Singapore’s central bank wants all FinTech AI use cases subject to independent review

Singapore's Monetary Authority (MAS), acting as the central bank and finance industry regulator, has issued new guidelines requiring all local financial institutions to undergo independent review of their artificial intelligence (AI) use cases before deploying such technologies. The island nation, known for its low taxes and strict regulations, is a leading finance hub in the world.

MAS, which aims to maintain its status as a premier finance center, warns that AI introduces additional risk due to its complexity and probabilistic nature. The regulator emphasizes that generative AI is even riskier because of inherent training data noise, non-representative data, and the model's inability to handle unseen scenarios.

To mitigate these risks, MAS advises financial institutions to expand their risk management frameworks to encompass AI, with board and senior management playing a crucial role in this process. Before deploying AI, MAS wants firms to have their AI use cases, including underlying systems and models, independently reviewed to ensure proper controls are in place.

Additionally, regulated entities must conduct technology and cybersecurity reviews to guarantee secure AI deployment. Ongoing monitoring of both internal systems and any third-party AI is also crucial, as AI dynamics and potential model staleness over time can lead to performance degradation. MAS holds financial institutions fully accountable for the AI used in their services, even if the technology is developed, operated, or provided by third parties.

To safeguard against third-party AI risks, MAS urges institutions to maintain updated inventories of all AI in use and develop contingency plans for high-risk applications, including alternative systems or manual processes to ensure business continuity in case of AI failure or unexpected behavior. The guidelines are set to take effect on October 7, 2027.