NEWS / NOV.2026
Singapore to begin studying a shared anti-fraud service for banks in early 2027
In early 2027, the Monetary Authority of Singapore will begin a study to design a shared service for banks to detect scams. An ongoing trial using transaction records from five banks aims to assess whether artificial intelligence can spot suspicious flows of funds that an individual bank misses.

Singapore has announced that a study of an interbank scam detection service will begin in early 2027
In early 2027, the Monetary Authority of Singapore (MAS) will begin a design study for an interbank scam detection service. Gan Kim Yong, Deputy Prime Minister and chair of MAS, announced the timetable on 18 November 2026 at the Singapore FinTech Festival.
The study aims to clarify who will be responsible for the service, called the Scam Detection Utility, and how it will operate. The work will initially focus on accounts and data in Singapore. The launch date remains unknown.
According to Gan Kim Yong, a victim’s money can pass through several banks in a matter of minutes. Suppose a customer of bank A transfers money to an account at bank B, and the holder of that account then sends the money to bank C. A sees the first transfer and C sees the second. B sees both transactions on its customer’s account, one as a credit and the other as a debit.
Combining transaction records from several banks would make it possible to trace funds beyond the transactions each bank can see. Two transfers made in quick succession can also have an ordinary explanation. A business might, for example, receive a payment from a customer and then pay a supplier.
On 4 May 2026, MAS announced a trial using actual transaction records from five banks. The trial, still underway in November, aims to assess whether artificial intelligence models can spot suspicious flows of funds in the combined records that a single bank would miss. MAS has disclosed neither a detection rate nor a figure for losses prevented.
Only authorised personnel can access the trial data. In the transaction records, a digital fingerprint generated by hashing replaces each account number. Only the bank that supplied the number can identify the corresponding account. MAS plans to delete the data at the end of the trial.
On 18 November, MAS and the Future of Finance Institute invited technology companies, AI labs, solution providers and research organisations to apply for access to synthetic data. Developed with the government agency GovTech, the dataset is intended to reproduce patterns of interbank transactions and scams observed during the trial. The teams selected will be able to use these simulated transactions to train and evaluate their own models.