{
  "$schema": "https://aigovernanceengineer.com/schemas/impact-assessment.v1.json",
  "assessment_id": "fria-2026-004",
  "type": "fria",
  "subject": "credit-afford-03@3.2.0",
  "assessor": "ai-governance-lead",
  "consulted": [
    "privacy-office",
    "legal",
    "customer-care",
    "consumer association liaison"
  ],
  "related_assessments": [
    "dpia-2026-011"
  ],
  "processes": "Consumer loan origination: the score supports the analyst's affordability decision for applications up to 30k EUR.",
  "period_and_frequency": "Continuous from go-live, about 1,500 applications a day, reviewed yearly.",
  "affected_categories": [
    "consumer loan applicants",
    "applicants over 60",
    "self-employed applicants"
  ],
  "risks": [
    {
      "right_or_interest": "non-discrimination",
      "description": "Scores that lean on age-correlated features decline older applicants more often.",
      "affected_group": "applicants over 60",
      "likelihood": 2,
      "severity": 4,
      "risk_register_id": "rk-0042"
    },
    {
      "right_or_interest": "access to credit and an effective remedy",
      "description": "An applicant cannot understand or contest a decline influenced by the score.",
      "affected_group": "declined applicants",
      "likelihood": 2,
      "severity": 3
    }
  ],
  "human_oversight_measures": "Analysts decide every application with the score, reason codes and data shown; overrides are logged and sampled monthly.",
  "mitigations": [
    {
      "measure": "Fairness eval gate on age bands, threshold 0.80",
      "addresses": [
        "non-discrimination"
      ],
      "owner": "model-validation",
      "status": "verified",
      "evidence": "https://evidence.example.org/test-reports/tr-credit-afford-03-3.2.0"
    },
    {
      "measure": "Reason codes in the decline letter and a human review on request",
      "addresses": [
        "access to credit and an effective remedy"
      ],
      "owner": "retail-credit-operations",
      "status": "in_place"
    }
  ],
  "governance_and_complaints": "Complaints about a decision reach customer care and, if they allege discrimination, open an AI incident; the AI governance committee reviews them quarterly.",
  "residual_risk": "Low, with the weekly fairness monitoring in the PMM plan.",
  "outcome": "proceed_with_mitigations",
  "authority_notification": {
    "required": true,
    "authority": "national market surveillance authority"
  },
  "approvals": [
    {
      "role": "head-of-retail-credit",
      "decision": "approve",
      "timestamp": "2026-08-19T16:00:00Z"
    }
  ],
  "review_triggers": [
    "new product line",
    "new country",
    "fairness ratio below threshold",
    "incident alleging discrimination"
  ],
  "next_review": "2027-08-19"
}