{ "source": "Cooked Index — occupational AI risk register", "page": "https://cookedindex.com/jobs/teaching-assistants-postsecondary/", "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": "Teaching Assistants, Postsecondary", "soc_code": "25-9044", "category": "Education", "us_employment": 164090, "median_annual_wage": 42910 }, "verdict": "EXPOSED", "risk_resistance": 36, "contested": false, "near_boundary": false, "dimensions": { "task_resistance": 8, "embodiment": 10, "liability_shield": 1, "trust_premium": 11, "judgment_accountability": 6 }, "reasoning": { "task_resistance": "Rubric-based grading of problem sets, drafting written feedback, and answering repeat forum questions are already being handed to LLM graders in large lecture courses, and what holds this at 8 rather than 4 is the live discussion section, wet-lab or studio supervision, and proctoring that still requires a body in a room at a fixed hour.", "embodiment": "A 10 fits the split day: office hours and grading happen at a laptop anywhere, but running a chemistry or biology lab section means handling reagents and equipment, enforcing PPE, and physically walking benches — done in a controlled campus space on a fixed timetable, not an uncontrolled field site, which is why it sits mid-scale rather than in the teens.", "liability_shield": "TAs hold no teaching licence and no credential beyond enrolled-student status; the instructor of record signs the grade roster, owns FERPA compliance and academic-integrity findings, and the lab PI holds the safety authority, so a TA's judgment can be overturned by an email.", "trust_premium": "Students genuinely bond with the TA who explains the proof that finally clicks and writes the recommendation-adjacent recognition, but the appointment turns over every semester or two and rosters get reshuffled, so the relationship rarely outlives the term — an 11 reflects real rapport with no continuity to protect it.", "judgment_accountability": "Grades are set against a rubric and a curve the instructor of record defines, suspected plagiarism gets referred upward rather than adjudicated, and syllabus and accommodation decisions are not the TA's to make; the discretion that exists is pacing a section and deciding how much partial credit an unusual solution earns." }, "rationale": "The modal postsecondary TA is a graduate student who grades problem sets and essays, writes feedback, holds office hours, answers course-forum questions, and runs discussion or lab sections. Grading, rubric application, feedback drafting, and 24/7 Q&A are exactly what LLM course assistants now do at usable quality and near-zero marginal cost, and universities are deploying them explicitly to cut grading load. What survives is the embodied part: standing in a lab or seminar room, reading confusion on faces, proctoring, and running hands-on demonstrations — plus the fact that TA lines exist partly as graduate funding, which insulates headcount from pure efficiency logic.", "outlook": "TA headcount will drift down as AI graders and course assistants absorb grading and Q&A, but sections, labs, and grad-funding politics keep a smaller, more instruction-facing version of the role alive through the 2030s.", "what_would_raise_it": { "levers": [ { "dimension": "trust_premium", "change": "Graduate-worker union contracts (UAW-affiliated locals at Columbia, Harvard, Michigan, the 48,000-worker UC system) adding explicit clauses on TA-to-student section ratios and bans on replacing bargained TA appointments with automated grading — the UC 2022 contract already fixed appointment counts and workload caps; an AI-displacement clause of the kind WGA won in 2023 would decouple headcount from efficiency further", "plausibility": "already happening", "would_add": 4 }, { "dimension": "trust_premium", "change": "Accreditor and federal 'regular and substantive interaction' enforcement: ED's distance-education rule requires instructor-initiated human interaction for Title IV eligibility, and if ED or regional accreditors (e.g. SACSCOC, WSCUC) issue guidance that AI course assistants do not count as RSI, human section leaders and graders become a compliance requirement for aid-eligible courses", "plausibility": "plausible", "would_add": 5 }, { "dimension": "task_resistance", "change": "Task-mix shift as auto-graded problem sets and forum Q&A are absorbed: the residual TA day becomes live section facilitation, in-person oral/whiteboard assessment, and academic-integrity adjudication of suspected AI submissions — a genuine two-tier job where the judgment tier is what remains", "plausibility": "already happening", "would_add": 3 }, { "dimension": "embodiment", "change": "Reversion to proctored in-person and oral examination as a response to AI cheating (already spreading in Australian universities under TEQSA guidance and in US STEM gateway courses), plus lab safety rules requiring a trained human present for hands-on wet-lab and machine-shop sections", "plausibility": "already happening", "would_add": 3 }, { "dimension": "judgment_accountability", "change": "Honor-code and Title IX procedures naming the section leader as the reporting party of record: if institutional AI-misconduct policy requires a named human TA to make the initial determination and testify at hearings — and courts continue to strike sanctions based on unverified AI detectors — the ambiguity call sits with an identifiable person", "plausibility": "plausible", "would_add": 4 } ], "ceiling_note": "No route to a real liability_shield: TAs hold no license, sign nothing legally operative, and grade under a faculty instructor of record who bears whatever accountability exists. Also note the funding logic cuts both ways — TA lines survive as stipends, but that same logic makes them a target when graduate cohorts shrink, and no lever here protects against enrollment-driven cuts." }, "adjudication": null, "employment_history": { "points": [ { "y": 2019, "emp": 137500, "wage": 32080 }, { "y": 2020, "emp": 138740, "wage": 36250 }, { "y": 2021, "emp": 121290, "wage": 38040 }, { "y": 2022, "emp": 135160, "wage": 38050 }, { "y": 2023, "emp": 145960, "wage": 43750 }, { "y": 2024, "emp": 155010, "wage": 44930 }, { "y": 2025, "emp": 164090, "wage": 42910 } ], "from": 2019, "to": 2025, "change_pct": 19.3, "comparable_from": 2019, "spans_soc_revision": false }, "pivots": [ { "slug": "substitute-teachers-short-term", "title": "Substitute Teachers, Short-Term", "verdict": "EXPOSED", "risk_resistance": 53, "median_wage": 41670, "overlap": 72, "skills_to_close": [ "Persuasion" ] }, { "slug": "philosophy-and-religion-teachers-postsecondary", "title": "Philosophy and Religion Teachers, Postsecondary", "verdict": "EXPOSED", "risk_resistance": 48, "median_wage": 80260, "overlap": 67, "skills_to_close": [ "Critical Thinking", "Systems Evaluation", "Active Learning", "Systems Analysis" ] }, { "slug": "secondary-school-teachers-except-special-and-career-technical-education", "title": "Secondary School Teachers, Except Special and Career/Technical Education", "verdict": "SAFE", "risk_resistance": 69, "median_wage": 72040, "overlap": 65, "skills_to_close": [ "Critical Thinking", "Monitoring", "Social Perceptiveness", "Active Learning" ] } ], "license": "https://cookedindex.com/terms" }