Mathematical Science Teachers, Postsecondary

25-1022.00
Median wage $79,940/yr47,670 employed (US)Rank #401 of 923 scored · top 43% by substitution

Teach courses pertaining to mathematical concepts, statistics, and actuarial science and to the application of original and standardized mathematical techniques in solving specific problems and situations. Includes both teachers primarily engaged in teaching and those who do a combination of teaching and research.

Sub-scores

0–100 · band = confidence interval from rater disagreement

Substitution30
Exposure27
Augmentation61

Substitution — the headline: capability discounted by cost, barriers and adoption.

Exposure — technical capability alone, regardless of whether anyone deploys it.

Augmentation — how much AI assists without replacing. High here + moderate substitution = a changing job, not a disappearing one.

Tasks on the substitution scale

23 rated tasks, binned by substitution score.

Position among all scored occupations

Distribution of 923 occupation scores; the marker is this occupation.

Tasks with substitution ≥ 70

9%

Run 1.0.0-draft.1 · computed 2026-08-05 · rater panel: claude-sonnet-5, claude-haiku-4-5-20251001 · intervals span rater disagreement.

Why this score

The five weighted dimensions of the composite, averaged across this occupation's tasks (importance-weighted, panel mean). Exact weights and formulas: /api/v1/methodology.

Task automatabilityw 35%28

panel mean rating 2.1/5 → substitution pressure 28/100

Technical feasibility todayw 20%27

panel mean rating 2.1/5 → substitution pressure 27/100

Cost vs. human wagew 15%32

panel mean rating 2.3/5 → substitution pressure 32/100

Adoption barriersw 20%inverted — strong barriers lower the score37

panel mean rating 3.5/5 (barrier strength) → substitution pressure 37/100

Sector adoption velocityw 10%24

panel mean rating 2.0/5 → substitution pressure 24/100

Task breakdown (23 tasks)

Substitution pressure per task, weighted by O*NET importance in the composite. Expand a task for the full rater audit trail — every rating, every model, every rationale.

Maintain student attendance records, grades, and other required records.

87

CI 79–95 · exposure 87 · augmentation 75 · importance 4.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510015/5Postsecondary institutions have achieved near-universal adoption of automated attendance and grade-tracking systems; this is standard practice in higher education globally.
Sector adoption velocityclaude-sonnet-54/5Higher education has broadly adopted digital LMS and gradebook systems for years, making this one of the most digitized administrative tasks in the sector.
Augmentation potentialclaude-haiku-4-5-202510014/5AI assists faculty by automating routine data entry, flagging attendance anomalies, and generating analytics dashboards, freeing time for pedagogical focus while the instructor remains responsible for grades and decisions.
Augmentation potentialclaude-sonnet-54/5AI-enabled gradebooks and attendance systems significantly reduce administrative burden on instructors, freeing time for teaching while the instructor still reviews and finalizes records.
Task automatabilityclaude-haiku-4-5-202510015/5Record management (attendance, grades, transcripts) is inherently structured and digital; current systems can fully automate data entry, updates, and maintenance with no quality loss and substantial time savings via learning management systems and administrative software.
Task automatabilityclaude-sonnet-54/5Recording and tallying attendance and grades is a structured, rules-based clerical task that LMS/gradebook software and AI-integrated systems can already handle end-to-end with minimal human input.
Adoption barriersclaude-haiku-4-5-202510012/5Light adoption barriers exist (institutional IT governance, data privacy compliance, faculty preference for specific interfaces), but no legal requirement mandates human record-keeping; most institutions already use automated systems.
Adoption barriersclaude-sonnet-52/5Some institutional policy requires instructor sign-off or verification of final grades, but the mechanical recordkeeping itself has few legal or licensing barriers.
Cost vs. human wageclaude-haiku-4-5-202510015/5LMS and student information systems cost institutions per-student per-year subscriptions that are orders of magnitude cheaper than hiring staff to manually maintain these records.
Cost vs. human wageclaude-sonnet-55/5Automated recordkeeping via existing software costs a tiny fraction of the instructor time it would take to do this manually, especially at scale.
Technical feasibility todayclaude-haiku-4-5-202510015/5Mature, deployed products (Canvas, Blackboard, Banner, Workday) reliably perform grade and attendance tracking at scale across thousands of institutions daily.
Technical feasibility todayclaude-sonnet-54/5Learning management systems (Canvas, Blackboard, etc.) already automate attendance tracking and gradebook calculations reliably in production at most universities today.

Prepare course materials, such as syllabi, homework assignments, and handouts.

76

CI 76–76 · exposure 75 · augmentation 100 · importance 4.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510013/5Higher education adoption of AI for course material generation is accelerating but remains in pilot and selective production phases; full institutional embrace is still emerging rather than widespread in the sector.
Sector adoption velocityclaude-sonnet-53/5Higher education is adopting generative AI for course prep at a moderate pace, with growing but uneven use across departments and institutions.
Augmentation potentialclaude-haiku-4-5-202510015/5AI dramatically enhances instructor productivity by drafting initial materials, enabling rapid iteration, personalization, and accommodation of diverse learning styles while faculty retain full pedagogical control and customization.
Augmentation potentialclaude-sonnet-55/5AI substantially speeds up drafting of syllabi, assignments, and handouts, letting instructors focus on refining and verifying mathematical content.
Task automatabilityclaude-haiku-4-5-202510014/5AI can generate substantial portions of syllabi, homework assignments, and handouts from learning objectives and course parameters, achieving significant time savings. However, final review and customization by instructors to align with institution-specific requirements and pedagogical philosophy typically remains necessary.
Task automatabilityclaude-sonnet-54/5LLMs can draft syllabi, homework sets, and handouts from a topic outline or textbook chapter with substantial time savings, though instructors still need to review for accuracy and alignment with course goals.
Adoption barriersclaude-haiku-4-5-202510012/5No legal requirement mandates human authorship of course materials, and institutions have no regulatory bar to automation. Organizational inertia and faculty preference for control over their curriculum design offer some friction but are not hard barriers.
Adoption barriersclaude-sonnet-52/5No licensing requirement blocks AI-assisted material creation, though institutional policies and instructor accountability for accuracy create mild friction.
Cost vs. human wageclaude-haiku-4-5-202510015/5The cost of AI inference and integration is negligible compared to the instructor labor cost of designing syllabi, assignments, and handouts from scratch, particularly at the hourly rates of postsecondary faculty.
Cost vs. human wageclaude-sonnet-55/5Generating draft materials via AI costs a fraction of a cent to a few dollars in compute versus hours of faculty or TA time at academic wages.
Technical feasibility todayclaude-haiku-4-5-202510014/5LLMs and specialized educational tools reliably generate course materials in production use by many universities. Error rates on factual accuracy and pedagogical appropriateness exist but are manageable with instructor review, making deployment practical at scale.
Technical feasibility todayclaude-sonnet-54/5Widely deployed tools (ChatGPT, Copilot, dedicated ed-tech assistants) are already used by instructors to generate drafts of syllabi and problem sets in production settings, though math content requires verification for correctness.

Participate in student recruitment, registration, and placement activities.

51

CI 30–71 · exposure 53 · augmentation 63 · importance 3.0/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510013/5Higher education has adopted ATS and CRM tools widely, but deep adoption of agent-based or autonomous recruitment is still in pilot phase at most institutions. Adoption is uneven and slower than in corporate recruiting, reflecting sector conservatism and budget constraints.
Sector adoption velocityclaude-sonnet-52/5Higher education is a moderate-to-slow adopter of AI for administrative/student-facing processes, with pilots in chatbot-based advising but limited deep integration into faculty recruitment/placement duties.
Augmentation potentialclaude-haiku-4-5-202510014/5AI can significantly assist faculty and staff by automating scheduling, filtering applications, flagging job matches, and producing outreach templates, freeing educators to focus on relationship-building, career counseling, and placement judgment. Productivity gains are substantial while human oversight remains central.
Augmentation potentialclaude-sonnet-53/5AI can assist with drafting recruitment materials, managing registration data, and analyzing placement test results, providing useful support while faculty retain decision-making roles.
Task automatabilityclaude-haiku-4-5-202510015/5AI can automate substantial parts of recruitment (email outreach, event coordination reminders), handle registration workflows (form processing, data entry, initial screening), and assist placement matching (resume parsing, job-skill alignment, interview scheduling). These steps easily exceed 50% time savings at comparable quality compared to manual human effort.
Task automatabilityclaude-sonnet-52/5This task involves relational, administrative, and evaluative activities (advising prospective students, admissions decisions, placement testing coordination) that require institutional judgment and personal interaction, limiting end-to-end automation.__AI can support parts like data processing but not the whole task.
Adoption barriersclaude-haiku-4-5-202510013/5Universities often prefer human faculty involvement in recruitment and placement for relationship and institutional context; there are soft norms around personal outreach, but no hard legal requirement that a licensed human must perform these tasks. Some organizational friction around trust in AI judgment exists, but regulatory barriers are minimal.
Adoption barriersclaude-sonnet-53/5No licensing barrier exists specifically for this task, but institutional policy, accreditation, and personalized advising expectations create moderate friction against full automation.
Cost vs. human wageclaude-haiku-4-5-202510014/5AI-powered CRM, email automation, and ATS solutions cost a fraction of dedicated full-time human recruiters or placement coordinators on a per-task basis. Setup costs exist but ongoing inference and processing are orders of magnitude cheaper than loaded salary for equivalent volume.
Cost vs. human wageclaude-sonnet-52/5While software can reduce some administrative overhead, the faculty labor cost for this task is a small fraction of their role, and AI tools still require human oversight and integration into institutional systems, keeping costs roughly comparable rather than dramatically cheaper.
Technical feasibility todayclaude-haiku-4-5-202510013/5Deployed systems like CRM platforms (HubSpot, Salesforce) with AI features, applicant tracking systems (ATS), and chatbots handle parts of this work in production, but full end-to-end automation of recruitment-to-placement with reliable judgment on candidate fit remains limited. Material gaps remain in relationship-building and nuanced placements.
Technical feasibility todayclaude-sonnet-52/5Some products exist for chatbot-based recruitment outreach and automated placement test scoring, but faculty involvement in recruitment/registration/placement is still largely manual and institution-specific with no mature end-to-end product replacing faculty participation.

Compile bibliographies of specialized materials for outside reading assignments.

49

CI 25–72 · exposure 45 · augmentation 75 · importance 2.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Higher education has been slower to adopt AI-driven course preparation tools compared to information and finance sectors; faculty adoption of AI for pedagogy remains limited and tentative, with pilots more common than production use. Conservative institutional cultures and manual workflows still dominate.
Sector adoption velocityclaude-sonnet-53/5Higher education has moderate AI adoption for research and content support tasks, with pilots and individual faculty use common but no widespread institutional standardization yet.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist by surfacing candidate papers, auto-organizing by topic, and suggesting related works, meaningfully speeding up the research phase while the instructor remains the final arbiter of what appears in the bibliography. This is useful assistance but not transformative productivity gain.
Augmentation potentialclaude-sonnet-55/5AI substantially speeds up literature discovery and bibliography drafting, letting instructors quickly generate and refine reading lists while retaining final judgment on suitability and rigor.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can generate lists of references and identify relevant papers for mathematical topics, compiling a curated bibliography that reflects a specific instructor's pedagogical intent, course level, and specialized focus requires subjective judgment and domain expertise that current systems struggle with reliably. The task lacks the structured, repetitive character needed for substantial time savings at equal quality.
Task automatabilityclaude-sonnet-54/5Compiling bibliographies of specialized reading materials is a well-structured search-and-list task that current AI systems can perform quickly by leveraging literature databases and topic knowledge, saving significant instructor time.4=but human curation for course-specific fit still adds value.
Adoption barriersclaude-haiku-4-5-202510014/5Academic institutions maintain strong norms around faculty curation of reading materials, and instructors retain significant autonomy and responsibility for course design and content. Substituting AI-compiled bibliographies may conflict with institutional expectations that faculty exercise professional judgment and accountability for student learning materials.
Adoption barriersclaude-sonnet-51/5There is no licensing, regulatory, or liability barrier to using AI for compiling reading lists; it's an administrative/academic support task with no legal requirement for human execution.
Cost vs. human wageclaude-haiku-4-5-202510012/5An instructor spends perhaps 2–3 hours manually compiling a bibliography; current AI tools (API calls, subscriptions) plus the overhead of verification and correction often approach or exceed the cost of direct human effort, especially when factoring in hallucination correction.
Cost vs. human wageclaude-sonnet-54/5AI-based bibliography compilation via LLMs or citation tools costs a small fraction of an instructor's or research assistant's time-equivalent, though occasional verification labor is needed.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed product reliably handles this task end-to-end; citation aggregators (Google Scholar, ResearchGate) and LLM-based tools can assist but produce inconsistent results, missing papers, hallucinated citations, and require substantial human verification. Production reliability remains low for ensuring quality and completeness.
Technical feasibility todayclaude-sonnet-53/5Tools like AI-assisted literature search and citation managers (e.g., Elicit, Semantic Scholar AI, ChatGPT with browsing) can generate bibliography drafts today, but accuracy issues like hallucinated citations require verification, limiting reliability in production use.

Compile, administer, and grade examinations, or assign this work to others.

45

CI 39–51 · exposure 42 · augmentation 75 · importance 4.6/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Postsecondary education is a laggard sector for automation; while learning-management systems are common, actual deployment of AI for end-to-end exam compilation and grading remains limited, with most institutions retaining human control over assessment rigor.
Sector adoption velocityclaude-sonnet-52/5Higher education is a slower-adopting sector for grading automation; while some pilots for auto-grading exist, postsecondary math instruction has not seen deep, widespread AI-driven exam administration.
Augmentation potentialclaude-haiku-4-5-202510014/5AI can meaningfully assist instructors by drafting exams, organizing items by difficulty, suggesting partial-credit rubrics, and flagging unusual grade distributions, thereby raising instructor productivity while maintaining human judgment on final assessment decisions.
Augmentation potentialclaude-sonnet-54/5AI meaningfully assists in generating question banks, formatting exams, and pre-grading objective items, saving instructor time while the instructor retains final grading authority.
Task automatabilityclaude-haiku-4-5-202510012/5AI can auto-generate and grade multiple-choice or computational exams at scale, but postsecondary mathematics exams often require nuanced grading of proofs, derivations, and reasoning—tasks where current AI struggles with consistent interpretation of correctness standards set by human instructors.
Task automatabilityclaude-sonnet-53/5AI can draft exam questions and grade objective or even short-answer/essay responses with reasonable accuracy, but compiling a coherent exam aligned to course objectives and grading proofs/mathematical reasoning still requires human oversight for validity and fairness.
Adoption barriersclaude-haiku-4-5-202510013/5Academic integrity, accreditation standards, and departmental oversight create friction; additionally, faculty often prefer direct control over assessment design and judgment on borderline answers, slowing substitution despite technical capability.
Adoption barriersclaude-sonnet-53/5No licensing requirement blocks AI use, but academic integrity, grading fairness, and institutional grading policies create moderate friction, especially for high-stakes exams.
Cost vs. human wageclaude-haiku-4-5-202510014/5Once integrated, AI-driven exam platforms have very low per-exam inference costs compared to instructor time spent on compilation, administration, and grading, though setup and oversight infrastructure add overhead.
Cost vs. human wageclaude-sonnet-53/5AI tools can cut time on drafting and objective grading significantly, but the need for human verification of grading accuracy on math work keeps overall cost savings moderate rather than order-of-magnitude.
Technical feasibility todayclaude-haiku-4-5-202510013/5Deployed systems exist for objective grading (multiple-choice, fill-in-the-blank) and basic exam generation, but these are typically narrow in scope; robust production systems that handle the full pipeline of compilation, administration, and grading of rigorous mathematics exams remain limited.
Technical feasibility todayclaude-sonnet-53/5Products like AI-assisted grading tools and question generators exist and are used in some courses, but reliable automated grading of mathematical proofs or multi-step derivations remains narrow and error-prone in production.

Evaluate and grade students' class work, assignments, and papers.

40

CI 25–55 · exposure 38 · augmentation 63 · importance 4.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Postsecondary education is a relatively laggard sector for automation; most mathematics departments continue manual grading or basic script-based checking. Pilot adoption exists but production deployment of AI grading for substantive mathematical work remains minimal and confined to large, digitally mature institutions.
Sector adoption velocityclaude-sonnet-53/5Higher education is adopting AI-assisted grading tools steadily, especially for STEM homework platforms, though full-scale grading automation for postsecondary math courses remains a pilot-to-moderate stage.'},
Augmentation potentialclaude-haiku-4-5-202510013/5AI can meaningfully assist faculty by pre-screening submissions, flagging anomalies, suggesting partial-credit categories, and organizing submissions for review, thereby reducing time spent on routine administrative work. However, the human expert remains essential for interpreting mathematical reasoning and assigning grades.
Augmentation potentialclaude-sonnet-54/5AI substantially speeds up grading of routine problems, generates rubric-aligned feedback, and flags common errors, meaningfully augmenting instructor throughput while they retain final oversight.'},
Task automatabilityclaude-haiku-4-5-202510012/5AI can assist with grading routine problems (e.g., multiple choice, mechanical calculations) and flag submissions for review, but cannot reliably evaluate complex mathematical reasoning, proof writing, or nuanced partial credit decisions that require deep understanding of pedagogical intent. Full end-to-end automation at equal quality remains infeasible.
Task automatabilityclaude-sonnet-53/5AI can grade well-structured problem sets and check numerical/symbolic answers with high accuracy, but grading proofs, partial credit reasoning, and open-ended mathematical argumentation still requires human judgment for most postsecondary-level rigor.ed rating middling.'},'automatability_note'):
Adoption barriersclaude-haiku-4-5-202510014/5Faculty retain strong autonomy and professional authority over assessment; institutions require human judgment on grading rubrics, learning outcomes, and appeals; accreditation and student-facing transparency expectations create friction against full delegation; and student and faculty preference for human grading (especially in mathematics) adds organizational resistance.
Adoption barriersclaude-sonnet-52/5No licensing requirement mandates a human grade coursework, but academic integrity, grade appeals, and institutional policy often require instructor accountability for final grades.'},
Cost vs. human wageclaude-haiku-4-5-202510012/5Basic automated grading tools have low per-task cost, but integration, calibration, and the human review overhead required to catch errors and ensure pedagogical validity substantially increase total cost. For postsecondary mathematics, where nuanced grading is critical, AI cost-plus-oversight often exceeds a faculty member's marginal time cost.
Cost vs. human wageclaude-sonnet-53/5For routine computational assignments AI grading is much cheaper, but for proof-based or conceptual work requiring careful evaluation, human oversight cost narrows the gap significantly.'},
Technical feasibility todayclaude-haiku-4-5-202510012/5While basic automated grading systems (e.g., LMS plugins, simple answer checkers) are deployed in some institutions, they handle only constrained problem types and require significant human oversight for correctness and fairness. Reliable production systems for comprehensive mathematical assessment across proof-based and applied work do not exist at scale.
Technical feasibility todayclaude-sonnet-53/5Automated grading tools exist for homework platforms (e.g., WebAssign, Gradescope AI-assisted rubrics) and are used in production for routine problem sets, but grading proofs and complex written work remains largely manual or lightly AI-assisted with instructor review.'},

Develop department and course schedules.

39

CI 25–52 · exposure 38 · augmentation 75 · importance 3.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Higher education is a traditionally slow-adopting sector; while some large institutions have experimented with AI scheduling, most departments still rely on manual processes or legacy systems, and there is limited evidence of rapid, widespread production deployment.
Sector adoption velocityclaude-sonnet-52/5Higher education administration adopts new software slowly due to bureaucracy, though scheduling tools have existed for years without full replacement of human oversight.
Augmentation potentialclaude-haiku-4-5-202510014/5AI can meaningfully augment this task by generating candidate schedules, flagging conflicts, optimizing room utilization, and modeling trade-offs, helping department chairs or administrators make faster, more informed decisions while retaining final human judgment over institutional and interpersonal factors.
Augmentation potentialclaude-sonnet-54/5AI-assisted scheduling tools significantly speed up draft schedule creation and conflict detection, letting the human focus on final approval and exceptions.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist with constraint satisfaction and calendar optimization, developing schedules requires balancing faculty preferences, room availability, student demand, prerequisite sequencing, and institutional politics—most of which involve qualitative judgment and stakeholder negotiation that AI cannot fully automate without substantial human oversight.
Task automatabilityclaude-sonnet-53/5Scheduling optimization is a well-defined constraint-satisfaction problem that AI/software can largely handle, though negotiating faculty preferences and departmental politics still requires human judgment.6.5 percent overlap with real automation potential is moderate.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5.6.5
Adoption barriersclaude-haiku-4-5-202510014/5Scheduling decisions affect faculty working conditions, student access, and institutional operations; faculty governance structures, collective bargaining agreements, and the expectation that tenured faculty have input into their own schedules create substantial organizational and contractual barriers to full automation.
Adoption barriersclaude-sonnet-52/5No licensing requirement to build schedules, but institutional policy, union rules, and faculty preference negotiations create moderate organizational friction.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI scheduling tools require significant setup, maintenance, and human validation, while a department administrator or faculty committee spending a few days per year on scheduling remains cost-competitive for small to medium departments.
Cost vs. human wageclaude-sonnet-53/5Scheduling software licenses plus administrative oversight cost roughly comparable to the time a department chair or coordinator would spend, without dramatic order-of-magnitude savings.
Technical feasibility todayclaude-haiku-4-5-202510012/5Some scheduling software exists, but current AI-based solutions handle only narrow, well-defined constraints and rarely integrate seamlessly with the full complexity of departmental operations; most production systems remain rule-based and require extensive manual input and adjustment.
Technical feasibility todayclaude-sonnet-53/5University scheduling software (e.g., academic scheduling optimization tools) is deployed widely, but final decisions still involve manual override and departmental negotiation.

Select and obtain materials and supplies, such as textbooks.

35

CI 23–47 · exposure 33 · augmentation 63 · importance 3.6/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Educational institutions are relatively slow to automate procurement processes, with human faculty involvement remaining standard practice. Adoption of AI-assisted procurement in higher education remains minimal, with most institutions still relying on traditional channels and human decision-making for textbook selection.
Sector adoption velocityclaude-sonnet-52/5Higher education adopts AI tools unevenly, and administrative/procurement tasks like this are not a primary focus of current AI rollouts in academia.
Augmentation potentialclaude-haiku-4-5-202510013/5AI could usefully assist by summarizing textbook features, comparing prices across vendors, or flagging options that meet search criteria, reducing the time faculty spend on initial research. However, the core decision of fit and institutional approval still rests with humans, so augmentation is helpful but partial.
Augmentation potentialclaude-sonnet-54/5AI can efficiently summarize textbook options, compare content coverage, and check pricing/availability, meaningfully speeding up the selection process for instructors.
Task automatabilityclaude-haiku-4-5-202510012/5This task involves selecting and procuring materials, which requires domain knowledge, vendor evaluation, and decision-making. While AI could assist with searching catalogs or generating shortlists, the actual selection (comparing textbooks across pedagogical fit, cost, institution policy) and obtaining (order placement, vendor negotiation) still require human judgment and institutional authorization.
Task automatabilityclaude-sonnet-53/5AI can research, compare, and recommend textbooks and materials, and even draft procurement requests, but final selection often involves curriculum fit judgments and institutional approval processes that limit full automation.
Adoption barriersclaude-haiku-4-5-202510014/5Educational institutions typically have formal procurement policies, approval chains, and often require authorized personnel to sign off on textbook selections due to budget and accreditation oversight. Many institutions require faculty sign-off, and some contracts require institutional authorization, creating legal and organizational friction.
Adoption barriersclaude-sonnet-52/5No licensing requirement blocks AI assistance, but faculty typically retain final authority over course materials due to academic freedom and curriculum standards.
Cost vs. human wageclaude-haiku-4-5-202510012/5The human labor involved (reviewing options, making informed selections, placing orders) is relatively low-cost compared to the value of correct material selection for a semester's course. AI systems would need integration into institutional procurement systems and oversight, making the all-in cost competitive with a staff member's time rather than dramatically cheaper.
Cost vs. human wageclaude-sonnet-53/5Using AI to search and compare materials is cheap, but the overall task is infrequent and low-cost already, so the savings relative to a professor's time are moderate rather than dramatic.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed product reliably handles end-to-end textbook selection and procurement for educational institutions at scale. Some systems can search library databases or generate recommendations, but they lack the contextual judgment about institutional policies, budget constraints, and pedagogical fit that real procurement requires.
Technical feasibility todayclaude-sonnet-52/5There are no widely deployed products specifically for postsecondary textbook selection; general AI assistants can support research but institutions still rely on manual review and committee processes.

Conduct faculty performance evaluations.

31

CI 7–55 · exposure 33 · augmentation 63 · importance 3.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Higher education adoption of AI for administrative functions lags most sectors; faculty performance evaluation remains governed by shared governance principles and tradition. Pilot adoption exists but production displacement is minimal and slow-moving.
Sector adoption velocityclaude-sonnet-52/5Higher education administration is a slower-adopting sector for AI in personnel decisions, with pilots limited to data aggregation rather than evaluative judgment.
Augmentation potentialclaude-haiku-4-5-202510014/5AI augmentation is already valuable here: automated literature indexing, citation analysis, student evaluation aggregation, and comparative benchmarking all enhance human evaluators' productivity without replacing the evaluator's judgment and accountability.
Augmentation potentialclaude-sonnet-53/5AI can help aggregate metrics like publication counts, citation data, and summarize student feedback, easing some administrative burden of the evaluation process.
Task automatabilityclaude-haiku-4-5-202510014/5AI can largely automate the collection, analysis, and synthesis of performance data (student evaluations, publication metrics, grant funding, teaching metrics). However, the final judgment and interpersonal feedback components require human discretion, so only ~60-75% time savings at equal quality is feasible end-to-end.
Task automatabilityclaude-sonnet-51/5Evaluating faculty performance requires nuanced judgment about teaching quality, research contribution, collegiality, and institutional fit that current AI cannot reliably assess end-to-end.'
Adoption barriersclaude-haiku-4-5-202510014/5Faculty governance traditions, union agreements, and institutional autonomy requirements create significant friction. Legal liability concerns over evaluation decisions, combined with institutional norms favoring human committee review, substantially protect the human role in formal summative judgments.
Adoption barriersclaude-sonnet-54/5Faculty evaluations are governed by tenure processes, union contracts, accreditation standards, and legal liability concerns requiring qualified human evaluators and committee oversight.
Cost vs. human wageclaude-haiku-4-5-202510014/5Automating data aggregation, metric calculation, and preliminary assessment synthesis is substantially cheaper than dedicated administrative time. Full AI integration would be several times cheaper than manual evaluation processes, though human oversight still adds cost.
Cost vs. human wageclaude-sonnet-51/5Since AI cannot substitute for the core judgment-based evaluation, there is no viable cost comparison—human evaluators remain necessary and AI offers no cost-saving substitute for the task itself.
Technical feasibility todayclaude-haiku-4-5-202510013/5Products exist that can aggregate and analyze faculty performance data, but no mature system end-to-end conducts evaluations at scale across institutions. Deployed tools assist with data compilation and comparative analytics, but human review and approval remain standard practice.
Technical feasibility todayclaude-sonnet-51/5No deployed product performs faculty evaluations autonomously; at most AI tools assist with summarizing data like student evaluation text or publication counts.

Write grant proposals to procure external research funding.

31

CI 25–38 · exposure 25 · augmentation 75 · importance 2.9/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Academia is a traditionally slower-adopting sector with strong norms favoring human expertise in proposal writing. While some institutions pilot AI writing tools, deep production adoption of autonomous proposal writing remains limited; most use remains experimental or assistive.
Sector adoption velocityclaude-sonnet-53/5Academia is a professional/knowledge-work sector with rising AI tool use for writing tasks, but grant writing specifically remains cautious due to funder scrutiny and integrity concerns, placing it in a middling adoption pace.
Augmentation potentialclaude-haiku-4-5-202510014/5AI can meaningfully assist faculty in drafting literature reviews, structuring proposal sections, refining language, and generating initial outlines, raising productivity and reducing writing burden while the faculty member retains control over strategy and final submission.
Augmentation potentialclaude-sonnet-54/5AI substantially helps with drafting narrative sections, editing for clarity, formatting to funder templates, and brainstorming framing, meaningfully speeding up the writing process while the researcher retains control over content and strategy.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can draft sections of grant proposals (background, literature review, methodology outline), the task requires substantial human judgment on research merit, institutional fit, funder priorities, and persuasive narrative that current AI systems cannot reliably produce end-to-end. A human expert must still direct strategy and heavily edit outputs to meet the ≥50% time-saving bar.
Task automatabilityclaude-sonnet-52/5AI can draft sections and generate boilerplate text, but crafting a competitive proposal requires original research framing, budget justification tied to real work, and strategic tailoring to funder priorities that current AI cannot reliably originate end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Grant proposals are subject to funder requirements, institutional review, and legal/compliance attestations that typically require faculty signature and accountability. Universities often have established grant-management processes and require human experts to certify submissions, creating organizational and regulatory friction.
Adoption barriersclaude-sonnet-52/5No licensing requirement exists, but funders and institutions expect PI authorship, intellectual ownership, and accountability for scientific claims, creating moderate organizational and reputational friction against pure AI authorship.
Cost vs. human wageclaude-haiku-4-5-202510012/5Grant writing requires senior faculty time and often institutional grant-writing staff or consultants. AI tools may reduce some drafting overhead but do not eliminate the need for expert human labor, and the cost of AI + human oversight is comparable to or higher than hiring specialized grant writers.
Cost vs. human wageclaude-sonnet-52/5AI can cut drafting time cheaply, but the overall cost is dominated by the researcher's expert time reviewing, verifying technical content, and iterating, so total cost savings versus a human-only process are modest.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed product reliably writes complete, competitive grant proposals at production scale. AI writing assistants exist but are used as drafting aids with heavy human oversight, not autonomous performers. Success rates and acceptance outcomes depend almost entirely on human expertise, not AI capability.
Technical feasibility todayclaude-sonnet-52/5Writing assistants (ChatGPT, Grammarly, specialized grant-writing tools) are used to draft and polish text, but no deployed product reliably produces fundable, technically sound grant proposals without heavy expert revision.

Plan, evaluate, and revise curricula, course content, and course materials and methods of instruction.

29

CI 25–34 · exposure 25 · augmentation 63 · importance 4.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Higher education has been slow to adopt AI for core instructional functions; adoption remains largely experimental with faculty pilots rather than systematic institutional deployment in curriculum design.
Sector adoption velocityclaude-sonnet-52/5Higher education adopts new instructional technology slowly due to governance structures, tenure-related incentives, and accreditation processes, so AI-driven curriculum tools remain in early pilot stages rather than broad production use.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist by drafting syllabi, generating example problems, suggesting content organization, and identifying gaps in course materials, genuinely improving productivity when faculty remain in control of final decisions.
Augmentation potentialclaude-sonnet-54/5AI is genuinely useful for brainstorming course content, generating practice problems, drafting learning objectives, and suggesting revisions, substantially speeding up an instructor's curriculum work while they retain final judgment.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can generate draft curriculum content and suggest revisions, the task requires substantial human judgment about pedagogical goals, student needs, and institutional context. AI cannot reliably perform the full evaluation and revision cycle that ensures quality instruction, falling well short of the 50% time-saving threshold.
Task automatabilityclaude-sonnet-52/5AI can draft syllabi, generate exercises, and suggest topic sequencing, but genuine curriculum planning requires institutional context, accreditation alignment, and pedagogical judgment that AI cannot fully replace end-to-end.4>50% time saving is not realistic across the whole task yet.
Adoption barriersclaude-haiku-4-5-202510014/5Curriculum development and evaluation are core faculty responsibilities embedded in academic governance, accreditation standards, and tenure/promotion processes. Institutional autonomy and faculty authority create strong organizational and regulatory barriers to full automation.
Adoption barriersclaude-sonnet-53/5No licensing requirement mandates a human perform curriculum design, but academic governance, faculty senate approval, and departmental review processes create real organizational friction against wholesale automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5Current AI tools require significant human oversight and rework when applied to curriculum planning. Integration costs and necessary faculty review time keep total-cost-of-ownership comparable to or higher than direct faculty labor.
Cost vs. human wageclaude-sonnet-53/5AI drafting assistance is cheap relative to faculty time spent on curriculum work, but the human review, departmental approval, and iterative revision still dominate the cost, keeping the ratio moderate rather than dramatically favorable.
Technical feasibility todayclaude-haiku-4-5-202510012/5AI tools exist for content generation and basic curriculum templating, but deployed products lack the contextual understanding and nuanced judgment required to independently plan, evaluate, and revise postsecondary mathematics curricula at reliable production scale.
Technical feasibility todayclaude-sonnet-52/5Tools like ChatGPT are used informally by instructors for drafting course materials, but no deployed product reliably performs full curriculum design and revision cycles in production at scale for postsecondary math departments.

Keep abreast of developments and technological advances in the mathematical field by reading current literature, talking with colleagues, and participating in professional conferences.

29

CI 21–38 · exposure 17 · augmentation 63 · importance 3.8/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Academic institutions are relatively slow adopters of automation; staying current through reading and conferences remains a valued, human-centric professional practice. AI augmentation tools exist (literature summary, preprint alerts) but deep adoption of AI for this task is still emerging.
Sector adoption velocityclaude-sonnet-53/5Academia has moderate AI tool adoption for research assistance (e.g., literature search AI), though full integration into ongoing professional development practices is still emerging.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can meaningfully assist by summarizing recent papers, identifying relevant preprints, and organizing conference schedules, but the teacher must still read critically, evaluate novelty, and participate directly in conferences, so augmentation is useful but partial rather than transformative.
Augmentation potentialclaude-sonnet-54/5AI literature summarization, alerting services, and research assistants can significantly speed up staying current with new developments, even though human interaction and synthesis remain central.
Task automatabilityclaude-haiku-4-5-202510011/5This task fundamentally requires human judgment about which developments matter, critical evaluation of advances, and authentic peer interaction. AI cannot autonomously decide what a mathematical scientist should know or substitute for collegial discussion and professional networking.
Task automatabilityclaude-sonnet-52/5AI can help surface and summarize literature, but the core task of sustained professional engagement, judgment about relevance, and networking cannot be fully offloaded to AI today.'
Adoption barriersclaude-haiku-4-5-202510013/5Professional development and staying current are expectations embedded in academic employment norms and evaluation, and peer relationships require human presence. However, there are no strict legal or licensing barriers preventing AI assistance in literature scanning and summarization.
Adoption barriersclaude-sonnet-52/5No licensing barrier exists, but tenure/promotion norms and professional culture expect genuine human engagement in conferences and collegial discussion, creating soft resistance.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI curation tools cost far less than a teacher's time to read and synthesize, but the teacher must still evaluate outputs, attend conferences in person, and engage with colleagues—making end-to-end cost savings minimal compared to the human labor involved.
Cost vs. human wageclaude-sonnet-52/5AI-assisted literature review tools are cheap, but the task also requires human conference attendance and colleague interaction, which AI cannot substitute at low cost.
Technical feasibility todayclaude-haiku-4-5-202510012/5While AI can summarize papers and curate preprints, no deployed system reliably filters mathematical literature for relevance to a specific researcher's needs, nor can it genuinely participate in conference discussions or replace human networking that builds professional judgment.
Technical feasibility todayclaude-sonnet-52/5Tools like literature summarizers and research digest apps exist but are not deployed as complete replacements for a scholar's ongoing engagement with a field.

Advise students on academic and vocational curricula and on career issues.

26

CI 23–30 · exposure 25 · augmentation 63 · importance 3.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Postsecondary institutions have been slow to adopt AI-driven advising in production. While some pilot chatbots exist, the vast majority of academic advising remains human-delivered, reflecting institutional conservatism and concern over liability in educational contexts.
Sector adoption velocityclaude-sonnet-52/5Higher education is generally slow to adopt AI for personalized advising functions, with pilots more common than production-scale deployment.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist advisors by surfacing program information, degree requirements, prerequisite sequences, and employment data, raising productivity on information-retrieval parts of the task. However, the emotional, relational, and judgment-heavy aspects of advising remain in the human domain, limiting overall augmentation impact.
Augmentation potentialclaude-sonnet-54/5AI tools can meaningfully help advisors draft degree plans, summarize career pathways, and organize information, improving efficiency while the human remains central to judgment and relationship-building.
Task automatabilityclaude-haiku-4-5-202510012/5Advising students requires understanding individual circumstances, aspirations, and constraints—tasks at which current AI has significant limitations. While AI can retrieve program information and suggest broad pathways, it cannot replicate the nuanced judgment, relationship-building, and personalized guidance that constitute the core of this task, falling well short of a 50% time-saving threshold for equal quality.
Task automatabilityclaude-sonnet-52/5Advising involves personalized judgment about a student's specific goals, transcript, and career fit, which requires context AI lacks reliable access to and cannot fully replace end-to-end.subCategory Some drafting of advice is possible but full task completion at equal quality is not yet achievable.
Adoption barriersclaude-haiku-4-5-202510014/5Strong barriers exist: institutions have duty-of-care obligations to students, advisors often hold professional credentials, and there is significant organizational and regulatory reluctance to fully automate educational guidance. Students and parents typically expect human-to-human interaction for sensitive career and academic decisions.
Adoption barriersclaude-sonnet-53/5No strict licensing requirement for academic advising, but institutional policies, liability for bad career/curriculum guidance, and student preference for human mentorship create moderate friction.
Cost vs. human wageclaude-haiku-4-5-202510012/5Current AI advisory tools require significant human oversight, customization, and integration into institutional systems. The cost of deployment, maintenance, and the necessary human validation of recommendations means AI-assisted advising is not yet cheaper than direct human advising, especially when liability is factored in.
Cost vs. human wageclaude-sonnet-52/5While AI chat costs are low, integration with student records, accreditation requirements, and error correction keep effective costs closer to human advising costs when quality is held constant.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed advising system reliably replaces human academic advisors in production at scale. Chatbots and recommendation systems exist but have narrow scope and material error rates in understanding complex student needs, career trajectories, and institutional pathways.
Technical feasibility todayclaude-sonnet-52/5Chatbots and advising tools exist in some universities as supplements, but no deployed product reliably handles full academic/career advising without human oversight at scale.

Conduct research in a particular field of knowledge and publish findings in books, professional journals, or electronic media.

26

CI 21–30 · exposure 25 · augmentation 75 · importance 3.9/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5While higher education has adopted AI writing assistants for drafting, actual research automation in academic mathematics remains negligible. Researchers use AI for ancillary tasks, but novel research output is still primarily human-driven; adoption is limited to augmentation roles.
Sector adoption velocityclaude-sonnet-52/5Academia is adopting AI tools for writing assistance and literature review at a moderate pace, but true autonomous research generation is still experimental and rare in production use within mathematics departments.
Augmentation potentialclaude-haiku-4-5-202510014/5AI significantly augments research productivity by accelerating literature review, generating code for computational work, drafting manuscript sections, and suggesting refinements—keeping the researcher in full control of conceptualization and validation. Tools like language models and code assistants demonstrably improve output velocity while the human retains epistemic authority.
Augmentation potentialclaude-sonnet-54/5AI significantly aids literature search, drafting, code/proof verification, and manuscript preparation, meaningfully boosting researcher productivity even though the core research insight remains human-driven.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist with literature review, data analysis, and drafting sections, the core task of conducting novel research in mathematical sciences requires original conceptualization, problem formulation, and creative insight that current systems cannot reliably generate. AI cannot independently design research programs or validate mathematical proofs at the level required for publishable findings.
Task automatabilityclaude-sonnet-52/5Mathematical research requires original insight, proof construction, and creativity that current AI cannot reliably generate end-to-end; AI can assist with literature review, computation, and drafting but cannot autonomously conduct novel research and publish it.
Adoption barriersclaude-haiku-4-5-202510014/5Postsecondary research publication carries strong institutional and epistemic barriers: authorship requires accountability for originality and correctness, peer review gatekeeping is universal, and academic norms assign responsibility to the human researcher for the validity of findings. Institutional policy and professional ethics currently require human researchers to be the accountable originators.
Adoption barriersclaude-sonnet-53/5No licensing barrier exists, but peer review, academic credentialing, and institutional norms around authorship and originality create moderate friction against AI-generated research being accepted as valid scholarship.
Cost vs. human wageclaude-haiku-4-5-202510011/5The cost of AI systems capable of assisting research (compute time, specialized models, human oversight for validation) combined with the unreliability of current outputs makes this more expensive than the researcher's time for the core research task itself.
Cost vs. human wageclaude-sonnet-52/5While AI tools reduce costs for literature search and formatting, the actual research and proof-generation still require expensive expert human oversight, making the all-in cost comparable to or only marginally cheaper than human-led research.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed product performs end-to-end research and publication in mathematical sciences. AI tools exist for writing assistance and figure generation, but they lack the capacity to conduct actual research, formulate novel problems, or produce original mathematical contributions that journals accept.
Technical feasibility todayclaude-sonnet-52/5Products like AI writing assistants and math-solving tools (e.g., Wolfram Alpha, GPT-based theorem helpers) exist, but no deployed system autonomously conducts original mathematical research and produces publishable findings without heavy human direction.

Prepare and deliver lectures to undergraduate or graduate students on topics such as linear algebra, differential equations, and discrete mathematics.

24

CI 18–30 · exposure 25 · augmentation 75 · importance 4.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Postsecondary math teaching remains a traditional, human-centered sector with slow AI adoption; despite ed-tech investment, universities have not deployed AI systems to replace lecturer roles in production, and cultural and regulatory inertia is high.
Sector adoption velocityclaude-sonnet-52/5Higher education adopts AI tools for tutoring and content prep but has been slow to replace core lecture delivery, given tenure structures, accreditation, and pedagogical norms.
Augmentation potentialclaude-haiku-4-5-202510014/5AI can meaningfully assist human math instructors by generating problem sets, creating visual explanations of abstract concepts, drafting lecture notes, and answering student questions asynchronously, substantially raising instructor productivity while the instructor retains pedagogical authority.
Augmentation potentialclaude-sonnet-54/5AI substantially helps with generating lecture slides, practice problems, explanations, and preparation materials for topics like linear algebra and differential equations, boosting instructor efficiency.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can generate lecture content and outlines for mathematics topics, delivering effective lectures requires real-time interaction, adaptive explanation based on student comprehension, and pedagogical judgment that current AI systems cannot reliably replicate end-to-end. AI might assist in creating materials but cannot autonomously teach a classroom at equal quality.
Task automatabilityclaude-sonnet-52/5AI can help draft lecture content and explanations, but live delivery, adapting to student questions, pacing, and classroom presence require human execution that current systems cannot fully replace end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Postsecondary education is heavily regulated and institutionally gatekept; accreditation bodies, tenure systems, and faculty governance structures legally and contractually require human instructors to deliver and sign off on curriculum. Student contact and the credentialing authority of a qualified instructor present hard barriers to automation.
Adoption barriersclaude-sonnet-53/5Accreditation and institutional norms generally require credentialed faculty to teach and interact with students, though not a strict individual licensing requirement like law or medicine.
Cost vs. human wageclaude-haiku-4-5-202510012/5An AI system capable of generating and presenting lectures would still require significant infrastructure, oversight, course design, and integration costs, likely exceeding the cost of a teaching assistant or adjunct instructor's wage when all-in expenses are considered.
Cost vs. human wageclaude-sonnet-52/5While AI content generation is cheap, actual lecture delivery still requires paid faculty or substantial human oversight, so total cost savings are modest rather than order-of-magnitude.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed product reliably delivers full lectures to university students with demonstrated effectiveness comparable to human instructors. AI can generate notes or answer questions in narrow contexts, but production systems do not replace postsecondary math instruction at scale.
Technical feasibility todayclaude-sonnet-52/5Products like AI tutors and content generators exist for math topics, but no deployed product autonomously delivers full undergraduate/graduate lecture courses in production at scale.

Initiate, facilitate, and moderate classroom discussions.

13

CI 0–25 · exposure 8 · augmentation 38 · importance 4.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Educational institutions remain among the slowest sectors to adopt AI for core instructional tasks; classroom discussion facilitation is viewed as a uniquely human pedagogical function, and adoption of AI in live discussion settings is nearly absent.
Sector adoption velocityclaude-sonnet-52/5Higher education is adopting AI tools for grading and content generation, but live classroom facilitation remains largely untouched, with adoption concentrated in supplementary rather than core teaching functions.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could marginally assist by suggesting discussion topics or generating follow-up questions for an instructor to review, but it offers limited real-time augmentation during live moderation since the instructor must remain fully attentive and responsive to student dynamics.
Augmentation potentialclaude-sonnet-53/5AI can help instructors prepare discussion questions, summarize student responses, or provide real-time polling/analytics, meaningfully supporting but not replacing the live facilitation role.
Task automatabilityclaude-haiku-4-5-202510011/5Facilitating and moderating live classroom discussions requires real-time social intelligence, emotional awareness, and dynamic responsiveness to student contributions that current AI systems cannot reliably provide at equal quality. The core value lies in human judgment about when to redirect, validate, encourage, or challenge—skills deeply dependent on contextual understanding and interpersonal presence.
Task automatabilityclaude-sonnet-52/5AI can generate discussion prompts or simulate Q&A, but live in-person facilitation of dynamic classroom discussion, reading student engagement, and adaptive moderation requires real-time social and pedagogical judgment beyond current systems.
Adoption barriersclaude-haiku-4-5-202510015/5Teaching positions carrying classroom instruction are legally and institutionally protected roles requiring credentialed faculty oversight. Educational accreditation standards, institutional policies, and student/parent expectations all mandate human faculty leadership of academic discussions; automated moderation would face severe organizational and regulatory friction.
Adoption barriersclaude-sonnet-53/5No formal licensing barrier bars AI from assisting, but strong human-contact expectations, pedagogical norms, and institutional accreditation standards favor human-led classroom facilitation.
Cost vs. human wageclaude-haiku-4-5-202510011/5The cost of AI systems to facilitate discussions (including oversight, prompt engineering, and intervention correction) would exceed the loaded cost of an instructor's time, since the human educator is already present and paid to teach anyway.
Cost vs. human wageclaude-sonnet-52/5Since AI cannot yet perform this task independently, any cost comparison requires heavy human oversight, making the all-in cost comparable to or higher than a human instructor for this specific task.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product reliably manages live classroom discussion moderation as a primary task. While AI can draft discussion prompts or suggest topics offline, the interactive, real-time facilitation of academic debate among students requires human presence and judgment that current systems have not achieved in production educational settings.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously runs live classroom discussions for postsecondary math courses; existing tools are limited to chatbots or asynchronous forums, not real-time in-person moderation.

Maintain regularly scheduled office hours to advise and assist students.

8

CI 0–16 · exposure 5 · augmentation 63 · importance 4.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Postsecondary institutions remain traditional in their staffing and advising models; adoption of AI for core faculty advising duties is minimal and encountering institutional resistance.
Sector adoption velocityclaude-sonnet-52/5Higher education is a moderate-to-slow adopter of AI for direct student advising displacement, though AI tutoring aids are increasingly piloted alongside human office hours.
Augmentation potentialclaude-haiku-4-5-202510013/5AI tools can assist faculty by preparing student records, suggesting topic structures, or drafting follow-up resources, but the core advising interaction remains human-led and judgment-intensive.
Augmentation potentialclaude-sonnet-54/5AI tools (chatbots, math solvers, tutoring systems) can meaningfully assist students with questions outside office hours and help faculty prepare materials or triage common questions, augmenting the advising process.
Task automatabilityclaude-haiku-4-5-202510011/5Student advising requires responsive, contextual interaction addressing individual concerns, emotional support, and judgment-based guidance—capabilities that current AI cannot reliably replicate in real-time, synchronous office settings without human presence.
Task automatabilityclaude-sonnet-51/5This task inherently requires a live, present human to hold scheduled office hours and provide personalized advising; AI cannot occupy office hours or fulfill the institutional obligation of faculty presence.'
Adoption barriersclaude-haiku-4-5-202510015/5Educational institutions have strong cultural and sometimes regulatory expectations that faculty maintain direct human contact with students; many accreditation standards and student success metrics depend on documented faculty-student interaction.
Adoption barriersclaude-sonnet-54/5Office hours are often a contractual/institutional requirement tied to faculty employment and accreditation expectations of student-faculty interaction, creating strong organizational barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510011/5Deploying AI to replace office hours would require supervision and oversight that approaches or exceeds the cost of the human instructor simply holding the office hours themselves.
Cost vs. human wageclaude-sonnet-52/5While AI tutoring tools are cheap per query, they don't replace the human obligation of holding office hours, so cost comparison for the actual task is not favorable to full substitution.
Technical feasibility todayclaude-haiku-4-5-202510012/5While chatbots can answer routine math questions, no deployed product reliably handles the full scope of office hours: complex problem-solving, personalized academic planning, pastoral care, and the implicit human mentorship that defines the interaction.
Technical feasibility todayclaude-sonnet-51/5No deployed product substitutes for a professor's scheduled office hours as an institutional/administrative requirement, though chatbots exist for supplemental Q&A.

Participate in campus and community events.

5

CI 5–5 · exposure 0 · augmentation 25 · importance 2.7/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5There is no adoption pattern for automating faculty participation in events because the core value proposition (human presence, relationship-building, institutional representation) is inherently human-centric.
Sector adoption velocityclaude-sonnet-51/5Higher education is a slow-adopting sector for AI in general, and this specific interpersonal/physical task shows essentially no automation trend.
Augmentation potentialclaude-haiku-4-5-202510012/5AI might assist with event scheduling, itinerary planning, or follow-up communication, but these are peripheral to the task itself and offer limited productivity enhancement for the core participation activity.
Augmentation potentialclaude-sonnet-52/5AI can help with scheduling, event planning logistics, or drafting communications related to events, but offers little assistance for the actual participation itself.
Task automatabilityclaude-haiku-4-5-202510011/5Campus and community event participation is fundamentally a human-presence and relationship task requiring interpersonal interaction, networking, and contextual judgment that current AI cannot meaningfully substitute.
Task automatabilityclaude-sonnet-51/5Participating in physical campus/community events (attending, networking, representing the department) requires embodied presence and social judgment that current AI cannot perform end-to-end.'
Adoption barriersclaude-haiku-4-5-202510014/5Institutional and community expectations, professional norms, and the nature of academic service roles all mandate human presence; events specifically exist to foster human relationships and institutional visibility that cannot be delegated to AI.
Adoption barriersclaude-sonnet-54/5Community and campus engagement inherently requires human presence and relationship-building, and institutions expect faculty representation, creating strong structural barriers to automation.
Cost vs. human wageclaude-haiku-4-5-202510011/5AI involvement would require human attendance anyway, making any AI tool purely supplementary (e.g., scheduling assistance) and thus more expensive than the human performing the task directly.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute, so any AI cost is irrelevant; the human must attend, making AI more expensive/impossible as a replacement.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI system can autonomously attend, engage in, or represent a faculty member at campus or community events; this requires embodied presence and genuine human participation.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product substitutes for a faculty member's physical/social presence at events; this is not something products are built to do.

Collaborate with colleagues to address teaching and research issues.

4

CI 0–7 · exposure 0 · augmentation 38 · importance 3.6/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5There is virtually no evidence of AI adoption for genuine peer-level academic collaboration. Even digitally advanced institutions rely on human-to-human meetings and deliberation for research and teaching strategy.
Sector adoption velocityclaude-sonnet-52/5Higher education is a relatively slow-adopting sector for AI in core faculty governance and collegial processes, though some administrative tools are creeping in elsewhere.
Augmentation potentialclaude-haiku-4-5-202510012/5AI might assist by organizing discussion points or summarizing prior conversations, but the core value of collaboration—interactive problem-solving, trust-building, and mutual accountability—cannot be meaningfully augmented by current systems.
Augmentation potentialclaude-sonnet-53/5AI tools can help colleagues prepare materials, summarize research, draft proposals, or organize meeting notes, providing moderate support to the collaborative process itself.
Task automatabilityclaude-haiku-4-5-202510011/5This task fundamentally requires interpersonal negotiation, judgment about research and pedagogical strategy, and contextual understanding of individual colleagues' expertise and preferences. Current AI cannot meaningfully participate in genuine collaboration or reach consensual decisions with humans on complex professional matters.
Task automatabilityclaude-sonnet-51/5This task is inherently interpersonal and collaborative, requiring relationship-building, negotiation, and shared decision-making among colleagues that current AI cannot substitute for end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Academic governance and collegiality are deeply rooted in human judgment and professional accountability. Colleagues must take personal responsibility for research and teaching decisions, and institutional norms require human deliberation on these matters.
Adoption barriersclaude-sonnet-54/5Departmental governance, tenure/promotion processes, and academic norms require human faculty to jointly deliberate and decide on curriculum and research matters, creating strong organizational and cultural barriers to automation.
Cost vs. human wageclaude-haiku-4-5-202510011/5Even if partial automation were possible (e.g., AI drafting collaborative notes), the overhead of oversight and human validation would exceed the time saved, making this more expensive than direct human collaboration.
Cost vs. human wageclaude-sonnet-51/5There is no AI product performing this task, so no cost comparison favors AI; the human cost is the only real option.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI product performs genuine peer-level collaboration on teaching and research strategy. AI can summarize discussions or suggest ideas, but actual back-and-forth negotiation and consensus-building among academic colleagues remains beyond current system capabilities.
Technical feasibility todayclaude-sonnet-51/5No deployed product performs collegial collaboration on teaching or research issues; this remains a human social and professional interaction.

Supervise undergraduate or graduate teaching, internship, and research work.

3

CI 0–5 · exposure 0 · augmentation 50 · importance 3.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Adoption of AI for core academic supervision remains negligible; institutional, legal, and cultural norms strongly favor human faculty in direct supervisory roles, and sector digitization does not change the legal requirement for faculty oversight.
Sector adoption velocityclaude-sonnet-51/5Higher education adoption of AI for supervisory/mentorship roles is minimal; academia is slow-moving on core mentorship functions despite AI use in adjacent research tasks.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist faculty with administrative tasks like scheduling, progress tracking, and providing preliminary feedback on student work, moderately raising productivity in the supervision workflow while faculty retain core mentoring and accountability.
Augmentation potentialclaude-sonnet-53/5AI can help with tracking progress, generating feedback drafts, or analyzing research data, providing moderate assistance while the supervisory relationship itself remains human-led.
Task automatabilityclaude-haiku-4-5-202510011/5Supervising teaching, internship, and research work requires ongoing human judgment, mentorship, relationship-building, and accountability that cannot be automated end-to-end. AI cannot replicate the interpersonal guidance, career development, and institutional responsibility inherent to academic supervision.
Task automatabilityclaude-sonnet-51/5Supervision involves mentoring, evaluating student progress, giving personalized feedback, and building relationships that require sustained human judgment and accountability AI cannot replicate end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Academic institutions legally and contractually require licensed faculty or authorized humans to supervise students, manage internships, and oversee research. Institutional accreditation, duty of care, and faculty employment agreements all mandate human supervision.
Adoption barriersclaude-sonnet-54/5Institutional accreditation, faculty credentialing, and mentorship/oversight responsibilities (e.g., signing off on theses, evaluating interns) require a qualified human academic, creating strong structural barriers.
Cost vs. human wageclaude-haiku-4-5-202510011/5The cost of AI systems capable of even partial supervision assistance would not undercut the loaded wage of a postsecondary faculty member who bears legal and institutional responsibility for student supervision.
Cost vs. human wageclaude-sonnet-51/5Since no AI product substitutes for this supervisory role, there is no viable AI cost basis to compare against the human faculty cost.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product reliably performs academic supervision as a standalone function today. While AI can assist with scheduling or grading papers, actually supervising students' intellectual and professional development requires human presence and accountability that current systems cannot assume.
Technical feasibility todayclaude-sonnet-51/5No deployed product performs the supervisory role of overseeing student teaching, internships, or research; this remains a human-only function in academic practice.

Act as advisers to student organizations.

3

CI 0–5 · exposure 0 · augmentation 25 · importance 3.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Universities are slow to digitize student-facing advisory roles; adoption of AI for core advising functions remains minimal, and institutions prioritize human-led mentorship in this domain.
Sector adoption velocityclaude-sonnet-51/5Higher education advising roles show minimal AI adoption; this is a low-digitization, relationship-driven task with no evidence of displacement.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could assist with logistical tasks (scheduling, document drafting, information retrieval) but offers limited value in the core advisory functions of mentoring, relationship-building, and judgment that define this role.
Augmentation potentialclaude-sonnet-52/5AI could help with scheduling, drafting communications, or budget tracking for the organization, but offers little assistance for the core mentoring and advisory relationship.
Task automatabilityclaude-haiku-4-5-202510011/5Advising student organizations requires interpersonal judgment, mentorship, conflict resolution, and understanding of individual student needs and organizational dynamics—tasks that current AI cannot meaningfully automate end-to-end.
Task automatabilityclaude-sonnet-51/5Advising student organizations requires relational trust, mentorship, in-person presence at events, and institutional judgment that current AI cannot perform end-to-end.'
Adoption barriersclaude-haiku-4-5-202510015/5Student advisors are typically required by institutional policy; universities mandate human advisors for student organizations, and there is strong organizational and cultural expectation that a human holds this role.
Adoption barriersclaude-sonnet-54/5Institutions typically require a designated faculty/staff member to formally serve as advisor for liability, oversight, and accreditation purposes, creating strong organizational and administrative barriers to substitution.
Cost vs. human wageclaude-haiku-4-5-202510011/5Even if a narrow component (e.g., scheduling) could be partially automated, the labor cost of a human advisor is modest relative to any AI system that would need to replicate genuine mentorship and organizational oversight.
Cost vs. human wageclaude-sonnet-51/5There is no AI product performing this role, so no cost comparison favors AI; the task is bundled into faculty duties with no separate automation cost saved.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product reliably performs the full advisory role for student organizations; this requires sustained relationship-building, contextual understanding, and responsive guidance that exceeds current AI capabilities.
Technical feasibility todayclaude-sonnet-51/5No deployed product substitutes for a human faculty advisor role in student organizations; this remains a purely human, relationship-based responsibility.

Perform administrative duties, such as serving as department head.

1

CI 0–3 · exposure 0 · augmentation 38 · importance 3.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5There is zero adoption of AI as autonomous department heads in higher education. The role is inherently human-bound by governance structure and regulatory requirement.
Sector adoption velocityclaude-sonnet-52/5Higher education administration adopts AI slowly for governance and leadership roles, though some AI tools are used for scheduling, reporting, and communications support.
Augmentation potentialclaude-haiku-4-5-202510012/5AI can assist a department head with scheduling, data synthesis, or document drafting, but these are minor aids to a role whose core (hiring, discipline, strategy, advocacy) remains human-centered and irreducibly collaborative.
Augmentation potentialclaude-sonnet-53/5AI can help draft reports, summarize meeting notes, manage schedules, and analyze budget data, meaningfully assisting a department head's administrative workload.
Task automatabilityclaude-haiku-4-5-202510011/5Department headship involves strategic decision-making, personnel management, budget oversight, and stakeholder negotiation that require contextual judgment, emotional intelligence, and accountability. Current AI cannot autonomously perform these fiduciary and leadership functions.
Task automatabilityclaude-sonnet-51/5Department head duties involve interpersonal leadership, personnel decisions, budget authority, and institutional politics that require human judgment, accountability, and relationship management that AI cannot perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Legal and institutional barriers are absolute: universities require a human department head with legal standing to sign hiring recommendations, approve budgets, represent the department, and assume fiduciary liability. Authority and accountability cannot be delegated to AI.
Adoption barriersclaude-sonnet-55/5Department head is a formal institutional appointment requiring faculty governance, tenure status, and legal signing authority over budgets, hiring, and personnel—hard organizational and often contractual barriers exist.
Cost vs. human wageclaude-haiku-4-5-202510011/5A department head is a senior salaried role commanding a six-figure loaded cost. Current AI cannot replace this cost-effectively because it cannot manage people, budgets, or institutional relationships independently.
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute for the actual role, so any comparison is moot—human labor is the only viable option and AI adds cost as a support tool rather than replacing the function.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product performs department headship at scale. Administrative support tools exist (scheduling, document drafting), but they address only fragments of the task, not end-to-end leadership.
Technical feasibility todayclaude-sonnet-51/5No deployed product performs the role of an academic department head; AI is at best a scheduling or drafting aid for a human occupying this position.

Serve on academic or administrative committees that deal with institutional policies, departmental matters, and academic issues.

0

CI 0–0 · exposure 0 · augmentation 38 · importance 3.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5There is no meaningful adoption of AI to replace faculty on committees because the task is fundamentally non-automatable and legally restricted to humans in formal roles.
Sector adoption velocityclaude-sonnet-51/5Higher education governance structures are slow-moving and highly resistant to structural change; there is no observed trend of AI systems participating in committee governance.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could offer limited assistance by generating background briefings, summarizing policy documents, or preparing agenda analysis before meetings, but these are peripheral to the core committee work of deliberation and decision-making that remains fully human.
Augmentation potentialclaude-sonnet-53/5AI can help prepare briefing materials, summarize policy documents, draft meeting agendas/minutes, or analyze data for committee discussions, moderately aiding preparation and follow-through.
Task automatabilityclaude-haiku-4-5-202510011/5Committee service requires deliberation, consensus-building, judgment on institutional policy, and accountability for decisions that affect people. AI cannot meaningfully participate in these human governance functions or bear responsibility for outcomes.
Task automatabilityclaude-sonnet-51/5Committee service requires human presence, judgment, negotiation, and institutional relationship-building that AI cannot substitute for; no end-to-end automation is plausible today.
Adoption barriersclaude-haiku-4-5-202510015/5Strong legal and institutional barriers exist: only designated faculty members with formal appointment authority can serve on committees; institutional governance structures legally require human signatories and decision-makers with fiduciary responsibility.
Adoption barriersclaude-sonnet-55/5Institutional governance, shared governance norms, tenure/faculty status requirements, and accreditation standards typically require actual faculty members to serve, creating hard structural barriers.
Cost vs. human wageclaude-haiku-4-5-202510011/5AI systems have no meaningful cost advantage here because the task cannot be automated at all—a human must participate, and automating attendance or voting is not feasible or legally permissible.
Cost vs. human wageclaude-sonnet-51/5There is no AI product performing this role, so any comparison favors the human by default; AI cannot substitute for the seat itself.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product can attend committee meetings, engage in debate, build consensus, or vote on institutional policies. This task fundamentally requires human participation and judgment.
Technical feasibility todayclaude-sonnet-51/5No deployed product performs committee membership or governance participation on behalf of a faculty member; this is not a task AI systems attempt to fulfill.

Related occupations — Educational Instruction & Library

How to read this

A high substitution score does not mean this job disappears — it means a large share of its current tasks face replacement pressure, so the mix of tasks is likely to change. High augmentation alongside substitution typically means the occupation reorganizes around the protected tasks. Wide confidence intervals mean the rater panel disagreed: treat those scores as open questions, not verdicts.

What would change this score

New model capabilities (automatability, feasibility), falling inference costs (cost ratio), regulation and licensing shifts (barriers), and measured sector adoption (velocity) all re-enter at every index release. Each release is recomputed, versioned and kept queryable — scores are claims with a date on them, not permanent labels.