AI Lending Could Expand Credit Access for Underbanked Americans

The core architecture of the U.S. credit assessment system dates to the Eisenhower administration from 1953 to 1961. Its longstanding reliance on traditional credit histories has made it difficult for people with limited banking records to obtain loans. AI can analyze a broader range of alternative data, making it a potentially important tool for expanding financial inclusion.

Former U.S. Comptroller of the Currency Rodney Hood recently said AI-powered lending could help about 36 million underbanked Americans gain access to credit while reducing subjective bias in manual reviews. However, the lack of transparency in model decisions and the risk that training data may perpetuate discrimination remain contentious issues for regulators and the industry.