On 5 December 2024, the Monetary Authority of Singapore published a review of observations on risk management of artificial intelligence models for financial institutions. In particular, AI testing requirements focus on robustness through sensitivity analysis, sub-population assessments and stress testing, as well as fairness assessments and adversarial simulations to ensure resilience and fairness. The review notes that performance thresholds, pre-deployment checks and ongoing monitoring for data and model drift are also necessary to ensure consistent, reliable and fair AI behaviour. In addition, the review notes that third-party AI introduces challenges that require compensatory testing and enhanced contingency planning.
Original source