Money FM Q&A: Using AI to Assess Credit Risk
06 August 2020

Artificial Intelligence is going to be part of decision making that determines credit risk even how financial products are marketed to you, what safeguards are in place to ensure that AI does not perpetuate or exacerbate biases and that human factors of fairness are also facts and into decisions driven by machine intelligence. It may be interesting to hear the monetary authority of Singapore is working with banks and technology firms to develop measures to judge customers fairly when AI is used to assess credit risk.

The steps are being taken to promote the responsible use of AI and data analytics. Dong Shou, Chief Operating Officer at ADVANCE.AI who believes Singapore can take the lead in this regard, and he will provide us some understanding and insights related to this field.

MONEY FM: How AI and facial recognition technology is currently being used by banks or financial services or payment and retail and e-commerce businesses for credit risk assessment?

DONG SHOU: Let me start with how AI has being deployed in the financial services and banking industry for now. Essentially AI is not like magical, it’s like mathematical work models, which can be changed to detect high risk behaviors, based on certain kind of data, for example, your repayment behavior, loans you have, as well as other traditional data such as age, your job information, income level, and all the family sides come into the play. The advantage of AI is that it can be changed to do this kind of vast amount of data and documents faster, cheaper, safer, and with more accuracy than human agents.

In the emerging markets, like Indonesia,the Philippines, Vietnam, even India, where we do a lot of business, up to 80% of the people we deal with don’t have any bank account or history. But that doesn't mean they are not quite worthy, as many of them are small business owners.

So there, AI can help financial service companies determine credit risk and profile based on other kinds of data, such as behavior, for example, how you use your smartphone, how you shop online, how much you pay for your data service, what is the social graph data looks like, etc. In e-commerce and retails specifically, AI and facial recognition can be changed to scan merchant business registration licenses, banking statements, and even like a product if it’s original and high quality. So, all these come together can help determine the credit risk before they are allowed to sell on major e-commerce and platforms. In summary, by adopting AI and applying AI into massive autonomous data, we can give credit and risk analysis to people, who are not being verified by financial institutions yet.

That’s how we see the AI technology and how we believe the importance of AI to the financial industry, especially in the financial inclusion.

MONEY FM: Why these became a push towards understanding fair governance and trying to come up with a framework for fair credit assessment?

DONG SHOU: This is so called a FEAT framework, fair, ethical, accountable and transparent. We certainly believe that AI should be equal to everyone. It’s not like used only by powerful people and institutions. For example, can we explain what factors influence the credit score? Is it easily experimental and justifiable? Is it transparent to everyone involved both internally and to customers? Is AI being biased due to certain factors? Who is accountable for this process? And for all these questions, we need to bear them in mind when we design the algorithm. It’s not like a black box that no one else knows about what had happened or why it will happen like this, but it should be transparent to people.  

In the end, trust is the most important thing. We need to educate and help people understand how AI works. on this at the first place, Singapore has a national initiative called Veritas, to propose the responsible use of AI and data among financial institutions to market financial products to people. A white paper will be released by the end of this year for all banks and the financial institutions to study, and that could be a part of this to make sure AI is fair to the financial institutions and to customers.

MONEY FM: It seems like the landscape is constantly changing, for example, can you help us understand how consumer behavior and even trends in banking, what people expect from their banks and financial services has changed because of COVID-19?

DONG SHOU: Because of this Covid-19, people actually don’t really want to go to a very crowded place for shopping or bank service. We found a trend that will happen after the COVID-19. One is the way that people deal with banks and other financial institutions, fewer people will visit the physical branch and many started using digital banking for the first time. Meanwhile, many banks also reduce the number of physical stores because they want to reduce the physical touch. And this allows the business to rethink, how are they going to make online operations and businesses more efficient with AI technology. In Indonesia for example, we are helping many banks to operate in a digital way. That’s the first change that we see.

The second change is the decline of using cash. For example, in Indonesia, most people used to bring cash to shop and purchase stuff. But because of this COVID-19, we find they have to use a lot of payment tools, E-wallet, Alipay to replace cash. They can basically shop online, pay stuff and transfer money online. In that case, like in many other countries, people will reduce the use of cash, and instead, we are using these online payment tools to finish all this kind of transaction.

Lastly, we see the change in online shopping and delivery. Because people just don’t want to spend too much time offline to do shopping so that they go online for shopping and try to use some APPs for food delivery. All these kinds of behaviors will be a new normal after COVID-19. That’s something I firmly believe.

MONEY FM: How is ADVANCE.AI working with banks to ensure good governance? As for the listener, Advance.AI is a data driven financial technology company.

DONG SHOU: Let me give you some examples by working with some banks in Indonesia, as it’s our biggest market.  We offer a bunch of services related to AI and specifically AI to the financial industry. We offer credit scoring; computer vision and risk management services. For credit scoring, traditional methods like traditional score cards don’t really apply to those people who don’t really have any credit history. We have to design some alternative database to credit scoring models that like financial history, social network, e-commerce and behavior data like all these. That’s something we call alternative database credit scoring.  

The second is computer vision. By computer vision, I mean technology like OCR to recognize the words on your ID card using AI, and facial recognition to compare the faces between yourself and the face on the ID card, and the lively detection to make sure the person behind the camera is alive, not like a robot. All these we call computer vision. By using computer vision technology, bank, e-commerce, they can onboard customers online, and do this kind of ID verification. They don’t need to have people face to face to do it. It’s very easy to onboard new customers, especially after the COVID-19 situation.

The third business line that we are doing is called data risk management. With the credit scoring, with the computer vision technologies, banks do need to have a kind of decision engine to make sure if they want to onboard or reject this new customer for bank accounting, bank account opening or like lending kind of business. We offer like an engine, a collection and even like a lending system for our customers, basically they adopt this end to end solution. So just adopt it and try to start online lending with our solution.

MONEY FM: We started by talking about SME(small medium enterprise) and how it's working with banks and it's working with tech firms to create a framework, when it comes to assessing credit risk and how AI is used in that regard.  I want to ask what needs to happen, do you think so that the drive towards good governance,on the one hand, does not hamper innovation of companies like yours?

DONG SHOU: Actually, we are applying for a virtual banking license in Singapore, as we are one in the top batch. We are actually waiting for the result on September this year.

So for the reason that we want to apply for this virtual bank license, is that we believe Singapore is a great market to be more creative on scoring for SME.  That’s why we apply for this wholesale banking license like this. We believe like using our risk management capability, credit scoring capabilities, AI capability, we do can help to renovate the way that we do for lending to our SME customers, like a lot of mama-papa shops here are in Singapore.

Especially after COVID-19, I think a lot of SME, they are hungry for the capital to operate their business. Banks, they also need someone to help them to basically to do like as at a bridge, to bridge this small medium like a source. The bank wants someone to help them to do this kind of risk analysis and credit scoring things. So that’s the first thing that we can help with. And regarding a question, I do feel like Singapore is actually doing a lot of initiatives to make this happen, make this whole thing happen. So for example, they have a program called the Veritas. And we are actively discussing to be part of the program as we hope as a Singapore AI company, we should be on the list. We are eager to be very actively involved, and try to contribute our knowledge, our expertise in Singapore, try to help Singapore finance, you call it like innovation.

MONEY FM: You bring with a lot of experience there, understand you worked as a senior engineer in amazon over in Seattle, before taking on this role. Why are you so confident and optimistic that Singapore can take a lead when it comes to AI framework for credit risk?

DONG SHOU: The first reason is that I feel like Singapore is a hub, especially in Southeast Asia. And as a heart, you can attract a lot of talent, a lot of capital. That’s why we have staff in ADVANCE.AI, and many others here in Singapore at the headquarter.

The second reason is I do believe its government is very clear that AI will be the next big thing to boost the economy here. The Singapore government actually put a lot of effort into supporting the startups, especially AI startups, fintech startups. Even in education, a lot of high school students start learning AI. Also, in elementary school, they start putting a lot of courses about AI. All these make Singapore be a very good place around AI innovation. That’s the second reason.

The third reason is its geographic location, very close to all these countries, Indonesia, the Philippines, Vietnam, like within two hours, right? All these countries, they are also talking about financial inclusion. They want to basically use AI to have more people to be more accessible to financial services, while they may have difficulty in terms of talents in this industry. And that’s why Singapore has been the hub, they can offer technologies, talents and products to serve these markets.

So, for these three reasons, we believe Singapore is the best place in this region. To apply AI to renovate a lot of industries, including traditional industries even like retail, even like logistics. Right? Actually, all these industries can be improved by using AI, I do feel a lot of confidence in that.

Follow us:
contact us, message
contact us, email