A class action alleged Wells Fargo's ML credit scoring routed minority applicants to worse tiers
A consolidated class-action lawsuit (In re Wells Fargo Mortgage Discrimination Litigation, Case 3:22-cv-00990) alleged that Wells Fargo's Enhanced Credit Score system, identified by a plaintiffs' expert as a supervised machine learning model, systematically assigned Black, Hispanic, and Asian mortgage applicants to higher-risk credit tiers, resulting in disproportionate denials and less favorable loan terms compared to white applicants. The plaintiffs sought to represent a class of approximately 119,100 minority borrowers who applied for mortgages between 2018 and 2022. A federal judge denied class certification in August 2025, though individual claims may still proceed.
A supervised machine learning model for credit scoring encoded historical racial disparities into automated risk tiers, turning past bias into present-day denials.
Key facts
- What
- A consolidated class-action lawsuit (In re Wells Fargo Mortgage Discrimination Litigation, Case 3:22-cv-00990) alleged that Wells Fargo's Enhanced Credit Score system, identified by a plaintiffs' expert as a supervised machine learning model, systematically assigned Black, Hispanic, and Asian mortgage applicants to higher-risk credit tiers, resulting in disproportionate denials and less favorable loan terms compared to white applicants.
- Incident date
- May 1, 2024
- Who
- Wells Fargo
- Failure mode
- Policy Violation
- AI surface
- Agentic Workflow
- Severity
- High
What happened
Plaintiffs filed a consolidated class-action lawsuit in February 2022 alleging that Wells Fargo's automated mortgage underwriting system, including its Enhanced Credit Score model and CORE platform, engaged in digital redlining against non-White applicants between 2018 and 2022. Bloomberg data showed Wells Fargo approved only 47% of Black homeowners' refinance applications and 53% of Hispanic applicants, compared to substantially higher rates for white applicants, making it the only major bank to deny more Black applicants than it accepted in 2020. The plaintiffs moved for class certification on April 25, 2024, seeking to represent 119,100 minority borrowers, but Judge James Donato denied certification on August 5, 2025, ruling the plaintiffs failed to establish commonality among the individual lending decisions. Individual plaintiffs may still pursue separate claims.
What broke inside the model
- 01 · TriggerA prompt pushes against a deployment boundary.
- 02 · Model stepThe model produces the disallowed output.
- 03 · Control gapNo enforcement blocks it at generation time.
- 04 · FailureThe output crosses the policy line.
- 05 · ConsequenceA limit the business set is breached in public.
The output crosses a policy boundary the deployment had defined.
Wells Fargo's Enhanced Credit Score supervised machine learning model disproportionately routed minority applicants into higher-risk credit tiers by relying on features such as average months in file, recent inquiries, and major derogatories that encoded historical racial disparities. The plaintiffs' expert testified that the model was capable of algorithmic bias, and it funneled minority applicants into heightened underwriting scrutiny even when external systems from Fannie Mae and Freddie Mac had initially approved them. The automated pipeline lacked sufficient safeguards to detect and correct the disparate impact across racial groups.
What it cost
Sources
- Court FilingIn Re Wells Fargo Mortgage Discrimination Litigation Consolidated Amended Complaintbencrump.com
- PressMortgage underwriting algorithm at heart of Wells Fargo's digital redlining casefinance.yahoo.com
- PressWells Fargo won't face class-action mortgage discrimination lawsuitbankingdive.com
Cite this entry
https://failureindex.ai/failures/class-action-alleged-wells-fargo-mlAI Failure Index. "A class action alleged Wells Fargo's ML credit scoring routed minority applicants to worse tiers" (FI-0086). Realm Labs. https://failureindex.ai/failures/class-action-alleged-wells-fargo-ml (indexed Jun 4, 2026).Data fields CC-BY 4.0, prose citation permitted. Incident ID FI-0086. Full dataset at /data.
Note from Realm Labs, the Index steward
How Realm would have caught this
- Prism
- OmniGuard
Realm compares what the model is about to output or do against the policy that governs the deployment, in real time, and can deny or redact the action before it takes effect, which is the gap an after-the-fact review never closes in time.