The Consumer Financial Protection Bureau (CFPB) this week announced the approval of a new rule that aims to govern “current and future applications of complex algorithms and artificial intelligence used to estimate the value of a home.” The rule specifically targets automated valuation models (AVMs), according to an announcement from CFPB Director Rohit Chopra and deputy director Zixta Martinez.
The importance of home appraisals to the homebuying and selling processes necessitate a stronger eye toward the use of burgeoning AI across a wide swath of U.S. businesses. The bureau expressed concern that the use of such tech could impact the equitable application of home appraisals.
“Over the years, players across the real estate and mortgage industry have made use of computer models to estimate a property’s value,” Chopra and Martinez wrote. “On popular real estate websites, many people even track their own home’s value generated from these algorithmic appraisal tools. As these models grow in complexity to incorporate more and more variables, they can resemble what many people often refer to as artificial intelligence.”
AI technology has demonstrated that it is not immune to biases that can impact its output, and the importance of that fact grows significantly when it is applied to something as important as an appraisal, they said.
“While these computer models can provide critical insight for buyers, sellers, and lenders, they cannot be inaccurate or discriminatory,” they said. “It can be tempting to think that computer models can take bias out of the equation, but they can’t.”
Building on previous work the CFPB has undertaken to allow more consumers to challenge appraisals they believe to be inaccurate — as well as its recent criticism of the structure and transparency of The Appraisal Foundation — the rule takes aim at AVMs and will impose stricter quality control standards for their use.
“Under the final rule, institutions that engage in certain credit decisions or securitization determinations must adopt policies, practices, procedures, and control systems to ensure that AVMs used in these transactions to determine the value of mortgage collateral adhere to quality control standards designed to ensure a high level of confidence in the estimates produced by AVMs,” the rule states.
Additionally, institutions must “protect against the manipulation of data; seek to avoid conflicts of interest; require random sample testing and reviews; and comply with applicable nondiscrimination laws.”
Beyond the CFPB, other agencies involved with drafting the final rule include the Federal Deposit Insurance Corp. (FDIC), the National Credit Union Administration (NCUA) and the Federal Housing Finance Agency (FHFA).
