.svg)
In August 2025, Bank Negara Malaysia (BNM) launched a public consultation to explore how artificial intelligence (AI) can be used responsibly across the financial sector. The discussion focuses on how Malaysia can embrace innovation while maintaining strong governance and consumer protection.
This is an important moment for the financial services industry. AI has the power to transform how financial institutions detect risk, serve customers, and combat financial crime. But to harness that power responsibly, we need shared standards, clear definitions, and a collaborative approach between regulators, financial institutions, and technology providers.
The need for harmonised definitions of AI
One of the key questions in the BNM Consultation was on the need for a formal sector-specific regulatory definition of AI. A clear, sector-specific definition of AI would bring much-needed consistency to Malaysia’s financial sector. Without one, institutions risk applying inconsistent or overly narrow interpretations, which can create confusion and loopholes.
Too often, many organisations label simple automation or rules-based tools as AI. While automation improves efficiency, it does not meet the threshold of AI systems that learn, adapt, and make data-driven decisions. Distinguishing between true AI and automation is essential for ensuring that both regulators and institutions apply the right standards of oversight, accountability, and transparency.
The rise of agentic AI ,systems composed of autonomous agents capable of interacting and adapting with minimal human input, makes this even more urgent. These technologies promise huge gains in efficiency but also raise new risks around explainability and accountability.
Greater uniformity of AI definitions across jurisdictions is essential. For global financial institutions, inconsistent definitions of AI create operational inefficiencies and compliance challenges. For example, the UK government defines agentic AI as ‘Artificial intelligence (AI) agents are small, specialised pieces of software that can make decisions and operate cooperatively or independently to achieve system objectives. Agentic AI refers to AI systems composed of agents that can behave, interact and adapt autonomously in order to achieve their objectives.’
Harmonised definitions, developed in consultation with regional and international regulators, would enable firms to adopt AI in a globally consistent way, benefiting regulatory alignment and industry-wide trust.
A clear regulatory definition, rooted in compliance and governance principles, would ensure that AI is deployed responsibly and that regulators can differentiate between tools that require closer oversight and those that do not. AI used in customer service, for example, poses different challenges than AI applied in anti-money laundering (AML) transaction monitoring or credit scoring.
The state of AI in Malaysia’s financial sector
AI adoption in Malaysia is gaining momentum, especially in financial crime compliance. According to the Napier AI / AML Index 2025 - 2026, Malaysia lost 5.04% of its GDP to money laundering in 2024, despite spending $1.95 billion on compliance – a cost much higher than what similar sized economies have spent last year. By adopting AI-powered compliance systems, the country could save up to $0.56 billion USD proving a clear sign of the potential impact of responsible AI adoption.
In financial crime compliance, AI enables institutions to detect suspicious activity more effectively, reduce false positives, and adapt faster to new money laundering and terrorism financing typologies. This improves both the efficiency and accuracy of AML processes, freeing compliance teams to focus on genuinely high-risk activity rather than being overwhelmed by alerts.
At the same time, the sector must manage challenges around explainability, fairness, and human oversight. AI should always enhance, not replace, human judgment especially where accountability and consumer outcomes are at stake. Maintaining transparency and auditability in AI systems is essential to building long-term trust.
Malaysia’s technology-neutral regulatory framework provides a strong foundation for AI innovation, but targeted guidance on issues like bias, explainability, and governance would further strengthen responsible adoption.
Preparing for the future of AI: what next for the region?
To prepare for AI-driven transformation, financial institutions should focus on three key areas: governance, accountability, and collaboration.
- Establish strong oversight structures, including board-level accountability and ethics committees.
- Implement model risk management covering explainability, validation, retraining, and drift detection.
- Strengthen data governance ensuring data integrity, quality, and fairness.
- Maintain human-in-the-loop decisioning to keep accountability where it belongs.
AI is a rapidly changing field, with new techniques and models pushing boundaries being released multiple times each year. BNM’s engagement with the sector should happen at least annually, with flexibility for more frequent consultations as AI technology evolves. Engagement does need to be meaningful, ensuring that concerns are heard and any recommendations are released quickly to support financial service providers, so that they can use the tools responsibly.
Feedback loops between regulators and financial institutions will help ensure that policy guidance keeps pace with innovation. Collaboration is key. Shared testing environments, joint pilots, and industry working groups would allow institutions and technology providers to develop safe, explainable, and transparent AI models together. This approach helps everyone understand what ‘good’ looks like creating confidence, consistency, and fairness.
Napier AI’s collaboration with the UK Financial Conduct Authority (FCA) offers a good example. Through the Supercharged Sandbox and the use of synthetic data for money laundering detection, we are testing how new AI tools can be developed safely under regulatory supervision. Adopting a similar model in Malaysia would create a balanced environment for innovation and consumer protection.
AI regulation offers a ‘win-win-win' opportunity in Malaysia. Regulators gain better oversight, institutions gain operational efficiency, and consumers gain greater trust and inclusion. Realising this potential depends on a shared commitment to governance, transparency, and collaboration.
By establishing harmonised definitions, building meaningful engagement, and embedding compliance-first principles into every stage of AI development, Malaysia can lead the region in responsible AI adoption.
.jpg)
Read the Napier AI / AML Index to learn more about regulation in Malaysia.
Photo by Esmonde Yong on Unsplash
