On June 1, 2023, the Consumer Financial Protection Bureau (CFPB) announced a new measure aimed at increasing accountability for automated systems and models, often referred to as artificial intelligence (AI). The CFPB is introducing a proposed rule aimed at enhancing the fairness and accuracy of algorithm-driven home appraisals. This action is part of broader efforts to ensure that AI and algorithms adhere to established legal standards.
Working alongside federal entities such as the Federal Reserve, FDIC, FHFA, NCUA, and the Office of the Comptroller of the Currency, the CFPB’s proposed rule seeks to guarantee that automated home valuation processes are equitable and free from discrimination. Although not a complete solution, this proposed regulation acknowledges the potential risks of algorithmic bias in property appraisals. It aligns with the objectives of the Interagency Task Force on Property Appraisal and Valuation Equity (PAVE) to foster accurate property valuations, challenging those derived from biased computational models.
Automated valuation models (AVMs) present a technological approach to property appraisal, contrasting with traditional methods that rely on human assessment. Despite the perception that these models remove human bias through mathematical algorithms, they can inadvertently perpetuate existing prejudices. The data input and algorithm design may reflect the biases of their creators, making bias more challenging to identify and address under the guise of objectivity.
The consequences of inaccurate or biased algorithmic appraisals are significant, potentially leading to overvaluation, risking foreclosure, or undervaluation, limiting homeowners’ access to equity. These systemic biases can harm not only individual homeowners but also entire communities by distorting market dynamics and affecting local tax revenues. Fair and unbiased property appraisals are essential for all involved in the home buying and selling process.
The proposed rule aims to establish basic precautions to mitigate the risks associated with AVMs. It would require institutions using these models for property valuations to implement measures that ensure the reliability of their estimates and guard against data manipulation. Additionally, it mandates that companies maintain policies to prevent conflicts of interest, conduct random testing, and adhere to nondiscrimination laws.
This initiative is part of the CFPB’s ongoing efforts to address the challenges posed by AI technologies that risk discriminatory impacts and threaten financial stability. For instance, the CFPB has issued guidelines emphasizing that lenders must understand how their AI models function to comply with federal laws. The bureau has also collaborated with other agencies to underscore the importance of developing and utilizing AI systems in accordance with legal standards, highlighting the potential dangers of misuse.
As AI technologies evolve, they pose risks to civil rights, fair competition, and consumer protection. It is crucial that these technologies abide by existing laws to prevent legal evasion and consumer harm. The CFPB remains committed to monitoring the use of automated systems and enforcing consumer financial protection laws, ensuring the rights of homeowners, buyers, and sellers are protected, whether the issues arise from human or machine actions.
Check the original article on this link.
Comments