KUALA LUMPUR—Bank Negara Malaysia (BNM), the nation’s central bank, and Payments Network Malaysia (PayNet) are developing a sophisticated, AI-driven system to combat the country’s escalating crisis of online financial scams. Slated for a 2026 rollout, the new technology represents a major technological escalation in Malaysia’s battle against digital financial crime, aiming to stop fraudulent bank transfers before victims lose their money.
The initiative is a direct response to the staggering financial toll of online fraud, which has become a national concern. The system, utilizing cutting-edge machine learning and predictive analytics, is intended to act as a digital sentinel watching over every transfer made through the country’s payment network.
The New AI Shield: How It Will Work
The core of the new system is its ability to analyze vast amounts of data in the brief moments between a user initiating a transaction and the funds being credited. It will leverage machine learning (ML), large language models (LLMs), and predictive analytics to identify unusual or high-risk activity that might signal a scam, such as sudden, large transfers to previously unknown accounts or transfers matching mule account profiles. BNM Deputy Governor Datuk Marzunisham Omar explained that when the system identifies a threat, it will issue an immediate alert to the user. This alert forces a pause, prompting the user to confirm the transfer details—a simple but effective way to interrupt the scammer’s flow and protect the consumer.
Fighting the Tide of Scams: The Context
The central bank’s urgent investment in AI is driven by alarming statistics on financial losses. The Home Ministry reported that financial losses due to scams reached ₹355 crores (about $335 million U.S.) in 2024. This trend of rising online fraud has put immense pressure on financial institutions and regulators to implement more robust defenses. The AI system is designed to provide a much-needed layer of proactive, automated security that current, more static defenses have been unable to match against increasingly sophisticated criminal networks.
FCRF Launches CCLP Program to Train India’s Next Generation of Cyber Law Practitioners
A Shared Data Approach: The ‘FinancialGPT’
In a parallel effort to strengthen cybersecurity, PayNet is developing its own specialized AI known as FinancialGPT. This federated AI model is being trained exclusively on local Malaysian financial data. PayNet’s strategy ensures that the intelligence used to detect fraud is highly relevant to local scam patterns, while also addressing critical data privacy concerns. By keeping the training data entirely within Malaysia, PayNet aims to improve the precision of fraud detection while ensuring sensitive information remains protected, creating a domestically focused shield against cross-border financial crime.
A Broader Digital Push
The new fraud detection system is part of a wider, multi-pronged digitalization agenda led by BNM to modernize Malaysia’s financial sector. This broader push includes measures to enhance transparency and accelerate responses to crime. In August 2024, BNM and PayNet launched the National Fraud Portal, a key tool designed to automate scam reporting, trace stolen funds, and enable faster data sharing among banks. Furthermore, the central bank launched the Digital Asset Innovation Hub in June 2025. This hub serves as a secure, regulatory sandbox for financial institutions and fintech firms to develop and test new digital solutions, creating a framework for responsible, technology-driven innovation across the industry.
