Economy & Banking★ Must Know

RBI proposes draft Model Risk Management framework to govern AI/ML use by banks and NBFCs

Wednesday, 24 June 2026 5 min readSource: ReutersArticle 1 of 14

The Reserve Bank of India (RBI) on June 24, 2026 released a draft 'Guidance on Regulatory Principles for Model Risk Management, 2026', its first comprehensive framework for governing models — including Artificial Intelligence (AI) and Machine Learning (ML) systems — used by banks, NBFCs and other regulated entities. The draft requires regulated entities to put in place board-approved model risk-management policies, maintain a model inventory, ensure independent validation of all models (including third-party ones), and adopt AI/ML-specific safeguards such as human oversight, 'kill-switch' mechanisms and bias testing. It follows a risk-tiering and lifecycle approach proportionate to each entity's scale and complexity, and places accountability on the regulated entity even for outsourced/third-party models. Stakeholder comments are invited until July 24, 2026.

Key Facts & Details

9 points

1
The RBI released the draft 'Guidance on Regulatory Principles for Model Risk Management, 2026' on June 24, 2026.
2
It governs all models including AI and ML systems used by banks, NBFCs and other regulated entities.
3
Entities must have board-approved policies, a model inventory and independent validation of every model.
4
AI/ML safeguards include human oversight, kill-switch mechanisms and bias testing.
5
Accountability rests with the regulated entity even for third-party/outsourced models.
6
It is a draft for public consultation, with comments open until July 24, 2026.

Deep Dive

+
Financial firms increasingly use models for credit scoring, fraud detection, pricing and customer service, raising risks from flawed or opaque AI systems.
+
'Model risk' is the potential for losses arising from errors in a model's design, data, assumptions or misuse.
+
The framework follows a use-case-based, risk-tiered approach, applying stricter controls to higher-impact models.

Exam Focus

Examiners may test the regulator (RBI), the draft's name (Guidance on Regulatory Principles for Model Risk Management, 2026), its scope (AI/ML and other models in banks/NBFCs), key safeguards (human oversight, kill-switch, bias testing) and the consultation deadline (July 24, 2026).

Exam Relevance & Angle

AI governance in finance is a fast-rising Banking & Financial Awareness and current-affairs theme; the RBI as issuer, the named Model Risk Management guidance, and its AI/ML safeguards (human oversight, kill-switch, bias testing) are topical, testable hooks across banking and RBI-grade exams.

Target Exams

SBI POIBPS POIBPS RRB OfficerRBI Grade BNABARD Grade ASSC CGLRRB NTPCLIC AAOUPSC CSEState PCS

Background & Context

The Reserve Bank of India, India's central bank and the regulator of banks and NBFCs, has been progressively building rules for the responsible use of technology in finance, including its earlier FREE-AI committee work and digital-lending norms. A 'model' in finance is any quantitative method that processes data to produce an output — a credit score, a fraud flag, a price — and model risk is the danger of loss from a faulty or misused model. As banks and NBFCs adopt AI and machine learning at scale, regulators worldwide are moving to ensure such systems remain explainable, fair, supervised and accountable. The RBI's draft mirrors this global shift toward formal AI governance, balancing innovation with consumer protection and financial stability.

Related GK Concepts

Must Know

Model risk managementArtificial Intelligence / Machine Learning (AI/ML)Board-approved governanceIndependent model validationReserve Bank of India (RBI)