The global credit industry has reached a major turning point this year, with the relentless march of artificial intelligence and machine learning having a massive impact on banks and other lenders’ abilities to effectively assess risk, detect fraud, and generally create a safer, more efficient, and more inclusive financial structure for everyone.
This impact will be the biggest felt in the “traditional credit gap.” This gap refers to those people unable to access credit as they lack the usual metrics lenders use to verify their eligibility. For decades, the traditional credit score has been the primary factor used in the decision-making process. There are millions of people without a traditional credit score who are essentially off the radar to regulated lenders.
This exposes people to the predatory practices of the informal lending market.
How AI Is Changing the Rules
AI is changing the fundamental rules for how a potential customer is reviewed. Creditworthiness can now be ascertained through various alternative data sources such as logs that show mobile payments and behaviours, utility bill payment, business cash flow data, and app engagement patterns, to name a few.
The basic concept is that AI can detect unique patterns of behaviour that can predict a potential customer’s approximate risk when lending. These AI models continue to evolve with real time data feeds, ensuring up to date assessments.
South Africa’s Credit Inclusion Challenge
One of South Africa’s largest credit providers aims to implement machine learning to help bridge their credit gap. South Africa has one of the most significant gaps of any nation, with almost 50 percent of residents lacking formal access to regulated credit.
Brett van Aswegen of short term loan giant Wonga believes the AI age can help develop a new generation of ‘financial identities’ from these alternative data sources that will pave the way for future inclusion in the regulated marketplace
The focus here, of course, is on ensuring sustainable and safe lending practices that protect both the customer and the loan provider. However, Wonga is also investing significantly in financial education that can help thousands of South Africans avoid dependency on credit altogether, counterintuitive for a short term provider but something that Brett feels strongly about.
Global Momentum in AI Driven Lending
Wonga’s effort is part of a global shift towards AI driven inclusion. We’ll now look at Tala, another financial service operating across Africa, Mexico, India, and the Philippines. They have already delivered over six billion dollars in credit to more than ten million customers.
They have accomplished this with the help of ‘Tala InSight’, their AI tool that helps personalize credit limits, which results in larger loans, lower levels of default, and better customer lifetime value.
AI and Fraud Prevention
Outside of risk assessment and short term loan accessibility, AI is having a massive impact on fraud detection and prevention. These real time systems are able to flag unnatural behaviours by using a huge data pool of authentic customer data to build typical behaviour patterns of legitimate users. Everything from when, how, and where a person buys can be used to build this profile.
The Challenges Ahead
Of course, the system is not perfect, with major challenges including the risks of privacy and transparency, as well as machine bias, especially if an AI model has been trained on historical and unequal data records.
It is essential that these models continue to evolve and implement key safeguards such as regular bias audits (humans will still be needed), transparency in data use policy, AI decision narratives that have clear decision reasoning, and real time compliance with evolving financial regulations. The main goal, however, stays the same: democratize access to credit without sacrificing ethics or fairness.
No one can predict what the next ten years in the credit market will look like. Machine learning models have advanced at a staggering pace in the last two years alone. Investment in this time increased from thirty-five billion dollars to over one hundred and twenty-six billion. It’s a reasonable bet that transactions are going to become much faster, and credit usage in general will become more regulated and safer to use.
True success will depend on the effective collaboration of banks modernizing their infrastructure, fintechs innovating for local markets, and financial regulators keeping their head above the rising tide of machine insights to steer the market in the best direction for the greatest number of people, which may not always align with what the companies investing billions into the technology want. Constant vigilance and transparency will be paramount.
