MAS partners banking industry to tap AI, machine learning to combat financial crime

[SINGAPORE] The Monetary Authority of Singapore (MAS) on Monday (May 4) said it will harness artificial intelligence and machine learning techniques to combat financial crime.

To enhance scam-detection capabilities, the central bank will work with partners in the banking industry, the Government Technology Agency of Singapore, as well as the Singapore Police Force.

MAS is conducting a proof-of-value (POV) to explore the use of AI and machine learning in pre-emptive scam detection, as part of broader efforts to apply these techniques to industry-wide use cases.

The POV will lay the groundwork for deeper industry collaboration, as well as enhance and complement financial institutions’ existing efforts to prevent and counter financial crime, said MAS.

It added that the POV will use data from five banks to build more robust and accurate AI and machine learning models that can identify higher-risk transactions and accounts.

Such “prompt identification could help to enable timely assessment, intervention and reduction of customer losses to scams”, the authority noted.

To support the POV, MAS has provided industry partners with a “secure data sharing environment governed by policies and protocols to safeguard customer information”.

Similarly, it has established a framework with industry participants to ensure that the data shared is protected and used responsibly.

Data used in the POV will remain confidential and protected with cryptographic techniques, the central bank added.

SEE ALSO

Director and company secretary to be charged over suspected involvement in money-laundering activities

Will the step forward in frontier AI mean a step backward in cybersecurity?

For example, bank account numbers will undergo hashing, an algorithmic process that substitutes input data with a unique set of generated values. This ensures that only the contributing bank can identify the actual account numbers.

Further, only authorised personnel will be able to access the data within a controlled setting that will be constantly monitored throughout the POV, said MAS.

All data used will be deleted at the end of the exercise.

After assessing the POV’s effectiveness, MAS said it may expand the scope and sophistication of the AI and machine learning models used by incorporating broader data sets and a wider range of use cases.