible AI credit scoring.
Circular Released
Mandatory compliance timeline established for all commercial banks and NBFCs.
Official Gazette & Regulatory References
Frequently Asked Questions & Legal Interpretations
Q:Can a bank reject a loan solely based on an AI model output?
Under RBI guidelines, the applicant has a right to know the key factors contributing to an adverse decision, and an escalation mechanism for human review must be provided.