Bank Negara Malaysia (BNM) and Payments Network Malaysia (PayNet) are developing an AI-driven fraud detection system targeted for rollout in 2026, according to The Sun.

The system will use machine learning, large language models and predictive analytics to identify suspicious transactions before they are completed.

BNM deputy governor Datuk Marzunisham Omar said the system will alert users to potential scams and prompt them to confirm transactions, forming part of BNM’s wider push to curb mule accounts and online fraud.

He added that the project aligns with BNM’s broader digitalisation agenda, which includes the Digital Asset Innovation Hub launched in June 2025.

The hub serves as a secure space for regulators, financial institutions and fintech firms to develop and test new digital financial solutions.

Marzunisham also cited a 2024 AI survey by the Asian Institute of Chartered Bankers showing that over 70 percent of Malaysian financial institutions have implemented or are exploring AI and generative AI tools.

The industry is now developing a shared framework to guide responsible adoption.

PayNet is simultaneously working on a federated AI model known as FinancialGPT to strengthen fraud detection and enhance payment system security.

PayNet chief of staff Dr Endry Lim Zhen Wan said the model will be trained entirely on local financial data to improve cybersecurity and ensure sensitive information stays within Malaysia.

The initiative comes amid rising online scams. The Home Ministry reported RM1.12 billion in losses during the first half of 2025, following RM1.58 billion in 2024.

BNM and PayNet had earlier launched the National Fraud Portal in August 2024 to automate scam reporting, trace stolen funds and enable faster data sharing between banks.

The portal supports the National Scam Response Centre in coordinating efforts to safeguard consumers.

Featured image: Edited by Fintech News Malaysia, based on image by ilygraphic via Freepik