The top U.S. consumer financial watchdog is sharpening his focus on lenders’ use of artificial intelligence when making credit decisions, signaling that fresh restrictions loom for financial firms.
Rohit Chopra, director of the Consumer Financial Protection Bureau, indicated that new regulation may be coming soon. The CFPB’s broad mandate means that everyone from banks to online lenders and mortgages-servicing firms could be affected.
“We are working on a very broad waterfront when it comes to AI, especially the use of data from a discrimination perspective from all facets of the underwriting process,” said Chopra during an interview late Wednesday in his office near the White House. One big challenge is technology that creates content such as images and text based off a user’s prompt, he said, referring to generative AI.
The CFPB has already taken steps to stamp out abuses. The agency last month proposed a rule for home appraisals computed by algorithms. The agency also recently issued a warning about chatbots used by financial firms could run afoul of rules, and has started coordinating with European counterparts on the issue.
Chopra, however, said that efforts were just getting started on biometrics, facial recognition and voice cloning. “We’ve proposed a rule on appraisals determined by AI, and there’s going to be more to come on that,” he said.
The CFPB’s efforts dovetail with a broader push by the Biden administration on AI. Although officials have touted possible benefits, they’re increasingly warning of needs for stricter regulation.
On Monday, SEC Chair Gary Gensler said a concentration among a handful of companies developing widely used artificial intelligence models or data sets could become a destabilizing factor in the financial sector. The Wall Street regulator next week plans to propose new regulations to address conflicts of interest related to brokerages or money managers using AI with clients.
The Fed’s Michael Barr followed on Tuesday, saying that lenders need to ensure that artificial intelligence tools don’t perpetuate biases and discrimination in credit decisions.
After warnings from the White House about the technology, several top firms, including Microsoft Corp., Alphabet Inc.’s Google and OpenAI, are planning to offer voluntary commitments on safeguards as soon as Friday.
During the wide-ranging interview on Wednesday, Chopra, a longtime consumer advocate, said his efforts would address not only current AI issues, but also future flash points.
Open banking
The CFPB has been a political lightning rod since its creation following the 2008 global financial crisis. Democrats have touted the agency as a necessary check on corporate power and a guardian of ordinary Americans with whom its popularity may grow if the U.S. economy goes into a recession. But it has also been reviled by some Republicans and business groups who claim the agency lacks accountability.
In addition to AI, Chopra also said the CFPB was getting ready to release new requirements on how financial firms should share client data, a topic known as open banking. He said the goal of that effort is to make it easier for Americans to share personal financial information held by banks with third parties, including financial-technology firms.
As it writes new rules to be proposed in October, Chopra said the CFPB will take into account standards that parts of the industry already have set. “Our role will help, in many ways, to operationalize those standards,” he said. “But if the standards are hijacked by any one set of actors, if it’s too skewed, the whole thing falls apart.”
(With assistance from Lydia Beyoud and Anna Edgerton.)
