AI Denied My Credit or Insurance in the US. What Are My Rights?

AI drives most US credit and insurance decisions. When it denies you or charges you more, federal and state laws give you specific rights to know why, challenge the decision, and seek remedies for algorithmic discrimination.

Key Takeaways

ECOA and the Fair Housing Act prohibit credit and insurance discrimination based on protected characteristics — and apply to AI-driven decisions in exactly the same way as human decisions.
When AI denies you credit or offers less favourable terms, ECOA requires adverse action notices stating specific reasons. The CFPB has made clear that 'our model' is not specific enough — actual factors must be disclosed.
The Fair Credit Reporting Act gives you the right to a free copy of your credit report and to dispute inaccurate information. If AI credit decisions are based on errors in your credit report, correcting those errors is often the fastest path to resolution.
The CFPB has brought enforcement actions against lenders for discriminatory AI models and inadequate adverse action notices. Complaints at consumerfinance.gov/complaint are taken seriously.
Colorado SB 21-169 (effective 2023) was the first US state law specifically addressing AI in insurance — prohibiting use of external consumer data that results in unfair discrimination. Several states have similar proposals pending.
Complaint routes: CFPB for credit, HUD for mortgage discrimination, state insurance commissioner for insurance, or a fair lending attorney who works on contingency.

"仅供参考。本文不构成法律、监管、财务或专业建议。如需具体指导，请咨询合格专家。"

How AI is used in US credit and insurance decisions — and what protects you

AI is now deeply embedded in US credit and insurance decision-making. Mortgage underwriting, credit card approvals, auto loan pricing, personal loan decisions, insurance underwriting, claims processing, and fraud detection all rely on machine learning models. The "black box" defence — that the institution does not know exactly how the AI reached its decision — is not a defence. Under US federal law, you have specific rights when AI plays a role in credit or insurance decisions about you, and these rights apply regardless of the technology used.

The Equal Credit Opportunity Act (ECOA) — your right to specific reasons

ECOA, implemented through Regulation B, is the foundational federal consumer protection law for credit decisions. ECOA prohibits creditors from discriminating against any applicant on the basis of race, colour, religion, national origin, sex, marital status, age, receipt of public assistance, or exercise of rights under the Consumer Credit Protection Act. The Act applies to all credit transactions — origination, servicing, debt collection, credit line changes, and account terminations.

When a creditor takes an "adverse action" against you — denying credit, denying a credit limit increase, increasing your interest rate adversely, terminating an existing credit arrangement — you have the right to a written statement of specific reasons. The CFPB has been unambiguous: under ECOA and Regulation B, generic adverse action checklist items are not sufficient when AI is used. The reasons must be specific, accurate, and reflect the principal factors actually considered by the model.

The CFPB's September 2023 Consumer Financial Protection Circular and its earlier May 2022 Circular together establish that creditors using AI must explain the specific reasons for adverse actions. If a creditor reduces your credit line based on behavioural spending data, the explanation must provide details about the specific negative behaviours that led to the reduction — not just a generic reference to "purchasing history." If a creditor's AI model is too opaque to produce specific accurate reasons, the creditor cannot lawfully use it for ECOA-covered decisions. Generic notices that hide the actual reason are violations.

The Fair Credit Reporting Act (FCRA) — additional adverse action protections

FCRA, implemented through Regulation V, provides additional adverse action notification requirements. FCRA's adverse action notice provisions apply not just to credit but to insurance, employment, and housing decisions made on the basis of information about your creditworthiness. When adverse action is based wholly or partly on a credit score obtained from a consumer reporting agency, you must receive disclosure of: the key factors that adversely affected the score; the name and contact information of the consumer reporting agency; and additional content specified by the regulation. FCRA's reach is broader than ECOA — it covers insurance underwriting, employment background checks using credit information, and rental application decisions.

Disparate impact liability — why AI itself can be the discrimination

One of the most consequential principles in current US fair lending law is that the decision to use AI itself can create disparate impact liability under civil rights statutes. The CFPB has explicitly stated, citing court decisions including Huskey v. State Farm Fire & Casualty Co., that courts have held that the choice to use algorithmic or machine-learning tools can itself constitute a policy producing bias under disparate impact theory. This applies to ECOA compliance and the Consumer Financial Protection Act.

The practical implication is significant: if an AI underwriting model produces measurably worse outcomes for applicants in a protected class (Black applicants, Hispanic applicants, women, applicants in protected age categories), the creditor cannot defend the disparate impact by saying "the algorithm is neutral. The choice to use the algorithm is itself the policy that produced the disparate impact. To defend it, the creditor must show the model serves a legitimate business need and no less discriminatory alternative was reasonably available. The CFPB expects ongoing testing for disparate treatment and disparate impact" and consider "less discriminatory alternatives" using both manual and automated techniques.

Recent enforcement — Massachusetts AG fair lending settlement

On 10 July 2025, Massachusetts Attorney General Andrea Joy Campbell announced a settlement with a student loan company over allegations its AI underwriting models resulted in unlawful disparate impact based on race and immigration status. The Massachusetts AG's investigation reviewed the lender's algorithmic rules, judgmental discretion in loan approval, and internal communications. The scoring model used (prior to 2017) included a Cohort Default Rate — the average rate of loan defaults associated with specific higher education institutions — which the AG asserted produced disparate impact disfavouring Black and Hispanic applicants in violation of ECOA. The model also included immigration status (knocking out applicants without a green card), which the AG said created risk of disparate outcome on the basis of national origin. The settlement prohibits both factors going forward.

This settlement is significant because state-level enforcement is filling the space left by reduced federal fair lending enforcement priorities. Even where federal regulators step back, state attorneys general — particularly in Massachusetts, New York, California, and Colorado — are actively enforcing fair lending laws against algorithmic discrimination.

Colorado AI Act — the state-level addition to credit and insurance rights

The Colorado Artificial Intelligence Act (effective in February 2026) is a major state-level addition to consumer rights. It applies specifically to AI systems used to make "consequential decisions" — explicitly including financial services and insurance. The Act requires AI developers and deployers to use reasonable care to protect against algorithmic discrimination, conduct annual impact assessments for high-risk AI systems, provide consumers with notice when an AI system is used to make a consequential decision affecting them, and explain the basis of the decision and what data was used. The Act creates enforcement powers for the Colorado Attorney General. While the Act has been subject to some delay debates, the operative framework remains the most comprehensive state-level AI consumer protection regime currently in force.

Insurance specifically — NAIC Model Bulletin and state action

For insurance, the National Association of Insurance Commissioners (NAIC) Model Bulletin on the Use of Artificial Intelligence Systems by Insurers (December 2023) has been adopted in some form by over 20 US states as of 2026. The bulletin sets expectations for insurers' AI governance programmes, including risk management, vendor oversight, testing and validation, transparency, and accountability. Adopting states include Connecticut, New York, Pennsylvania, Texas, California, Maryland, Illinois, and others. State insurance regulators are increasingly examining insurers' AI use in underwriting, pricing, claims processing, and fraud detection — with discrimination risk as a primary enforcement focus.

What to do if you believe AI made a wrong decision about your credit or insurance

Request your adverse action notice in writing and read it carefully. Generic, non-specific reasons indicate a potential ECOA violation — the creditor must provide specific principal reasons. Request your credit report from all three nationwide consumer reporting agencies (free annual reports via annualcreditreport.com). Inaccuracies in your credit file may be the underlying cause of an algorithmic adverse decision; the FCRA gives you the right to dispute inaccurate information.

If you believe the decision was discriminatory or the explanation insufficient, file a complaint with the CFPB at consumerfinance.gov. Complaints are reviewed and can prompt enforcement action. For insurance-specific issues, file with your state insurance commissioner. For state-law violations, complaints to your state attorney general can be effective — Massachusetts, New York, California, Colorado, and Illinois have been most active. Consider consulting a fair lending or consumer rights attorney, particularly if you suspect disparate impact discrimination — disparate impact cases often involve patterns across many applicants, and class action representation may be available.