Malaysia’s banking and development finance sector is embracing AI faster than many institutions are ready to fully trust it, according to a new study from the Asian Institute of Chartered Bankers, Ecosystm, and the AICB Chief Risk Officers’ Forum. The “AICB-Ecosystm AI in Practice” report found that while AI is increasingly showing up in areas like KYC onboarding, fraud detection, anti-money laundering, and counter-terrorism financing checks, and general employee productivity, only a quarter of respondents trust AI-generated outputs enough to actually act on them for high-impact business decisions.
The study launched alongside AICB’s 4th Malaysian Banking Conference and 2nd Bank Audit Conference, drawing on responses from 87 senior leaders across commercial banks, digital banks, Islamic banks, and development financial institutions in Malaysia, supplemented by executive roundtables and interviews.
On readiness, the report found 44% of Malaysian banks and DFIs sit in a Developing stage, having moved past pure experimentation but still dealing with fragmented capabilities across data, skills, and operating models. Only 15% have reached an Established level, and just 2% qualify as Advanced, meaning AI is fully woven into decision-making and actively contributing to competitive advantage.
Specific gaps holding institutions back as well, as just 26% have a defined strategy tying AI to business goals, even as 44% are already building custom AI solutions, a combination that risks fragmented initiatives that are hard to scale or replicate. Specialized AI technical skills are in short supply at 79% of institutions, and only 20% actively promote AI-driven decision-making across their workforce, pointing to capability gaps that run throughout these organizations rather than sitting in any one department.
Meanwhile, about 53% of organizations are still relying on fragmented or ad hoc governance rather than consistent, risk-based frameworks for deciding controls, approvals, and oversight across different AI use cases. Only 33% have structured AI governance and model risk management in place, and just 27 percent apply formal AI risk tiering to match oversight to the level of risk involved.
AICB Chief Executive Edward Ling framed the findings as a shift in the underlying question facing the industry, saying Malaysia’s banks and DFIs have moved past debating whether AI belongs in financial services and are now grappling with whether they have the judgment, ethics, governance, and professional capability to use it responsibly in decisions touching customers, risk, and institutional performance.
Dr Chong Han Hwee, Chairman of the AICB Chief Risk Officers’ Forum and Group Chief Risk Officer at RHB Malaysia, noted that AI’s risks don’t live solely inside the model itself, but emerge across the whole ecosystem, spanning data quality, how people actually use the tools, the decisions AI informs, and how all of that shifts over time.
Sash Mukherjee, VP Industry Insights at Ecosystm, added that as AI moves into higher-risk use cases, financial institutions are looking for more clarity around model risk management, explainability, third-party AI, and data governance, and that regulation alone won’t keep pace with the technology, making ongoing collaboration between industry and regulators just as important as formal rules.
AICB said the findings offer a useful benchmark for the sector as institutions move from AI pilots toward responsible, enterprise-wide implementation, and reflect the Institute’s broader commitment to building industry capacity for the future of banking.
