미래 금융 시스템의 구축 - 신뢰할 수 있고, 연결되며, 회복력 있는 시스템

Good morning to all. It is a great pleasure to participate in the Global FinTech Festival, one of the largest FinTech events in the world.

Let me start by noting that we are all standing on the cusp of great technological transformation over the next 10 years. Three technologies stand out.

The AI transformation is already upon us. Tokenisation may take a little more time to scale, perhaps another few years. Quantum computing is a little further out. Experts estimate 5-10 years. But it is not too early to prepare now for quantum resilience.

All three technologies have far reaching consequences but let me focus on AI. The technology is advancing the most rapidly, and adoption is spreading the fastest.

Model performance is clearly advancing rapidly. It is well known that AI models already perform at expert or specialist level in many defined tasks, such as in coding, graduate-level science and mathematics, and general knowledge work.

Where there are still gaps to human performance are in areas requiring complex interpretation, strategic judgement and decisions, and human interaction skills.

Corporate adoption of AI is progressing rapidly. While a high proportion of companies report using AI, a much smaller proportion report significant productivity gains.

The proportion of firms reporting productivity gains are likely to increase as employees and organisations become better users of AI through training and process and product redesign.

In time to come, we can expect AI to become a foundational capability for organisations. Those that use it well and safely will innovate faster, serve customers better, and unlock new opportunities.

The challenge for regulators is to support innovation for competitiveness and growth; and also steer it on a course that is safe and sustainable.

Innovation must be founded on trust and stability if it is to scale.

In Singapore's financial sector, we are seeing wide AI adoption. Common use cases include fraud detection, credit underwriting, risk management, regulatory compliance, marketing, customer service and document processing. These use cases have moved beyond pilots and are being deployed at scale.The largest, well-managed financial institutions need no encouragement. They are in rapid adoption. Our focus with them is good governance — around safety, guardrails, and accountability.AI should not only be a competitive tool wielded by the largest financial institutions. We should avoid a winner-takes-all dynamic if we are to maintain a competitive and stable financial system to support the public and the economy.We want the benefits of AI to spread across the whole industry. Our aim should be for a sustainable productivity uplift for the entire sector.

To that end, MAS launched Pathfin.ai. This is a platform and programme to share and match validated AI solutions across the industry.Smaller financial institutions can lower the cost and effort of finding and deploying effective AI solutions. If a solution has been validated and works, we want to make it easier for others to find and implement it.We now have over 300 participants and a growing number of successful matches.

AI also gives us an opportunity to solve problems at the system level — problems no single financial institution can solve alone.Detecting and disrupting scams and fraud is an example. MAS is working with the law enforcement agency, and the banking industry. We are testing different AI models, drawing on cross-bank and public-private data, to improve the detection of suspicious accounts and transactions in near-real time.The aim is to detect sooner, intervene faster and reduce losses. We expect to have findings from this work by the end of this year.

To propagate sound AI governance and risk management in the financial industry, we have found it very useful for MAS to collaborate with the industry.The first step was taken in 2023. MAS and the industry collectively published a Gen AI risk framework to establish a common understanding and baseline.The second step was taken in 2025. We published jointly with the industry two AI Risk Management Handbooks to promote good practices. These recommended practices were applicable across the banking, insurance, capital markets sectors.MAS has also issued a set of Guidelines for AI Risk Management for public consultation. These guidelines set out supervisory expectations for governance and risk management, as well as AI life cycle controls and capabilities.The Guidelines will work in tandem with the Risk Management Handbooks. The first will set out the what. The latter will set out the how. For example, the Guidelines require financial institutions to perform risk materiality assessments, while the Handbook provides examples of how such assessments can be implemented.Updating good AI risk governance practices will be important as the technology progresses fast.The latest joint initiative with industry is SAFR — Safeguards for Agentic Finance at Runtime. As AI ag