In December 2020, Upstart Network (“Upstart”), the NAACP Legal Defense Fund (“LDF”), and the Student Borrower Protection Center (“SBPC”) entered into an agreement to appoint Relman Colfax to serve as an independent fair lending Monitor to evaluate and make recommendations regarding the fair lending implications of Upstart’s lending platform, and to issue a series of periodic reports on its findings and recommendations. Those reports are published below, upon release.
Upstart’s lending platform relies on Machine Learning-based Artificial Intelligence (“ML” and “AI”) models and non-traditional applicant data—including data related to borrowers’ higher education—to underwrite and price consumer loans. LDF and the SBPC raised concerns with Upstart that the use of educational criteria can lead to discriminatory lending outcomes, particularly for communities of color, leading to the appointment of an independent fair lending Monitor.
In April 2021, we issued an Initial Report, which provides a summary of legal principles and fair lending testing, and a descriptive history of the events leading up to the Monitorship.
On November 10, 2021, we issued a public Second Report, providing further detail regarding the methodology and fair lending tests conducted to date.
The Third Report, issued in September 2022, explains application of those tests to a recent version of Upstart’s Model, including identifying what would likely have been a viable less discriminatory alternative model. Before the analyses were completed, Upstart updated its Model. Accordingly, instead of recommending adoption of that specific less discriminatory alternative model, the Report recommends that Upstart apply the methodologies to its existing algorithms and resulting Model or to any imminent upcoming model updates, and to future model updates. If a viable less discriminatory alternative model is identified, the Report recommends that Upstart adopt that alternative. The Third Report separately does not find quantitative evidence that variables in Upstart’s model are functioning as close proxies for certain protected characteristics. At the same time, the Report provides recommendations to mitigate future potential age-proxy risks specifically related to non-traditional variables. A Confidential version of each Report has been provided to the parties. Public versions of each report are available below.
In our capacity as Monitor we have engaged Sentrana to serve as a consultant. Sentrana is a leading firm in the field of machine learning and artificial intelligence. We have also engaged Dr. Bernard Siskin of BLDS. Dr. Siskin is an expert on the use of statistical analyses to measure discrimination in the financial services industry.