What does BNM AI Feedback Statement say about AI in banking?

Published Nov 19, 2026 · Last reviewed Nov 19, 2026

Bank Negara Malaysia published its Feedback Statement on Artificial Intelligence in the Malaysian Financial Sector on 19 November 2026, answering 113 responses to its August 2025 discussion paper. Its central decision is that 'BNM is not introducing AI-specific regulation at this stage': existing requirements on governance, risk management, consumer protection and Shariah compliance continue to apply to AI-enabled activities, applied proportionately with stronger safeguards for higher-risk uses such as credit decisioning. It sets out emerging practices for consumer outcomes, agentic AI and Shariah governance, and from 2027 BNM will explore with industry an AI Responsible Innovation Forum and a testing pathway for AI solutions.

OFFICIAL TEXT: bnm.gov.my ↗ · FINAL · MALAYSIA (BNM / SC)

Checked 2026-11-19 · Malaysia (BNM / SC), bnm.gov.my
DocumentBNM AI Feedback Statement — Feedback Statement: Artificial Intelligence (AI) in the Malaysian Financial Sector
Issued byBank Negara Malaysia (BNM), Malaysia's central bank and the regulator of banks, insurers and takaful operators, development financial institutions and payment firms, with the Securities Commission Malaysia (SC) for capital market entities
TypeReport
StatusFinal
PublishedNov 19, 2026
Applies toFinancial service providers regulated by BNM, including banks, insurers and takaful operators, development financial institutions and payment service providers. Non-binding: BNM says its illustrative examples 'do not introduce new requirements, thresholds or approval arrangements', while the existing policy documents it lists continue to apply
Official sourcebnm.gov.my ↗
Use casesAI governance (general) · Model risk management · Generative & agentic AI · Credit scoring & underwriting · Fair lending & discrimination · Third-party & vendor AI · Fraud detection · Cybersecurity · Data & privacy

What are the key points of BNM AI Feedback Statement?

  • Respondents: 113, about 80% of them financial service providers (37 banks, 42 insurers and takaful operators, 12 payment firms, 22 others). Use cases most often reported as deployed, piloted or explored: operational efficiency and process automation (64% of respondents), customer service and personalisation (61%), fraud and AML (52%) and compliance (48%).
  • Risks most cited: data governance (privacy, leakage through prompts, data quality), adversarial attacks (prompt injection, synthetic identities used to bypass e-KYC controls), model risk and governance (accountability, validation, bias), AI-enabled fraud (deepfakes, AI phishing), third-party dependency and concentration on a few foundation-model providers, and workforce impact.
  • Existing rules apply: 'AI-enabled activities presenting risks similar to those posed by conventional activities remain subject to the same applicable regulatory requirements.' Policy documents named as a starting point include RMiT, Corporate Governance, Risk Governance, Responsibility Mapping, Fair Treatment of Financial Consumers, Outsourcing, Capital Adequacy Framework, Management of Customer Information and Permitted Disclosures, the Data Management and MIS Framework and Complaints Handling.
  • Proportionality: stronger safeguards for higher-risk AI use cases, 'including proportionate testing, monitoring and human oversight, supported by a documented rationale and periodic reassessment'. Illustrations: department-level approval may suffice for meeting-summary tools, while credit decisioning may warrant more senior approval, and models that materially influence customer outcomes or financial decisions, such as credit underwriting, may warrant more comprehensive testing and independent validation.
  • Supervision: in the near term BNM will deepen its understanding of AI adoption, governance practices and key third-party dependencies and build multidisciplinary capability within existing supervisory structures; these insights may inform further supervisory engagement, 'potentially including thematic reviews'.
  • Industry frameworks: providers may consider the seven principles of the AI Governance Framework issued in 2025 by the Chief Risk Officers' Forum of the Asian Institute of Chartered Bankers and endorsed by the Association of Banks in Malaysia, and may consider assessing their practices against the forum's AI Model Risk Management Framework once it is issued.
  • Agentic AI practices: map the decisions and actions each agent may take with clear institutional accountability, calibrate access rights and transaction limits to risk tolerance, set thresholds for human approval and escalation, monitor for privilege escalation and 'autonomy drift', give customers avenues for review and redress, and build suitability and affordability safeguards into agentic AI used for financing decisions.
  • Consumer and Shariah practices: boards set expectations for responsible AI use, providers keep enough understanding to 'explain, challenge and correct AI-driven decisions', and clear boundaries prohibit use cases where material risks of discrimination, unfair treatment or consumer harm cannot be adequately mitigated; in Islamic institutions AI supports rather than replaces human Shariah decision-making.

What did BNM AI Feedback Statement change for banks?

The statement closes BNM's 2025 AI consultation without new rules. BNM kept its technology-neutral approach and instead mapped AI onto existing policy documents, illustrated proportionate controls and set out emerging practices for consumer outcomes, agentic AI and Shariah governance. From 2027 BNM will explore with industry an AI Responsible Innovation Forum, whose first priority area is cyber resilience, a Testing Pathway for AI Solutions and improved access to cross-institution data and synthetic data.

Is Bank Negara Malaysia introducing AI regulation?

Not at this stage. The feedback statement of 19 November 2026 says BNM is not introducing AI-specific regulation and will support responsible AI adoption through existing regulatory expectations, proportionate supervision and ongoing engagement. Existing requirements on governance, risk management, consumer protection and Shariah compliance continue to apply to AI-enabled activities, and BNM will progressively provide greater regulatory clarity as use cases and risks evolve.

What does BNM expect from banks using agentic AI?

The statement lists emerging practices rather than rules: clear accountability and mapped boundaries for each agent's actions, access rights and transaction limits calibrated to risk tolerance, human approval and escalation thresholds, ongoing monitoring for privilege escalation and autonomy drift, customer review and redress, and controlled experimentation before expanding an agent's scope or autonomy.

What is BNM's AI Responsible Innovation Forum?

ARIF is a structured forum that BNM will explore with industry from 2027 onwards, convening financial service providers, fintechs, technology providers, researchers and public-sector partners in thematic chapters. Its first priority area is strengthening cyber resilience against AI-enabled threats such as phishing, deepfakes and synthetic identity fraud, including possible joint red-teaming and controlled testing.

DateDocumentStatus
Jul 8, 2026BNM Governor AICB Nexus 2026 speech — Governor's Opening Address at the Inaugural AICB Nexus 2026: Advancing Banking, Governance and Financial Innovation in Malaysia and AsiaFinal
Nov 28, 2025BNM Risk Management in Technology (RMiT) — Risk Management in Technology (RMiT) policy documentIn force
Aug 5, 2025BNM AI Discussion Paper — Discussion Paper on Artificial Intelligence in the Malaysian Financial SectorFinal
Apr 15, 2024BNM e-KYC Policy Document — Electronic Know-Your-Customer (e-KYC) policy documentIn force
Mar 27, 2024BNM Fair Treatment of Financial Consumers — Fair Treatment of Financial Consumers policy documentIn force
Aug 1, 2023SC Guidelines on Technology Risk Management — Guidelines on Technology Risk Management (Securities Commission Malaysia)In force

Follow every document these regulators publish

when one of these regulators moves, the next morning's brief says so · six sourced stories · 7 am PT · free

plus every regulation, bank and agent page update, the morning after · leave any morning