{
  "node_id": "us-cfpb-circular-2023-03-adverse-action-ai-creditors",
  "title": "CFPB Consumer Financial Protection Circular 2023-03 - Adverse Action Notices When Using AI and Complex Credit Models",
  "domain": "AI Governance & Law",
  "version": "1.0.0",
  "last_updated": "2026-05-14",
  "bluf": "CFPB Consumer Financial Protection Circular 2023-03 (19 September 2023) builds on the 2022-03 Circular and clarifies that creditors using AI and other complex algorithmic models in credit decisions cannot satisfy the Equal Credit Opportunity Act and Regulation B adverse action notice requirements by relying on a closed list of generic checkbox reasons that do not reflect the actual factors driving the denial. The Circular specifies that creditors must provide accurate and specific reasons, even when those reasons fall outside the standard checklists referenced in the Regulation B model adverse action notice forms. Examples cited include where AI models consider unconventional factors (consumer spending patterns, social media activity, behavioural data, third-party data), the creditor must disclose those factors specifically rather than substitute a generic checkbox. The Circular reinforces that adverse action specificity is a strict legal requirement that does not bend for model complexity, third-party vendor opacity, or proprietary system claims. Enforcement authority is the CFPB and parallel regulators under ECOA.",
  "paywall": {
    "status": "LOCKED",
    "unlock_cost_usd": "0.01",
    "skyfire_id": "41779894-ece2-4163-9761-b3b1b76e19b0"
  },
  "crosswalks": {
    "_available_keys": [
      "industry_mapping",
      "related_frameworks"
    ],
    "_note": "Full crosswalk values included in vault response"
  },
  "dependencies": [
    "us-cfpb-circular-2022-03-adverse-action-complex-algorithms",
    "frb-sr-11-7-model-risk-management"
  ],
  "primary_citations_count": 7
}