For banks
AI risk management for banks
The AIRG applies to every bank in Singapore. Most of the discipline you already have.
AI turns up across a bank: in credit scoring and lending, fraud detection and anti-money-laundering, trading and treasury, operations, and customer service. The MAS AI Risk Management Guidelines (AIRG) apply to banks as they do to all financial institutions, so each of these uses needs to be identified, rated for materiality, and controlled in proportion to the harm a failure would cause.
The good news for banks is that you start from the strongest base. Banks have the deepest model risk management heritage in the sector, and AI risk is largely an extension of it: the 2024 MAS thematic review that fed the AIRG was a review of banks. The work is extending your existing model, third-party and technology risk disciplines to cover AI, rather than building something new. A fuller guide for banks is coming.
Where to start
- The complete guide to AI risk management in finance in Singapore - the whole picture in one place.
- The AIRG explained - the guidelines your bank is supervised against.
- The AIRG in practice - identification, inventory, materiality and lifecycle controls as a working system.
- The AIRG and third-party AI - where the final tightened most, and where most banks' real exposure sits.
I developed the AIRG while leading AI risk supervision at MAS, and now advise banks independently. I am building a fuller guide for banks; subscribe to get it, and future updates on the AIRG.