Singapore’s Monetary Authority has initiated a joint project to harness the use of AI and machine learning (ML) techniques to fight financial crime and enhance scam detection capabilities.
MAS said in a press release the proof-of-value (POV) will involve partners in the banking industry, the Government Technology Agency of Singapore, and the Singapore Police Force.
At the heart of the project is a plan to bring together data from five participating banks and store it in a secure environment. The data pool will then be used by experts to develop AI and ML models that can help identify higher-risk transactions and accounts.
“Prompt identification could enable timely assessment, intervention and reduction of customer losses to scams,” MAS said in the press release, adding that the project will help build more robust and accurate AI/ML learning models.
MAS has already provided industry partners with a secure data-sharing environment for the project. The system is governed by policies and protocols to safeguard customer information. This will ensure that all data used in the project is protected and used responsibly, the regulator said.
Data protection will also be supported by the use of cryptographic techniques. All bank account numbers used in the project will be hashed (a one-way algorithmic process which substitutes input data with a unique set of generated values). This will ensure that only the contributing bank can identify actual account numbers.
“Data access will be restricted to authorised personnel within a controlled setting that will be continuously monitored throughout the POV.”
Industrywide collaboration
MAS says it will assess the findings and the project and, depending on its effectiveness, it could expand the scope and sophistication of AI/ML models used. The regulator said the project will help establish industrywide collaboration to fight financial crime by using AI and ML.
Future expansion could also involve incorporating broader datasets, and a wider set of use cases across the financial sector in order to strengthen defenses against criminal activity.
The initiative is the latest in a series of efforts by MAS to enable and encourage AI adoption across the financial sector. In March this year, the watchdog announced the publication of an Artificial Intelligence (AI) Risk Management Toolkit for the region’s financial services sector.
The toolkit is developed through joint collaboration between 24 leading banks, insurance companies, capital market firms, and other industry partners, and confirms the completion of Phase 2 of Project MindForge, an initiative launched by the regulator in 2023 aimed at supporting organizations deploying AI technologies in their services and operations.
MAS said at the time the toolkit “provides financial institutions with resources for managing AI-related risks across traditional AI, generative AI, and emerging agentic AI technologies.”
And in November last year, the regulator launched a consultation paper on proposed guidelines for AI risk management within the financial sector. It said at the time the guidelines will apply to all financial institutions.
MAS’s announcement comes at a time of increased scrutiny and unease around the use of advanced AI models in the financial and other sectors. Regulators in the UK, Australia, and other jurisdictions have raised concerns about cybersecurity risks posed by frontier AI models such as Anthropic’s Claude Mythos.
Last week, senior IMF officials singled out Mythos in a blog post and warned that such advanced AI models could destabilize the financial system if not dealt with carefully, or if they fall into the wrong hands.