AI Governance in Malaysian Finance: Human Accountability is Key

The rapid integration of artificial intelligence across Malaysia’s financial sector presents a profound structural challenge: how to reconcile the efficiency gains of AI with the immutable need for human accountability. Bank Negara Malaysia (BNM) Governor Abdul Rasheed Ghaffour recently articulated this imperative, stressing that financial institutions must retain ultimate human responsibility for AI-driven decisions, despite over 70% of Malaysian financial service providers already employing AI applications.

The Structural Challenge of Autonomous Systems

At its core, the mechanism of AI adoption in finance moves beyond mere technological deployment; it necessitates a re-evaluation of decision-making authority. While AI excels at processing vast datasets for fraud detection, credit risk assessment, and customer service, the underlying first principle remains that machines can inform, but humans must be accountable. Governor Ghaffour underscored that AI should be viewed primarily as a governance issue, not solely a technological one.

This perspective introduces a critical framework: the ‘Human Accountability Framework’ for AI. In increasingly complex AI models, financial institutions must be capable of explaining outcomes, challenging decisions, and maintaining clear lines of responsibility. Boards and senior management are therefore tasked with prioritizing AI governance, linking its deployment to explicit business objectives, measurable value, and a defined risk appetite, rather than delegating ultimate responsibility to algorithms.

Evidence and Ecosystemic Imperatives

The evidence of AI’s pervasive growth in Malaysia’s financial sector is clear, as indicated by responses to BNM’s Discussion Paper on Artificial Intelligence. While current applications span internal efficiencies and risk management, Abdul Rasheed urged institutions to broaden AI’s scope to address wider industry challenges, including scams, fraud, and cyber threats. This expansion requires an ecosystem-wide cooperative effort, exemplified by collaborative initiatives involving BNM, the Royal Malaysia Police, financial institutions, and PayNet, which collectively enhance safeguards and intelligence sharing.

Realigning Organisational Structures for AI Governance

A common pitfall in AI integration is assuming that technological advancement automatically translates to robust governance. What many overlook is the fundamental shift required in organizational structures. Risk, compliance, and internal audit teams must expand their roles significantly to ensure institutions can explain and stand behind AI-driven outcomes. This necessitates a proactive investment in human capital development, with initiatives like AICB’s Future Skills Framework designed to equip banking professionals with essential AI literacy, ethical judgment, and governance capabilities.

This commitment to a governance-first approach informs BNM’s broader strategy: fostering innovation while preserving trust and ensuring societal benefit. This encompasses proactive industry engagement, clear regulatory guidelines, and responsible experimentation. Looking ahead, the central bank is progressing with its Open Finance framework, slated for phased implementation from 2027 with support from PayNet, alongside pilot phases for its asset tokenisation roadmap via the Digital Assets Innovation Hub. The durable lesson is that technological progress in finance is only sustainable when firmly anchored by clear governance and human oversight.