Zaylani urges AI oversight and clearer limits on audit assurance
Central Bank of Malaysia Deputy Governor Adnan Zaylani Mohamad Zahid called for stronger AI governance and risk-focused internal audit. He spoke at the 11th ASEAN central bank internal audit network meeting in Kuala Lumpur on October 8, 2026.
Five risks to institutional trust
Zaylani identified five areas affecting central bank credibility: cross-border risks, persistent uncertainty, data governance, artificial intelligence and operational resilience.
Cyberattacks, digital assets and interconnected financial systems require cooperation across jurisdictions.
Decisions increasingly rely on incomplete information, while confidence in policy, supervision and market operations depends on reliable data.
AI can improve analysis and risk detection, but requires clear accountability, human oversight and explainable decisions.
Resilience means anticipating, withstanding, recovering from and adapting to cyber incidents, operational disruptions and market stress while maintaining critical functions.
Assurance has boundaries
Internal audit must retain independent assurance over controls, governance and compliance while helping boards identify emerging vulnerabilities, Zaylani said.
It cannot anticipate every event, examine every risk or guarantee outcomes.
Boards, management and auditors need agreement on assurance scope and limits.
The network welcomed Papua New Guinea representatives for the first time.
Its programme covered AI auditing, cyber threats and recovery, cybersecurity readiness, data management and the Global Internal Audit Standards.
Technology cannot replace judgment
Zaylani placed institutional judgment, rather than technology alone, at the centre of effective assurance.
He said independence, impartiality, professional scepticism and integrity cannot be automated.
Audit credibility therefore rests on both better tools and explicit limits.