{
  "id": "workflow:legal-ai-bias-audit",
  "_kind": "Workflow",
  "_file": "workflows/workflows/workflows-thin-domains-fill.yaml",
  "_cluster": "workflows",
  "attributes": {
    "displayName": "Legal AI Bias Audit",
    "workflowKind": "governance",
    "triggerType": "scheduled",
    "typicalCadence": "semi-annually",
    "complexity": "cross-team",
    "description": "Audits AI systems used in legal workflows for bias and fairness --\nevaluating contract review model performance disparities across\ndocument types, jurisdictions, and languages, testing e-discovery\nrelevance models for demographic and topic bias in responsive\ndocument identification, assessing predictive coding consistency\nbetween reviewers and model predictions, reviewing litigation outcome\nprediction models for protected-class correlation leakage, auditing\nlegal research tool ranking algorithms for citation bias, evaluating\nexplainability and transparency of AI-assisted legal reasoning, and\nverifying compliance with emerging AI governance regulations (EU AI\nAct, state-level requirements). Produces bias audit report, fairness\nmetric dashboard, and mitigation recommendations. Excludes model\nretraining.\n"
  },
  "outgoingEdges": [
    {
      "from": "workflow:legal-ai-bias-audit",
      "to": "role:ml-engineer",
      "kind": "involves_role",
      "attributes": {}
    },
    {
      "from": "workflow:legal-ai-bias-audit",
      "to": "role:security-reviewer",
      "kind": "involves_role",
      "attributes": {}
    },
    {
      "from": "workflow:legal-ai-bias-audit",
      "to": "role:data-scientist",
      "kind": "involves_role",
      "attributes": {}
    },
    {
      "from": "workflow:legal-ai-bias-audit",
      "to": "skill-area:ml-fine-tuning",
      "kind": "requires_skill_area",
      "attributes": {}
    },
    {
      "from": "workflow:legal-ai-bias-audit",
      "to": "skill-area:data-quality",
      "kind": "requires_skill_area",
      "attributes": {}
    },
    {
      "from": "workflow:legal-ai-bias-audit",
      "to": "domain:legaltech",
      "kind": "applies_to_domain",
      "attributes": {}
    },
    {
      "from": "workflow:legal-ai-bias-audit",
      "to": "domain:legal",
      "kind": "applies_to_domain",
      "attributes": {}
    },
    {
      "from": "workflow:legal-ai-bias-audit",
      "to": "domain:data-science",
      "kind": "applies_to_domain",
      "attributes": {}
    },
    {
      "from": "workflow:legal-ai-bias-audit",
      "to": "responsibility:security-review",
      "kind": "triggers_responsibility",
      "attributes": {}
    },
    {
      "from": "workflow:legal-ai-bias-audit",
      "to": "responsibility:data-quality-monitoring",
      "kind": "triggers_responsibility",
      "attributes": {}
    },
    {
      "from": "workflow:legal-ai-bias-audit",
      "to": "org-unit:ml-platform-team",
      "kind": "performed_by_org_unit",
      "attributes": {}
    },
    {
      "from": "workflow:legal-ai-bias-audit",
      "to": "org-unit:legal-team",
      "kind": "performed_by_org_unit",
      "attributes": {}
    },
    {
      "from": "workflow:legal-ai-bias-audit",
      "to": "org-unit:compliance-team",
      "kind": "performed_by_org_unit",
      "attributes": {}
    }
  ],
  "incomingEdges": []
}