{
  "source": "Cooked Index — occupational AI risk register",
  "page": "https://cookedindex.com/jobs/education-administrators-all-other/",
  "methodology": "https://cookedindex.com/methodology",
  "notice": "Verdicts are re-examined as evidence accumulates. Re-fetch before relying on this; the page above always carries the current score.",
  "scored_at": "2026-08-11",
  "model": "claude-opus-5",
  "occupation": {
    "title": "Education Administrators, All Other",
    "soc_code": "11-9039",
    "category": "Management",
    "us_employment": 55130,
    "median_annual_wage": 95200
  },
  "verdict": "EXPOSED",
  "risk_resistance": 49,
  "contested": false,
  "near_boundary": false,
  "dimensions": {
    "task_resistance": 10,
    "embodiment": 7,
    "liability_shield": 6,
    "trust_premium": 12,
    "judgment_accountability": 14
  },
  "reasoning": {
    "task_resistance": "Roughly half the week — enrollment dashboards, catalog and schedule updates, Title IV or Perkins compliance filings, accreditation self-study narratives assembled from data you already report elsewhere — is drafting and reconciliation that a model does in a first pass, while the other half (sitting with an instructor whose evaluations collapsed, walking an accreditor through a finding, telling a funder why cohort completion dropped) does not survive being handed to software, which is what puts this at 10 rather than the 5 a pure reporting role would get.",
    "embodiment": "You are on campus — classroom observations, facility walk-throughs before a site visit, showing up in a correctional education wing or a hospital training unit where badge access and physical presence are the job — but none of it is skilled manual work, so the physical component is attendance and inspection rather than anything a machine would need hands to replicate.",
    "liability_shield": "Most positions in this bucket require a master's and administrative experience but no state-issued license that attaches personal liability the way a principal's or superintendent's certificate does; the institution's accreditation and its Clery/FERPA/Title IX exposure sit with the president or the general counsel, and a designated compliance officer's signature is institutional, not personal, which is why this lands at 6 rather than in licensed territory.",
    "trust_premium": "Accreditors, agency program officers, employer partners in a corporate training contract, and the faculty you supervise deal with you by name over multi-year cycles, and a reviewer who trusts your self-study reads it differently — but the relationship is instrumental to getting programs approved and funded, not the deliverable itself, so it sits at 12 rather than at the 16-plus of a role where clients follow the individual out the door.",
    "judgment_accountability": "When an accreditor issues a finding, when a program's enrollment no longer supports its faculty line, when a student grievance against an instructor is credible but not provable, you make the call under incomplete facts and it becomes the institution's position — there is no procedure manual for cutting a program or non-renewing an instructor, and the consequences land on people's employment and students' credentials."
  },
  "rationale": "This is a residual bucket — program directors, accreditation and compliance administrators, deans of specialized units, corporate and correctional education managers — so the modal worker splits time between paperwork (accreditation self-studies, enrollment reports, budget narratives, course schedules, grant compliance filings) and human work (supervising instructors, resolving student and parent disputes, negotiating with agencies and funders). The paperwork half is squarely in reach of current AI, which drafts policy language, reconciles data, and assembles compliance packets faster than a small admin office can. What holds is the person who owns the decision when a faculty member is failing, an accreditor raises a finding, or a program has to be cut — and who signs their name to it.",
  "outlook": "The reporting and scheduling side of the job compresses hard, thinning administrative support layers, while the smaller tier that owns personnel, compliance findings, and program-level bets remains and gets busier.",
  "what_would_raise_it": {
    "levers": [
      {
        "dimension": "liability_shield",
        "change": "Accreditor and federal rules that name a specific individual as personally accountable signatory: regional accreditors (e.g. SACSCOC, MSCHE) already require a named institutional accreditation liaison officer, and ED's program participation agreements require a designated Title IV compliance officer whose certification of enrollment/attendance data carries False Claims Act exposure. If accreditors add an explicit attestation that AI-generated self-study or assessment evidence was reviewed and adopted by the named officer, the sign-off becomes non-delegable.",
        "plausibility": "plausible",
        "would_add": 4
      },
      {
        "dimension": "liability_shield",
        "change": "State licensure/approval regimes for specific program types — correctional education program directors under state DOC contracts, clinical program directors requiring a licensed director of record (nursing programs under state boards of nursing already require a named, credentialed program administrator). Expansion of 'named director of record' requirements to more specialized units, or Clery/Title IX coordinator designations being folded into these roles, attaches personal liability.",
        "plausibility": "already happening",
        "would_add": 3
      },
      {
        "dimension": "judgment_accountability",
        "change": "If institutions adopt AI-governance policies requiring a human administrator to be the accountable decision-maker for any adverse action driven by algorithmic flags (academic integrity findings, early-alert dismissals, program discontinuation), the role's remaining work is concentrated in contested calls. Illinois HB 3773 and NYC Local Law 144-style rules on automated decision tools point this direction for employment; extension to student-facing decisions would do it here.",
        "plausibility": "plausible",
        "would_add": 3
      },
      {
        "dimension": "task_resistance",
        "change": "Genuine two-tier structure: if reporting, budget narratives, schedule building and compliance packet assembly are automated, what remains is faculty performance management, accreditor site-visit response, funder negotiation, and dispute adjudication — all low-volume, high-ambiguity. Task-mix shift alone raises the share of the day AI cannot do at usable quality, without any new law.",
        "plausibility": "already happening",
        "would_add": 3
      },
      {
        "dimension": "trust_premium",
        "change": "Narrow route only: accreditor site visits and agency/funder negotiations where the counterparty is itself a human committee that requires an accountable human interlocutor. Grant funders (state agencies, foundations) requiring a named human program officer of contact is common; buyers do not otherwise pay a premium for a human administrator, so the upside is small.",
        "plausibility": "plausible",
        "would_add": 2
      }
    ],
    "ceiling_note": "This is a residual SOC bucket, so any lever applies unevenly — the clinical/correctional/Title IV subsets can gain real liability shielding while corporate training managers and generic program coordinators have essentially no route. Aggregate movement is capped by that heterogeneity: the code as a whole is unlikely to leave the exposed band even if its regulated subsets harden considerably."
  },
  "adjudication": null,
  "employment_history": {
    "points": [
      {
        "y": 2017,
        "emp": 36190,
        "wage": 81630
      },
      {
        "y": 2018,
        "emp": 41000,
        "wage": 82850
      },
      {
        "y": 2019,
        "emp": 44550,
        "wage": 85450
      },
      {
        "y": 2020,
        "emp": 43580,
        "wage": 87580
      },
      {
        "y": 2021,
        "emp": 49970,
        "wage": 90560
      },
      {
        "y": 2022,
        "emp": 50180,
        "wage": 89130
      },
      {
        "y": 2023,
        "emp": 50690,
        "wage": 88460
      },
      {
        "y": 2024,
        "emp": 53330,
        "wage": 89040
      },
      {
        "y": 2025,
        "emp": 55130,
        "wage": 95200
      }
    ],
    "from": 2017,
    "to": 2025,
    "change_pct": 52.3,
    "comparable_from": 2019,
    "spans_soc_revision": true
  },
  "pivots": [],
  "license": "https://cookedindex.com/terms"
}