Regulation — housing-specific

ECOA + Regulation B — protected-class analysis + adverse-action notice for AI credit decisions

The Equal Credit Opportunity Act prohibits credit discrimination on protected-class grounds. Regulation B is the operational rulebook. Together they impose two separate duties on AI-driven mortgage decisions: no protected-class discrimination in outcomes, and accurate adverse-action notices when the AI says no.

What the regulation actually says

"It shall be unlawful for any creditor to discriminate against any applicant, with respect to any aspect of a credit transaction — (1) on the basis of race, color, religion, national origin, sex or marital status, or age …" 15 U.S.C. §1691(a) — the underlying statute

"A creditor shall notify an applicant of action taken … (2) The statement of reasons for adverse action … must be specific and indicate the principal reason(s) for the adverse action. Statements that the adverse action was based on the creditor's internal standards or policies or that the applicant, joint applicant, or similar party failed to achieve a qualifying score on the creditor's credit scoring system are insufficient." 12 CFR 1002.9(b)(2) — Regulation B, adverse-action notice specificity

What this means in plain English

Two obligations for AI underwriting:

No discrimination on protected-class grounds. Age, race, color, religion, national origin, sex, marital status. Applies to outcomes, not only intent.
Adverse-action notices must state specific reasons. "The model scored you below threshold" is not a valid reason. Reason codes must map to real, applicant-visible features of the decision.

What triggers the exposure in the sample

H3 decline / non-QM lane rate rose 14% → 61% silently. Age (30-50) is a protected class under ECOA. The near-prime + majority-minority zip combination adds race-proxy reach. Every H3 adverse-action notice issued during the drift window sits inside ECOA reason-code accuracy review: was the stated reason the actual reason the AI declined the file, or was it the model's baked-in group offset?

What the $499 Snapshot shows against this rule

Dated per-group differential in lane routing — the protected-class effects record
Model version pinning + decision-hash binding — the reason-code traceability record
Counterparty-question rehearsal on 5 sample decisions — the reason-code accuracy check
Independent-verifier declaration — the "reasonable person could have known" record

See the 5-decision reason-code drill →

How does this help me?

Reason-code failure is the fastest path to per-application civil money penalty under Reg B. Dated evidence changes the CFPB stance from "possible pattern-or-practice" to "documented isolated events."

Read: ECOA + Reg B -- what it saves you at CFPB exam or DOJ referral →

$499 Snapshot. 3 business days.

Independent-verifier determination scoped to your lender's AI surface + 3 fix-first items + signed declaration.

Buy $499

Snapshot credit applies to Baseline ($2,500) or Enterprise Attestation ($35-55K) upgrade within 30 days.