Bank Negara Malaysia (BNM) has published a feedback statement on artificial intelligence (AI) in Malaysia’s financial sector, confirming that it will apply existing regulatory frameworks rather than introduce dedicated AI regulation at this stage. Following its August 2025 discussion paper, the statement draws on 113 responses and sets out regulatory expectations, emerging practices and areas for collaboration. Respondents reported growing adoption, particularly for operational efficiency, customer service and fraud management, alongside challenges involving governance, data and external providers.

Existing requirements on governance, risk management, consumer protection and Shariah compliance continue to apply to AI activities. Financial service providers should calibrate safeguards to each use case’s risks and potential impact, with stronger testing, monitoring and human oversight for applications that materially affect customers or financial decisions. Institutions remain accountable for AI decisions and outcomes. Recommended practices for increasingly autonomous agentic AI include defined authority boundaries, access and transaction limits, human approval thresholds and customer redress. For Islamic financial institutions, AI should support rather than replace human Shariah decision making, with final decisions remaining subject to human review and judgment.

From 2027, BNM will work with industry to explore an AI Responsible Innovation Forum for shared learning on governance and emerging risks, with cyber resilience identified as a priority. Other potential initiatives include testing pathways for advanced AI applications and improved access to data, including synthetic data, with privacy and security safeguards. In the near term, supervisory work will examine AI adoption, governance and key dependencies on external providers, potentially informing thematic reviews where warranted by evidence and material risks.

Related updates