Josephine Teo on how AI uplifts Singapore's financial services industry

Speaking after visiting UBS's AI centre of excellence in Singapore, Josephine Teo describes how MAS pushes leading financial firms to set up AI centres of excellence. More than 30 financial institutions now run AI teams, with some exporting Singapore-built innovations to their global operations.

She emphasises bilingual AI talent — people who know data science and machine learning and also understand business processes. MAS has updated its model governance framework for generative AI and built Project Moonshot for red-teaming and benchmarking on top of AI Verify. Through the Pathfinder programme, companies that completed proofs of concept share results with peers, and total AI centres of excellence across sectors now exceed 50.

Caption language: en · Fetched: 2026-05-02

Singapore is a leading and trusted financial services hub and we'd like to help the companies that currently operate here uh to gain even stronger competitiveness as a way of growing this hub and strengthening it. I think one of the ways in which um uh we want to help is uh by using AI to a broader and a deeper extent and we can see that um the AI applications should serve a business purpose. The business purpose here is that uh the companies would like to be able to serve a wider client base. Um it's also because it is a growing region and there are business opportunities and how they are able to capture these new growth opportunities is an efficient in an efficient way is one of the driving forces.

Um now for the financial services sector the agency that leads its development is the monetary authority of Singapore and MEES has introduced a number of uh initiatives to try and help these companies to use AI to strengthen their competitiveness. uh among the initiatives that they have uh introduced is to encourage the setting up of AI centers of excellence by leading financial institutions here and the organization that we are visiting today UBS is a very good example of this. Another very important way to uplift the sector through the use of AI is to enable more to learn uh the skills that will help them to succeed uh in these endeavors.

uh one consistent theme that we keep hearing is that you need AI specialists, people who know data science, who know machine learning, engineering, but equally you need people who are steep in the understanding of the business processes. They understand the domain well. They understand the functions well. They understand the client relationships well. And when you bring the two together then something special happens for the company. This is the group that we have broadly referred to as AI bilingual. Today I met some of them. They are specialists in their respective areas. For example, in business operations. They know what are the pain points. They also know what are the opportunities for improvement and they acquire enough of AI skills and knowledge in order to apply it to their day-to-day uh activities.

And in this way I think we can strengthen the sector's ability to use AI and we will continue to press on in this area. The third part of um MAS's efforts is also in ensuring that uh there are sufficient guardrails and to this end they have uh put out governance principles that financial services institutions should observe when they are using AI and they will continue to uh strengthen the guidance uh to the players in in the sector. I think these efforts are showing uh good results. uh by now there are already more than 30 financial institutions in Singapore that have AI teams. Some of them uh have pitched it at a very high level of center of excellence where the innovations that are being born in Singapore are being deployed uh applied to their global operations.

We see this as a very good way of showcasing the capabilities here and to also uh promote the continued uh development of the AI ecosystem in Singapore. Okay. >> Yeah. >> So I guess sure does this partnership also mean for Singapore as like the country strives like you said become a global leader in AI and also in terms of the talent how can how how do such partnerships help grow talent in the sector in Singapore? Yes, I think we know that um the skills that you develop uh working on specific problems or specific opportunities uh will be so much more useful um in addition to the skills that are learned uh in a educational institution in an educational setting. Those give you some grounding but you need to be able to understand the specific context and issues that companies face in using the technology.

Uh you need to be able to understand the perspective of the clients. You need to be able to see the challenges that your co-workers face. So in real world applications I think skills uh that are developed are far more entrenched and they uh stay with a person for a much longer time and that's what we'd like to uh see more of. would like to see uh real applications that are being developed by companies uh that make a difference to their businesses and in the process of developing these AI enabled solutions. The people that they deploy then gain a different kind of experience that will also enable them to work on similar projects within the same organization or to apply their skills more generally. Yeah, >> minister, you mentioned rail.

>> What kind of regulatory challenges do you see the adoption of AI in financial services and uh how is government going to tackle these challenges? >> Yeah. Um the concerns with AI uh are quite similar across different settings. Yeah. For example, if AI is being used to make uh recommendations, uh will those recommendations result in costly mistakes? uh or if uh AI is being used to uh make assessment, right? Um is the assessment unbiased or does the AI uh carry with it certain discriminatory uh practices and and that get persists over time. So I think these two types of risks um are not unique to the use of AI in financial services. Uh what we have done in Singapore is to lean forward um to make companies more aware of these risks to begin with.

Um in our case we've had a model AI governance framework uh even in the first iteration of the national AI strategy and when we introduced the refresh uh national AI strategy uh we were working at the same time to uh also uh update our AI governance framework to take into account uh the different concerns as a result of the use of generative AI. I we've gone a step further. uh we have uh previously already introduced testing frameworks, software toolkits that were applicable in the use of classical AI but we created uh project moonshot as a way for us to say in the context of generative AI how do you test for these kinds of risks and what are the ways in which you can you know assure yourself that uh the risks have been managed to a adequate extent. So these efforts uh are very important and they continue to uh occupy our minds.

We don't do it alone. We work together with companies that are actually implementing AI solutions so that they also have a chance to test the uh guard rails that have been proposed uh to give feedback to IMDA that looks after this area so that we can continuously improve these guardrails. Okay, one more question. Yeah, sure. Do we expect any more centers of excellence between the government and the financial sector in the upcoming few months? >> I'm sure it will continue to grow uh across the different sectors. There are already more than uh 50 uh AI centers of excellence and um we are uh welcoming of the AI centers of excellence uh whichever sector u they come from.

The more important thing is the kind of work that they are doing uh and whether um the applications that they develop are impactful beyond the specific enterprise in Singapore to their global operations as well as if they can be also shared with the wider uh AI ecosystem in Singapore. So one good example is MES's Pathfinder program where they bring together uh companies that are interested in using AI enabled solutions and those who have uh successfully piloted proofs of concepts can then share with others uh who in a way will have a shorter learning curve. um and they can also better identify the partners that they can work with to implement these solutions.

So I think in this way the ecosystem can uh enhance its uh offerings and support to the participants and as a result help us to push forward with the AI adoption agenda in in a much more concerted and systematic way.