Earnest Operations LLC · Governance finding · severity 3 of 5
On 10 July 2025 Massachusetts Attorney General Andrea Joy Campbell announced a $2.5 million settlement with student-loan lender Earnest Operations, resolving allegations that its lending practices, including its use of AI underwriting models, risked disparate harm to Black, Hispanic and non-citizen applicants. According to the AG, Earnest's model used a "Cohort Default Rate" variable that penalised Black and Hispanic applicants more than White applicants, a "Knockout Rule" automatically denied applicants without at least a green card, and Earnest did not test its models for disparate impact. Under the assurance of discontinuance, Earnest denies the allegations, pays $2.5 million, drops both practices and must build an AI governance and fair-lending testing program and report on compliance.
Record last reviewed Oct 21, 2026 · Markdown
Date
Institution
Earnest Operations LLC
Country
United States
AI system
Algorithmic and AI underwriting models scoring student-loan and personal-loan applications for approval and pricing, with automated knockout rules
AI?
Described as AI by the institution or regulator
The institution, a regulator or a court called the system AI.
Category
Status
Settled
Severity
3 of 5
Impact
$2.5 million paid to the Commonwealth of Massachusetts · ≈ $2.5 million · Not disclosed; the AG said Massachusetts applicants across all three products were declined for citizenship reasons · The AOD does not state how many applicants were affected by the Cohort Default Rate variable or the knockout rule, and no restitution amount to individual borrowers is specified; the AG may use the payment for consumers or other purposes permitted by law.
What happened?
Earnest, a Delaware company formed in 2013, originates and refinances private student loans and made personal loans from 2014 to 2020. According to the attorney general's office, Earnest uses artificial intelligence models, specifically algorithmic models, to decide applicants' eligibility, loan terms and pricing. Its underwriting ran in three stages (prescreen decline, quick decline and risk score) and combined model scores with knockout rules and human judgment, according to the assurance of discontinuance (AOD).[1][2]
The AG alleged that Earnest's student-loan refinance model included the U.S. Department of Education's Cohort Default Rate for the applicant's school as a weighted input until 13 September 2017, and that this "resulted in disparate impact in approval rates and loan terms", with Black and Hispanic applicants more likely to be penalised than White applicants. Until 30 June 2023, Earnest automatically declined any applicant without at least a green card at the prescreen stage, without assessing creditworthiness.[1][2]
The AG also alleged that underwriters could override model recommendations without written standards until 2018, that Earnest trained models on "arbitrary, discretionary human selections" and designed its newer student-loan model partly to replicate those decisions, that it never tested its models for disparate impact, and that its adverse action notices sometimes gave inaccurate decline reasons because the model's actual reasons were missing from the drop-down menu staff used, in violation of ECOA and Regulation B.[2]
The settlement was filed as an assurance of discontinuance in Suffolk County Superior Court. Ballard Spahr's Consumer Finance Monitor and the ABA Banking Journal summarised the allegations and the AI governance obligations.[1][3][4]
When did it happen?
- Sep 13, 2017
Earnest stops using the Cohort Default Rate as a weighted input in its student-loan refinance model, per the AG's allegations.[2]
- Jun 30, 2023
Earnest ends its knockout rule automatically declining applicants without at least a green card, per the AG's allegations.[2]
- Jul 10, 2025
AG Andrea Joy Campbell announces the $2.5 million settlement with Earnest, filed as an assurance of discontinuance in Suffolk County Superior Court.[1]
What was the impact?
- Financial: $2.5 million paid to the Commonwealth of Massachusetts · ≈ $2.5 million[1][2]
- Customers: Not disclosed; the AG said Massachusetts applicants across all three products were declined for citizenship reasons[1][2]
- Other: The AOD does not state how many applicants were affected by the Cohort Default Rate variable or the knockout rule, and no restitution amount to individual borrowers is specified; the AG may use the payment for consumers or other purposes permitted by law.[1][2]
Severity 3 of 5: Rubric 3 (regulatory finding): a $2.5m state settlement over alleged disparate impact and automatic denials in AI-assisted underwriting, with fair-lending testing and model-governance obligations; the number of affected applicants was not disclosed.
How did Earnest Operations LLC respond?
In the assurance of discontinuance, "Earnest denies the AGO's allegations and further denies that it has violated Massachusetts or federal law" and says it agreed to the terms to avoid the uncertainty of litigation. No other public statement from Earnest was found as of 2026-10-21 (searched news coverage and the AG's release).[2]
What was the outcome, and did regulators act?
Earnest must pay $2.5 million within 30 days, discontinue the Cohort Default Rate variable and the immigration-status knockout rule, and build a written governance system for its AI models: fair-lending testing at least annually and on trigger events, a model inventory, documentation and data retention, an internal algorithmic oversight team, a channel for staff to report algorithmic bias, and a Model Risk Management policy. It must report regularly to the attorney general's office on compliance.[1][2]
Did the Massachusetts attorney general say Earnest's AI was biased?
The AG alleged that Earnest's AI underwriting models could lead to disparate harm to Black, Hispanic and non-citizen applicants, citing a Cohort Default Rate variable that penalised Black and Hispanic applicants more than White applicants and an automatic denial rule based on immigration status. Earnest denies the allegations; the settlement is not an admission.
Was Earnest's underwriting system actually AI?
The attorney general's office said Earnest uses artificial intelligence models, specifically algorithmic models, to decide eligibility, terms and pricing. The settlement defines AI models to include machine-learning algorithms used for automated underwriting decisions, while the specific variable at issue sat in an earlier scorecard model built on human-selected variables and weights.
Were borrowers compensated in the Earnest settlement?
Earnest paid $2.5 million to the Commonwealth of Massachusetts. The assurance of discontinuance lets the attorney general use the money for payments to consumers or other lawful purposes, but it does not set restitution for individual applicants, and the number of people affected has not been disclosed.
What did Earnest have to change?
Earnest must stop using the Cohort Default Rate variable and the immigration-status knockout rule, test its underwriting models for fair-lending problems at least annually, keep an inventory and documentation of its models, set up an internal algorithmic oversight team and a way for staff to report bias, and report on compliance to the attorney general.
- AG Campbell Announces $2.5 Million Settlement With Student Loan Lender For Unlawful Practices Through AI Use, Other Consumer Protection Violations official
- Assurance of Discontinuance, In the matter of Earnest Operations LLC (Suffolk Superior Court) official
- Massachusetts AG reaches settlement with student loan company, Earnest Operations LLC news
- Mass. AG reaches settlement with student loan firm for $2.5M over AI lending bias news
When AI goes wrong at a bank, the next morning's brief says so.
when one of these regulators moves, the next morning's brief says so · six sourced stories · 7 am PT · free
plus every regulation, bank and agent page update, the morning after · leave any morning