Why It Matters: When built and used properly, artificial intelligence and machine learning (AI/ML) models lower costs, increase efficiency, and expand fairness in the financial system. Financial technology companies’ use of AI/ML is a significant driver of competition in the marketplace that ultimately benefits consumers.
Here are five ways the responsible use of AI/ML is transforming financial services for the better:
- Expanding access to credit and reducing racial bias: Millions of American consumers and small businesses lack access to credit because of archaic and discriminatory systems. Traditional credit scores are known to correlate with protected class characteristics (e.g. race, gender, and ethnicity), which may have an outsized impact on historically marginalized groups. But, AI/ML models can better ingest and analyze large data sets, and adjust the weight of various modeling inputs, to reduce bias, thereby approving more borrowers. For example, AI/ML industry leader Zest AI helped one lender reduce the approval gap between Black and white borrowers by 50 percent.
- Breaking down barriers to investment opportunities: Digital investment advisory platforms use AI/ML to help everyday Americans take advantage of investment advice previously only available to the wealthy, allowing them to save for retirement or college, build wealth, and make smart choices about their money. While traditional advisors often charge a one percent annual management fee or require high minimum balances, digital advisors tend to charge a fraction of that. As of 2021, over 3.5 million Americans used a digital advisor, a 23.2 percent increase from the previous year.
- Identifying suspicious financial activity and combating fraud: AI/ML models can sort through large volumes of data to rapidly identify suspicious activities like fraud or money laundering, improving oversight and creating safer financial markets. These technologies can make compliance efforts more efficient and reduce costs. Eighty percent of fraud prevention experts say AI/ML can help reduce payments fraud. Using AI/ML for such “regtech” applications can enhance overall compliance and help safeguard the financial system.
- Giving consumers options with personalized banking: Banks and financial technology companies rely on AI/ML to deliver personalized services and meet consumers’ rising expectations for digital banking. More than seven in ten consumers (72 percent), and millennials in particular (79 percent), say that personalization is important to their banking relationships. AI/ML capabilities can help financial institutions anticipate customer needs and offer customized services like savings recommendations, bill pay reminders, personalized cashback offers, and more.
- Advancing identity verification for frictionless services: As more Americans access financial services digitally, robust and seamless identity verification is paramount. Consumers prioritize convenience, speed, and safety when making an online payment or applying for a loan, and delays can be a barrier to access. For example, almost 70 percent of online shopping carts are abandoned, with consumers citing verification challenges like creating an account (34 percent) and a lengthy checkout process (26 percent). AI/ML technologies like those offered by companies like Onfido can help process large amounts of data to verify consumer identities quickly and efficiently, reducing customer drop-off.
What’s Next: Looking ahead, there is a need for clear guidelines to continue fostering the responsible use of AI/ML, both for companies using the technology and policymakers and regulators overseeing them.
Companies using AI/ML must be able to understand, monitor, and explain the operations of their models. Regulators must have a sophisticated understanding of AI/ML models and explainability techniques. They must also work with industry and other stakeholders to develop explainability requirements for various use-cases of AI/ML.
More broadly, open banking enables consumers to securely permission and share their financial data, enabling AI/ML models to iterate and refine their analysis, leading to better outcomes. The CFPB is currently considering rule-making under Section 1033 of the Dodd-Frank Act to secure the open banking infrastructure and safeguard these consumer data rights, allowing the industry to continue innovating.