← the record
AIAAIC-0876

Upstart automated consumer lending discrimination

A loan decision model deployed by San Mateo-based AI consumer lender Upstart produced 'significant disparities' in how often loans were made to Black and non-Hispanic white borrowers. Law firm Relman Colfax discovered that Upstart assesses and predicts creditworthiness using so-called 'Alternative Data', including a person's educational and employment history. The AI lender claims its algorithmic lending model approved 30 percent more Black borrowers than traditional models in 2020, and that it's approach is more inclusive than the underwriting methods used by many banks. However, academic research published in January 2022 indicated discriminatory pricing exists in both traditional and fintech lending. In July 2022, Upstart directors were hit by a class-action investor lawsuit claiming it made false and misleading statements about its business, operations and prospects. System 🤖 Upstart 🔗 Operator: Upstart Holdings; Multiple Developer: Upstart Holdings Country: USA Sector: Banking/Financ ial services Purpose: Assess creditworthiness Technology: Machine learning Issue: Bias/discrimination Legal, regulatory 👩🏼‍⚖️ https://www.businesswire.com/news/home/20220513005551/en/INVESTOR-ALERT-Upstart-Holdings-Inc.-Investors-with-Substantial-Losses-Have-Opportunity-to-Lead-the-Upstart-Class-Action-Lawsuit---UPST https://financialservices.house.gov/uploadedfiles/hhrg-117-ba00-wstate-girouardd-20210507.pdf I nvestigations, assessments, audits 🧐 https://www.relmanlaw.com/media/news/1332_PUBLIC_Upstart_Monitorship_3rd_Report_FINAL.pdf https://www.relmanlaw.com/media/news/1182_PUBLIC%20Upstart%20Monitorship_2nd%20Report_FINAL.pdf https://www.relmanlaw.com/media/cases/1086_Upstart%20Initial%20Report%20-%20Final.pdf

Date it happened
2022-09-01
Organisation involved
Upstart Holdings; Multiple
Product, system or model
Upstart
Where this came from
Share this incident
XLinkedInFacebookWhatsAppEmail
Attribution

This incident was imported from AIAAIC and is used under CC BY-SA 4.0. Our additions to it — the structured fields, the translation, the checks against other reports — are published under the same licence.

This is a record of what was reported, not a finding that anyone broke the law. If it names your organisation and you believe it is wrong, the corrections process is free and open to everyone.