By Katanga Johnson | Bloomberg
Home mortgage lenders will have to meet a new set of standards designed to ensure that their tech-fueled appraisals don’t build in flaws that might produce faulty estimates or discriminate against borrowers.
The Federal Reserve and five other federal agencies announced the new demands in a final rule on Wednesday, saying they’re intended to “help ensure the credibility and integrity” of models used in home valuations for certain mortgages.
The regulation will introduce quality-control standards for automated-valuation models used by mortgage originators and secondary-market issuers, the agencies said in a statement.
The rule is very similar to the June 2023 proposal and was mandated as part of the Dodd-Frank Act that followed the 2008 global financial crisis. Regulators under the Biden administration have stepped up efforts to address potential abuses as artificial intelligence and other technologies become more prevalent in lending.
In January, three top US regulators warned banks that they can’t escape responsibility if their outsourced artificial-intelligence tools violate consumer-protection and fair-lending laws. Automated valuation models can rely on machine learning or artificial intelligence.
Lenders are increasingly using automated valuation models to gauge what a property is worth, according to the agencies’ statement Wednesday. Although this can help cut costs and turnaround times, lenders need to make sure that the valuations are accurate and don’t result in discrimination, the regulators said.
Consumer groups and regulators have complained for decades that figures used by many home lenders to make decisions are biased against credit-worthy minorities because of flaws embedded in commercial databases.
The Federal Deposit Insurance Corp., the Consumer Financial Protection Bureau, the Federal Housing Finance Agency, the Office of the Comptroller of the Currency and the National Credit Union Administration also announced Wednesday’s rule.
Originally Published:
