I don’t check every state Attorney General website every day, but I am fairly confident in saying that this is one of the first consumer protection enforcement actions from a state related to the use of artificial intelligence. Massachusetts Attorney General Andrea Joy Campbell has announced a $2.5 million settlement with student loan provider Earnest Operations, marking a significant enforcement action at the intersection of AI, fair lending, and consumer protection.
The settlement follows an investigation by the AG’s office that found Earnest used algorithmic underwriting and artificial intelligence models in ways that resulted in unlawful disparate impacts on Black, Hispanic, and non-citizen borrowers. The company was also accused of failing to adequately test its AI models for bias, using variables like “Cohort Default Rate” tied to specific educational institutions, and applying “Knockout Rules” that automatically denied applicants based on immigration status without evaluating their creditworthiness.
Campbell further alleged that Earnest issued vague or incorrect adverse action notices, preventing consumers from understanding why they were denied credit, and that the company lacked adequate governance policies to detect or prevent discrimination in its AI-based lending practices.
As part of the agreement, filed as an Assurance of Discontinuance in Suffolk County Superior Court, Earnest will:
- Pay $2.5 million to the Commonwealth.
- Cease using the “Cohort Default Rate” and the “Knockout Rule” for immigration status in underwriting decisions.
- Implement robust AI governance policies, including documentation, oversight teams, annual fair lending testing, and a whistleblower mechanism for reporting algorithmic bias.
- Provide ongoing compliance reporting to the AGO for at least three years.
The enforcement action noted, “Earnest failed to test its Models for disparate impact… and designed the SLO Model, in part or in whole, to replicate the arbitrary and unfair and deceptive SLR applicant decisions.” It further concluded that “Earnest knew or should have known that these acts were unfair or deceptive.”
Earnest has not admitted to any wrongdoing but has agreed to the terms to avoid litigation.
This case signals a new era of regulatory scrutiny, where the use of AI in financial services—especially when it affects marginalized communities—will not go unchecked. As AG Campbell put it: “This settlement will put lenders on notice that Massachusetts will not tolerate unlawful practices that harm our consumers.”
.




