{
  "id": "e9b271d4-8521-4f1a-9694-81d3d6e5a401",
  "rmf_schema_version": "1.0.0",
  "generated_at": "2026-06-15T11:00:00Z",
  "signed_by": "Compliance Lead, AiExponent",
  "sha256_hash": "4e9a3b8d1f2c6e7a0b5d8f3e2a1c9b8d7e6f5a4b3c2d1e0f9a8b7c6d5e4f3a2b",
  "audit_entry_hash": "a7c8e9f0123456789abcdef0123456789abcdef0123456789abcdef012345678",
  "disclosure": "This document was produced using RiskForge v1.1.3, question bank version 1.0.0. It represents the team's documented risk assessment and has not been reviewed by a qualified legal professional. It does not constitute legal advice under the EU AI Act or any other regulation.",
  "register": {
    "system": {
      "id": "sys_credit_eval_01",
      "name": "Consumer Credit Eligibility Screener",
      "version": "1.0.0",
      "category": "essential_services",
      "annex_iii_reference": "Annex III point 5(b) — AI systems intended to evaluate the creditworthiness of natural persons",
      "provider": "AiExponent Reference Deployment",
      "deployment_context": "Financial Services / Automated Credit Approval"
    },
    "dimension_summary": {
      "data_governance": { "status": "CONTROLLED", "residual_risk": "LOW" },
      "human_oversight": { "status": "CONTROLLED", "residual_risk": "LOW" },
      "transparency": { "status": "CONTROLLED", "residual_risk": "LOW" },
      "robustness": { "status": "CONTROLLED", "residual_risk": "LOW" },
      "discrimination": { "status": "CONTROLLED", "residual_risk": "ACCEPTABLE" },
      "privacy": { "status": "CONTROLLED", "residual_risk": "LOW" },
      "health_safety": { "status": "CONTROLLED", "residual_risk": "NEGLIGIBLE" },
      "fundamental_rights": { "status": "CONTROLLED", "residual_risk": "ACCEPTABLE" }
    },
    "risks": [
      {
        "id": "RSK-001",
        "dimension": "discrimination",
        "title": "Proxy demographic bias in credit scoring feature weights",
        "pre_mitigation_score": 16,
        "mitigation_id": "MIT-001",
        "mitigation_description": "Disparate impact analysis with re-weighing algorithm maintaining DI ratio >= 0.94 across protected classes.",
        "post_mitigation_score": 4,
        "residual_band": "ACCEPTABLE",
        "status": "MITIGATED"
      },
      {
        "id": "RSK-002",
        "dimension": "human_oversight",
        "title": "Unsupervised rejection without human contestability",
        "pre_mitigation_score": 20,
        "mitigation_id": "MIT-002",
        "mitigation_description": "Mandatory dual-human sign-off on automated adverse determinations pursuant to Article 14.",
        "post_mitigation_score": 4,
        "residual_band": "LOW",
        "status": "MITIGATED"
      }
    ]
  },
  "cross_references": [
    {
      "article_ref": "Art.9(2)(a)",
      "iso42001_ref": "Clause A.7",
      "nist_rmf_ref": "MEASURE 2.9",
      "risk_item_ids": ["RSK-001", "RSK-002"]
    },
    {
      "article_ref": "Art.10(2)(f)",
      "iso42001_ref": "Clause A.7.4",
      "nist_rmf_ref": "MANAGE 1.3",
      "risk_item_ids": ["RSK-001"]
    },
    {
      "article_ref": "Art.14(1)",
      "iso42001_ref": "Clause 6.1",
      "nist_rmf_ref": "GOVERN 1.5",
      "risk_item_ids": ["RSK-002"]
    }
  ]
}