The only practitioner-grade index of RBI, SEBI, IRDAI, DPDP Rules 2025 and the proposed Digital India Act — applied to AI/ML use-cases in Indian banking, capital markets and insurance. Curated by Hemant Sahay, 25+ years JAPAC banking technology.
RBI's guidance on the use of artificial intelligence and machine learning in financial services (FREE-AI committee report, December 2024) sets expectations across model governance, explainability, bias testing, customer redress and operational resilience. Banks deploying AI for credit decisioning, fraud detection or customer servicing must demonstrate human-in-the-loop oversight, periodic re-validation, dataset lineage, and bias-impact reports. Expect mandatory model-risk frameworks aligned to RBI MD-ITGRC 2023 plus the upcoming AI Master Direction.
SEBI's December 2024 circular on algorithmic trading for retail investors mandates broker certification, kill-switch controls, audit trails for every algo order, and a maker-checker workflow for algo deployment. AI-driven trading and robo-advisory must comply with the Investment Adviser Regulations + the upcoming SEBI AI Use-Case Inventory rule. Capital-market participants must classify their AI use across high/low risk tiers and notify SEBI for high-risk deployments.
IRDAI's 2025 sandbox guidelines invite insurers to test AI applications in underwriting, claims triage and customer servicing under a controlled regulatory environment. Insurers must maintain a Model Risk Register, document training data sources, ensure DPDP-compliant consent for sensitive personal data inputs (health, biometric), and submit periodic algorithmic-fairness reports. Combine with IRDAI's 2026 Cybersecurity Guidelines for end-to-end ISMS coverage.
The DPDP Rules 2025 (notified 13 Nov 2025) treat AI-driven processing of personal data as a high-risk activity. Data Fiduciaries deploying AI must: obtain granular consent for AI-based profiling, provide opt-out from automated decisions, conduct DPIAs every 12 months, log every model prediction touching personal data for 3 years, and notify the Data Protection Board of any automated-decision dispute involving over ₹50,000 monetary impact. Significant Data Fiduciaries face additional algorithmic-audit obligations.
The proposed Digital India Act (DIA) — successor to the IT Act 2000 — is expected to carve out a dedicated AI chapter covering high-risk AI use-cases (credit scoring, biometric ID, employment screening, healthcare diagnostics), with mandatory pre-deployment risk assessment, transparency obligations and grievance redress. BFSI players should map current AI deployments against the EU AI Act's risk taxonomy now — DIA's classification will likely converge.
5 questions · 60 seconds. Find out how your BFSI organisation scores against DPDP Rules 2025 + RBI/SEBI/IRDAI AI-readiness baselines.
Made with Emergent