Artificial intelligence is fundamentally changing how retailers and merchants attend to customers and even how they manage inventory, all via technology such as chatbots and facial recognition. And now, AI can potentially transform lending in the sector, too.
Banks, lenders and merchants with white-labeled lending offerings could eventually use AI to address credit risk assessment, personalize loan offerings, and enhance the customer experience. As the CEO of a fintech company at the forefront of embedded lending, I wanted to delve into what adopting AI in retail lending could look like — including the issues that may arise.
The Possible Advantages of AI Underwriting in Lending
Chatbots and virtual assistants automated with generative AI, where borrowers can receive loan recommendations and customer support, are already in existence. However, AI-driven processes could bring more advantages to the retail lending landscape.
Unlike the underwriting systems currently using algorithms, machine learning and data analysis for faster decisioning, AI would apply much wider data parameters and teach itself based on past performance by assigning different weights to various elements of a person’s background. This would lead to better outcomes when assessing consumers for financing.
For example, AI-driven algorithms could analyze data like financial records and industry trends to identify patterns and assess credit risk more quickly and more accurately — even during moments when application volume is high. These advancements may allow lenders to mitigate risks when determining consumer creditworthiness, streamline loan approvals, and encourage sustainable and well-informed lending decisions.
Also, by analyzing historical data, AI could allow lenders to offer customized financial products later in the customer journey, both in terms of loan amounts and interest rates.
Needing AI Regulation to Navigate Potential Breach of Customer Liberties
On the other hand, using both existing algorithms and newer AI technologies in retail raises legitimate concerns about a lack of human empathy and subjective decision making.
While today’s type of underwriting is often automated without much (or any) human intervention, AI allows for a much more sophisticated and dynamic underwriting process. AI could almost function as a “black box,” meaning we wouldn’t always be able to properly scrutinize the factors and calculations behind underwriting.
Some argue that this will exacerbate inequalities. If the generative AI relies solely on historical data to learn patterns, this could perpetuate biases and unfair treatment towards specific racial groups or marginalized communities, impacting borrower creditworthiness. For example, someone who has no credit record and lives in a lower-income neighborhood, yet is educated and deserving of credit, might not meet the specified AI criteria for receiving the credit.
Regulation is already paramount. In the lending sphere, it serves to protect the borrower from potential unfair loan practices. With the use of AI in lending, regulation would become even more critical.
To achieve precise, fair underwriting, AI would need to analyze millions of data points about an individual. This poses the risk of potentially exposing customers to big-brother-type practices. Regulation would safeguard the personal liberties of borrowers, counter potential discriminatory outcomes, and balance increased bank profitability with the protection of the individual.
As AI becomes more commonplace for many aspects of lending, banks and merchants must prioritize transparency, open communication, and data privacy. They need to clarify how AI, machine underwriting, and human expertise each play a role in their lending process.
Yaacov Martin is the CEO of Jifiti, a fintech company that powers white-labeled embedded lending solutions for banks, lenders and merchants worldwide.
View Original Article
