Education Teachers, Postsecondary

25-1081.00

Median wage $75,350/yr60,830 employed (US)Rank #349 of 923 scored · top 38% by substitution

Teach courses pertaining to education, such as counseling, curriculum, guidance, instruction, teacher education, and teaching English as a second language. 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

Substitution32

Exposure29

Augmentation64

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

24 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

13%

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%30

panel mean rating 2.2/5 → substitution pressure 30/100

Technical feasibility todayw 20%29

panel mean rating 2.1/5 → substitution pressure 29/100

Cost vs. human wagew 15%36

panel mean rating 2.4/5 → substitution pressure 36/100

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%30

panel mean rating 2.2/5 → substitution pressure 30/100

Task breakdown (24 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.

CI 90–95 · exposure 100 · augmentation 75 · importance 3.7/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	5/5	Postsecondary institutions have already widely adopted LMS and SIS platforms; automated record-keeping is now standard practice in higher education, with deep penetration in both community colleges and universities.
Sector adoption velocity	claude-sonnet-5	5/5	Higher education has near-universal adoption of digital LMS platforms that automate these recordkeeping functions already.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI augments faculty productivity by auto-populating records from multiple sources, generating statistical summaries of class performance, and alerting instructors to attendance patterns or grade outliers, while instructors retain final authority over accuracy and decisions.
Augmentation potential	claude-sonnet-5	4/5	AI-enhanced gradebooks and LMS analytics significantly speed up tracking, flagging at-risk students and automating calculations while instructors retain oversight.
Task automatability	claude-haiku-4-5-20251001	5/5	Record-keeping for attendance, grades, and student data is entirely routine and already highly automated in learning management systems (LMS) like Canvas, Blackboard, and Google Classroom. AI can extract data, verify completeness, flag discrepancies, and generate reports with >50% time savings at equal quality.
Task automatability	claude-sonnet-5	5/5	Recording attendance and grades is a structured data-entry task easily handled by LMS software and AI-enabled gradebook tools with minimal human oversight beyond initial data input.
Adoption barriers	claude-haiku-4-5-20251001	3/5	Some institutional friction exists: records must meet FERPA compliance, institutional integrations are required, and faculty often prefer direct control of grading. However, no legal licensing bar prevents automation, and adoption is widespread.
Adoption barriers	claude-sonnet-5	2/5	Some institutional policies require instructor verification of final grades, but the mechanical recordkeeping itself carries no licensing or liability barrier.
Cost vs. human wage	claude-haiku-4-5-20251001	5/5	LMS and SIS integration costs are fixed infrastructure expenses; marginal cost per record is negligible. Automated systems cost orders of magnitude less than dedicated clerical staff per transaction, with minimal ongoing oversight.
Cost vs. human wage	claude-sonnet-5	5/5	Automated recordkeeping software costs a small fraction of the faculty time it would take to manually track and record attendance and grades.
Technical feasibility today	claude-haiku-4-5-20251001	5/5	Multiple mature, production-deployed systems (LMS platforms, student information systems, automated grade-tracking tools) perform this task reliably at scale in institutions worldwide. Automated attendance tracking and grade recording are industry standard.
Technical feasibility today	claude-sonnet-5	5/5	Learning management systems (Canvas, Blackboard, Moodle) already automate attendance tracking, grade calculation, and recordkeeping reliably at scale in production across universities.

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

CI 76–76 · exposure 75 · augmentation 100 · importance 4.3/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	3/5	Adoption is moderate: many universities and online platforms experiment with AI drafting tools, but widespread production use remains limited. Faculty adoption varies widely by discipline and institution; pilots are common, but systemic replacement has not yet become standard practice.
Sector adoption velocity	claude-sonnet-5	3/5	Higher education is adopting generative AI for course prep at a moderate pace, with many pilots and growing informal use but no uniform institutional mandate yet.
Augmentation potential	claude-haiku-4-5-20251001	5/5	AI provides transformative assistance for course material preparation: instructors use AI to draft syllabi, generate diverse problem sets, and create multiple assignment variants, while retaining full control over pedagogical intent, learning objectives, and quality assurance. This is a textbook case of augmentation raising productivity.
Augmentation potential	claude-sonnet-5	5/5	AI strongly augments instructors by rapidly producing drafts of assignments, handouts, and syllabi that faculty then refine, saving significant preparation time.
Task automatability	claude-haiku-4-5-20251001	4/5	Current AI systems can generate syllabi, assignments, and handouts with high quality and significant time savings using prompt engineering and templates. While customization to institutional standards and learning objectives requires human oversight, the bulk of content creation can be automated, easily meeting the 50% time-saving threshold.
Task automatability	claude-sonnet-5	4/5	LLMs can draft syllabi, homework assignments, and handouts from a course description or learning objectives with substantial time savings, though instructor review/customization is typically needed.
Adoption barriers	claude-haiku-4-5-20251001	2/5	There are no legal or licensing barriers preventing AI use for preparing course materials. The main friction is institutional preference for human-authored content and faculty concerns about educational authenticity, but these are soft, not hard, barriers to substitution.
Adoption barriers	claude-sonnet-5	2/5	No licensing requirement dictates who writes course materials, though institutional policies, accreditation standards, and academic freedom norms create some review friction.
Cost vs. human wage	claude-haiku-4-5-20251001	5/5	AI inference costs for generating course materials are minimal (often under $1 per syllabus or assignment set), while instructor time typically costs $50–150+ per hour. The cost ratio strongly favors automation by an order of magnitude or more.
Cost vs. human wage	claude-sonnet-5	5/5	Generating a draft syllabus or handout via AI costs cents to a few dollars in compute versus hours of faculty time at academic wages.
Technical feasibility today	claude-haiku-4-5-20251001	4/5	LLMs like GPT-4, Claude, and specialized educational tools reliably generate course materials in production settings. Multiple universities and platforms deploy AI-assisted document generation; output quality is sufficient for immediate use or minor human refinement, though some institutions still require human review before deployment.
Technical feasibility today	claude-sonnet-5	4/5	Widely deployed tools (ChatGPT, Copilot, dedicated ed-tech platforms) are routinely used by instructors today to generate these materials at acceptable quality with light editing.

Compile bibliographies of specialized materials for outside reading assignments.

CI 67–81 · exposure 70 · augmentation 100 · importance 2.8/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	3/5	Higher education institutions are experimenting with AI tools for administrative and instructional support, including citation assistance, but widespread production use for bibliography compilation remains nascent. Adoption is faster in digitally mature institutions but lags overall sector penetration.
Sector adoption velocity	claude-sonnet-5	3/5	Higher education is a professional/information-sector context with growing but uneven adoption of AI research tools; many faculty still compile reading lists manually or with traditional library tools.
Augmentation potential	claude-haiku-4-5-20251001	5/5	AI powerfully augments faculty by rapidly generating initial bibliography drafts, suggesting interdisciplinary sources, and automating formatting, allowing educators to focus on curation, relevance judgment, and pedagogical integration. The human remains in control while productivity rises substantially.
Augmentation potential	claude-sonnet-5	5/5	AI tools substantially speed up literature discovery, summarization, and citation formatting, letting instructors focus on curating quality and relevance rather than manual searching.
Task automatability	claude-haiku-4-5-20251001	4/5	AI can reliably search academic databases, compile citations, format bibliographies, and filter for reading level with minimal human intervention. While verification of relevance and discipline-specific curation may require oversight, the core compilation task easily exceeds 50% time savings at comparable quality.
Task automatability	claude-sonnet-5	4/5	Compiling bibliographies of specialized materials is largely a search, curation, and formatting task that current AI (with search/retrieval tools) can do quickly, though a professor still needs to verify relevance and accuracy for their specific course.
Adoption barriers	claude-haiku-4-5-20251001	2/5	No regulatory or licensing requirement mandates human authorship of bibliographies. Institutional adoption friction is minimal; faculty autonomy and preference for personalized curation provide some resistance, but no legal or structural barrier prevents substitution.
Adoption barriers	claude-sonnet-5	1/5	There is no licensing, liability, or regulatory requirement that a bibliography be compiled specifically by a human instructor.
Cost vs. human wage	claude-haiku-4-5-20251001	5/5	AI inference cost for generating a comprehensive bibliography is typically a few cents to dollars, far below the 30–60 minutes a faculty member would spend manually searching and formatting (loaded wage $50–100/hour). The cost ratio favors automation by an order of magnitude or more.
Cost vs. human wage	claude-sonnet-5	4/5	An AI-assisted bibliography compilation costs a small fraction of the faculty time it would otherwise take, even accounting for verification overhead.
Technical feasibility today	claude-haiku-4-5-20251001	4/5	Deployed AI tools (ChatGPT, Claude, specialized citation managers with AI, Google Scholar integration) demonstrably perform bibliography compilation today. Products exist in production, though some require human review for accuracy and disciplinary fit; minor error rates on obscure materials remain.
Technical feasibility today	claude-sonnet-5	3/5	AI research assistants and citation tools (e.g., research databases with AI search, reference managers with AI suggestions) exist and are used, but hallucinated citations and gaps in specialized/niche literature coverage remain a real production issue requiring human verification.

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

CI 62–75 · exposure 70 · augmentation 100 · importance 4.3/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	4/5	Higher education has rapidly adopted online exam platforms and learning management systems with autograding features; many institutions now use AI-assisted exam generation and proctoring in production. Adoption is especially fast in large, digitized universities and online-first programs.
Sector adoption velocity	claude-sonnet-5	3/5	Higher education has adopted AI grading and exam-generation tools unevenly—common in large intro courses and some LMS platforms, but many institutions and disciplines still rely heavily on manual processes.
Augmentation potential	claude-haiku-4-5-20251001	5/5	AI significantly augments faculty productivity by generating question banks, formatting exams, instantly scoring objective items, and flagging outlier essays for review. Instructors maintain control over assessment design and subjective judgment while offloading mechanical labor.
Augmentation potential	claude-sonnet-5	5/5	AI substantially speeds up exam creation, question variation, rubric-based grading, and feedback generation, making it a strong productivity aid for instructors and TAs.
Task automatability	claude-haiku-4-5-20251001	4/5	AI systems can compile exam questions from course materials, generate and administer online assessments, and automatically grade objective items (multiple choice, short answer matching patterns) with high accuracy. Subjective grading (essays) remains challenging, but 50% time savings are achievable for mixed-item exams through automation of logistics and objective scoring.
Task automatability	claude-sonnet-5	4/5	Compiling questions, generating exams from item banks, and grading objective or short-answer responses via AI is well within current capability; grading complex essays or judging nuanced understanding still needs human oversight but the bulk of the task can be automated with existing tools.
Adoption barriers	claude-haiku-4-5-20251001	2/5	Few legal barriers exist to automating exam administration and objective grading; institutions already delegate grading to TAs and software. Main friction is institutional inertia, faculty preference for human control over assessment, and accreditation expectations around instructor oversight—not regulatory prohibition.
Adoption barriers	claude-sonnet-5	3/5	Faculty typically retain final authority over grading and academic integrity, and institutional policies often require instructor sign-off, creating moderate friction even though no strict licensing barrier exists.
Cost vs. human wage	claude-haiku-4-5-20251001	4/5	AI-driven exam compilation and objective grading cost significantly less than instructor time (minutes of compute vs. hours of faculty labor). Subjective grading oversight is still needed, but the cost ratio for the automatable portion (generation, administration, objective scoring) favors AI by an order of magnitude or more.
Cost vs. human wage	claude-sonnet-5	4/5	Automated exam generation and grading tools are inexpensive per use compared to faculty or TA time, especially for large classes with objective-format assessments.
Technical feasibility today	claude-haiku-4-5-20251001	4/5	Mature LMS platforms (Canvas, Blackboard, Brightspace) with AI-assisted grading plugins and automated test generation are in widespread production use in higher education. Learning Analytics Engines and exam platforms (e.g., Examplify, ProctorU with AI proctoring) handle administration reliably, though essay grading still requires human oversight.
Technical feasibility today	claude-sonnet-5	3/5	Products like AI-assisted quiz generators and automated grading systems (e.g., Gradescope, LMS auto-graders, AI essay scorers) are deployed in production, but reliability for open-ended or high-stakes assessment is still limited and often requires instructor review.

Select and obtain materials and supplies, such as textbooks.

CI 51–65 · exposure 58 · augmentation 75 · importance 3.4/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Higher education has been slow to adopt AI-driven procurement and material selection tools; most institutions still rely on manual faculty selection processes. While administrative systems are digitizing, automation of pedagogical material choice remains limited in practice.
Sector adoption velocity	claude-sonnet-5	2/5	Higher education administrative processes adopt AI tools slowly outside of grading/writing support; textbook selection remains a mostly manual, low-digitization process tied to procurement systems.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI tools can significantly assist faculty by rapidly identifying candidate textbooks, comparing prices and editions, and flagging materials that match learning outcomes, allowing instructors to focus on evaluative judgment rather than search legwork.
Augmentation potential	claude-sonnet-5	4/5	AI can efficiently suggest textbooks, summarize reviews, compare editions/costs, and align materials with learning objectives, meaningfully speeding up faculty decision-making while they retain final judgment.
Task automatability	claude-haiku-4-5-20251001	3/5	AI can partially automate this task by identifying suitable textbooks through search and comparison (e.g., matching curriculum requirements to available materials), but the final selection typically requires human judgment about pedagogical fit, budget constraints, and institutional preferences. End-to-end automation with 50% time savings is achievable for routine reordering of established materials but not for novel course design.
Task automatability	claude-sonnet-5	4/5	AI can research, compare, and recommend textbooks/materials based on course objectives, syllabi, and budget constraints, largely automating the selection research process, though final approval and procurement logistics may need human action.
Adoption barriers	claude-haiku-4-5-20251001	3/5	Some institutional procurement policies require human authorization and budget sign-off, and textbook selection often involves departmental or accreditation standards that mandate human review. However, there are no strict legal barriers preventing automation of material identification and ordering.
Adoption barriers	claude-sonnet-5	2/5	Some institutional procurement policies and budget authorization require faculty or administrative sign-off, but there's no licensing or legal requirement that a human personally select materials.
Cost vs. human wage	claude-haiku-4-5-20251001	4/5	AI-assisted procurement platforms and automated search/comparison tools are relatively inexpensive to operate compared to the faculty time saved in material research and selection, especially when integrated into existing institutional systems.
Cost vs. human wage	claude-sonnet-5	4/5	AI-assisted research and comparison of textbooks/materials is far cheaper than faculty time spent browsing catalogs, though actual purchasing/procurement still involves institutional systems and staff.
Technical feasibility today	claude-haiku-4-5-20251001	3/5	Procurement systems and some course management platforms offer semi-automated material suggestion and ordering features, but they operate with narrow scope and require significant human oversight to ensure appropriateness. Deployed products exist but typically function as assistants rather than autonomous agents.
Technical feasibility today	claude-sonnet-5	3/5	Tools like AI search assistants and curriculum-recommendation platforms exist and are used for material discovery, but no mature end-to-end product handles selection plus procurement (ordering, licensing, budget approval) reliably in production.

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

CI 51–54 · exposure 50 · augmentation 75 · importance 4.1/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Postsecondary institutions, especially traditional universities, lag in AI adoption for grading. Most institutions remain cautious, pilots are common but production deployment is rare, and cultural resistance from faculty is substantial. Adoption is slower than in high-digitization sectors.
Sector adoption velocity	claude-sonnet-5	3/5	Higher education is adopting AI grading and feedback tools steadily, particularly in STEM and writing-intensive courses, but adoption is uneven across disciplines and institutions with pilots more common than full-scale deployment.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI augmentation is already demonstrable: systems pre-score assignments, flag plagiarism, generate initial feedback, and free instructors to focus on high-order assessment and individualized feedback. This transforms instructor productivity on the mechanical aspects while instructors retain final judgment and personalization.
Augmentation potential	claude-sonnet-5	4/5	AI substantially speeds up grading and feedback generation, allowing instructors to review and finalize rather than grade from scratch, which is a major productivity gain while keeping the instructor in the loop.
Task automatability	claude-haiku-4-5-20251001	3/5	AI can automatically score objective assignments (multiple choice, fill-in-the-blank, formulaic calculations) and generate initial feedback on essays via NLP analysis of grammar, structure, and content coherence. However, evaluating higher-order thinking, nuanced argument quality, and discipline-specific rigor typically requires human judgment, meaning roughly half the work can be automated with setup.
Task automatability	claude-sonnet-5	3/5	AI can draft grades and feedback for structured assignments (essays, short answers, code) with substantial time savings, but nuanced grading against rubrics, academic integrity judgment, and final grade responsibility still require human review, especially for advanced postsecondary work.
Adoption barriers	claude-haiku-4-5-20251001	3/5	No legal license is required, but significant institutional and human barriers exist: faculty ownership of grading, student expectations for personalized feedback, institutional liability concerns over unfair grading, and faculty union agreements that may protect grading as core work. These create adoption friction despite technical feasibility.
Adoption barriers	claude-sonnet-5	3/5	No licensing requirement bars AI from assisting, but institutional academic integrity policies, grade appeal processes, and instructor-of-record accountability create meaningful friction against full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	4/5	Inference costs for AI grading (especially at scale across a semester) are very low—cents per assignment—compared to instructor labor (often $30–80/hour fully loaded). Even with oversight and manual review of edge cases, the cost ratio heavily favors AI.
Cost vs. human wage	claude-sonnet-5	4/5	AI-assisted grading tools are inexpensive per assignment compared to instructor or TA time, especially at scale for large courses, though integration and oversight add some cost.
Technical feasibility today	claude-haiku-4-5-20251001	3/5	Products like auto-grading systems, plagiarism detectors, and essay-scoring models (e.g., via LLMs) exist in production at some institutions. However, error rates remain material for complex assignments, and institutional adoption is inconsistent; most still rely on human grading, with AI used only for preliminary flagging or low-stakes tasks.
Technical feasibility today	claude-sonnet-5	3/5	Products like Gradescope, Turnitin, and LLM-based grading assistants are deployed in real courses, but accuracy varies by discipline and assignment complexity, and most institutions still require instructor sign-off on final grades.

Write grant proposals to procure external research funding.

CI 30–55 · exposure 38 · augmentation 88 · importance 3.2/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Higher education is moderately digitized but adoption of AI for grant-writing is nascent and cautious. Most institutions treat AI as an optional drafting tool, not a replacement. Adoption is slower than in pure information work because grant outcomes have immediate, measurable consequences (funding success/failure).
Sector adoption velocity	claude-sonnet-5	3/5	Higher education is moderately adopting AI writing tools for grant drafting, but institutional caution, integrity policies, and funder norms slow full integration compared to fast-moving sectors like finance or tech.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI augmentation is already common and high-value: tools assist with literature synthesis, outlining, boilerplate generation, and editing, materially reducing drafting time while faculty retain strategic control. Faculty using AI writing assistants report improved productivity and faster iteration without displacing human judgment on proposal merit.
Augmentation potential	claude-sonnet-5	5/5	AI substantially speeds up drafting, editing, formatting, and literature summarization for grant proposals, letting faculty focus on research design and strategic framing while staying fully in control of final content.
Task automatability	claude-haiku-4-5-20251001	2/5	Writing grant proposals requires domain expertise, originality, and strategic alignment with specific funding priorities that vary by funder. While AI can draft sections (literature review, methods outline), the core intellectual work—formulating novel research questions, justifying significance, and positioning the work competitively—remains heavily dependent on human judgment. Automation would not meet the 50% time-savings threshold for high-quality, fundable proposals.
Task automatability	claude-sonnet-5	3/5	AI can draft substantial portions of a grant proposal (background, literature synthesis, boilerplate methodology sections) but requires deep domain expertise, novel research framing, and strategic alignment with funder priorities that still need heavy human input and revision.
Adoption barriers	claude-haiku-4-5-20251001	3/5	No licensing requirement exists for the writing task itself, but institutional grant policies, funder guidelines, and reputational risk around proposal quality create friction. Universities encourage human oversight and often require faculty sign-off, creating partial organizational barriers to full automation.
Adoption barriers	claude-sonnet-5	2/5	No licensing requirement bars AI use in drafting, though institutional research offices, PI accountability for accuracy, and funder scrutiny of originality/plagiarism impose some friction.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	A faculty member's loaded cost (salary + benefits) for grant-writing is substantial, while AI inference is cheap; however, the output from unsupervised AI requires significant expert review and revision, adding back human time. The all-in cost (inference + human oversight + refinement) does not undercut the human's effort materially, especially for high-stakes funding.
Cost vs. human wage	claude-sonnet-5	3/5	AI subscription costs are low, but the professor's time for review, fact-checking, and strategic tailoring remains significant, so overall cost savings are moderate rather than order-of-magnitude given the high stakes of funding outcomes.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Current AI tools (e.g., ChatGPT, writing assistants) can generate boilerplate text and structural templates, but no deployed product reliably writes fundable grant proposals end-to-end. Institutions use AI as a drafting aid, not a substitute, because proposal quality directly determines funding outcomes and rejection is costly. Products exist for outline generation but lack the domain integration and funder-specific adaptation needed for reliable deployment.
Technical feasibility today	claude-sonnet-5	3/5	Products like ChatGPT, Grantable, and other AI writing assistants are used by faculty to draft proposal sections today, but reliability on technical accuracy, originality, and funder-specific nuance remains inconsistent, requiring substantial expert editing.

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

CI 25–34 · exposure 25 · augmentation 75 · importance 4.0/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Academic institutions adopt AI tools incrementally for narrow functions (e.g., generating discussion prompts), but curriculum planning remains largely human-driven due to tradition, governance requirements, and risk aversion around educational quality and institutional autonomy.
Sector adoption velocity	claude-sonnet-5	2/5	Higher education adopts new tools slowly due to shared governance, tenure structures, and academic culture, with AI use in course design still largely at the experimental/pilot stage despite being an information-sector job.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI can meaningfully assist instructors by generating content drafts, suggesting alternative delivery methods, identifying gaps in existing materials, and facilitating rapid iteration on course design—substantially raising human productivity while the instructor retains critical pedagogical and institutional oversight.
Augmentation potential	claude-sonnet-5	4/5	AI is quite useful for brainstorming course content, generating assessment items, drafting learning objectives, and suggesting revisions, meaningfully speeding up an instructor's curriculum work while they retain final judgment.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can draft course outlines, generate sample materials, and suggest pedagogical approaches, curriculum planning requires deep subject expertise, institutional knowledge of student populations, accreditation standards, and iterative stakeholder feedback that AI cannot fully coordinate. End-to-end curriculum redesign with 50% time savings at equal quality is not achievable with current systems.
Task automatability	claude-sonnet-5	2/5	AI can draft syllabi, generate content ideas, and suggest revisions, but the overall planning and evaluation of curricula requires institutional judgment, pedagogical expertise, and accreditation alignment that current AI cannot fully replace end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Curriculum design is subject to institutional governance, accreditation bodies, departmental review processes, and faculty senate approval, all of which mandate human expert judgment and collective decision-making, creating significant legal and organizational barriers to full automation.
Adoption barriers	claude-sonnet-5	3/5	Curriculum decisions typically require faculty governance, accreditation compliance, and departmental review, creating institutional friction, though there's no strict licensing requirement barring AI-assisted drafting.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	While AI inference is cheap, the integration cost and substantial human oversight required to validate curricular coherence, pedagogical soundness, and institutional alignment means total cost per complete curriculum cycle remains comparable to or higher than human-only effort.
Cost vs. human wage	claude-sonnet-5	3/5	AI assistance is cheap per query, but the human oversight, review, and institutional approval processes needed still consume significant faculty time, keeping total cost roughly comparable to traditional methods for the full task.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Products exist (ChatGPT, Claude, specialized ed-tech tools) that can assist with material generation and outline suggestions, but no deployed system reliably handles the full curriculum design cycle, including evaluation against learning outcomes and institutional fit, without significant human revision and oversight.
Technical feasibility today	claude-sonnet-5	2/5	Products like ChatGPT and specialized ed-tech tools are used informally by instructors for drafting materials, but no deployed product reliably performs full curriculum planning and evaluation autonomously in production.

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

CI 25–34 · exposure 25 · augmentation 63 · importance 3.8/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Postsecondary institutions have been slow to adopt AI for core advising functions; most deployments remain experimental or advisory chatbots that supplement rather than replace human advisors, reflecting both cultural conservatism and legitimate concern over liability.
Sector adoption velocity	claude-sonnet-5	2/5	Higher education is a relatively slow-adopting sector for replacing personal advising roles, though some institutions pilot AI chat advisors for basic queries.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can assist advisors by retrieving program requirements, generating degree-path suggestions, and surfacing relevant career data, meaningfully reducing preparation time; however, the task's reliance on human judgment and relationship-building limits transformative productivity gains.
Augmentation potential	claude-sonnet-5	4/5	AI can significantly assist advisors by summarizing student records, suggesting curricula pathways, and providing career information, boosting efficiency while the human retains the relationship and judgment.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can provide generic information about curricula and career paths, advising requires understanding individual student backgrounds, goals, aptitudes, and constraints—nuanced judgment that current systems cannot reliably deliver end-to-end. Partial automation of information retrieval and formatting is possible, but the core advising function remains human-dependent.
Task automatability	claude-sonnet-5	2/5	Advising requires understanding individual student history, motivations, and institutional nuance that current AI cannot reliably synthesize into personalized, trusted guidance end-to-end.apiKey Some scheduling/curriculum-mapping subtasks can be automated but the core advisory judgment cannot.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Advising is a core duty of faculty and professional advisors with fiduciary responsibility to students; many institutions require human sign-off on academic planning, and there is strong organizational and regulatory expectation that qualified humans perform or directly oversee student guidance.
Adoption barriers	claude-sonnet-5	3/5	No formal licensure required for most academic advising, but institutional policy, liability concerns for career/curriculum decisions, and student preference for human mentorship create meaningful friction.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI systems for this task are cheap to run, but require significant human oversight and intervention to be trustworthy; integrating them into advising workflows adds coordination costs that approach or exceed the savings from automation.
Cost vs. human wage	claude-sonnet-5	3/5	AI tools can cheaply handle informational queries, but comprehensive advising still requires human faculty time, so overall cost savings are moderate rather than dramatic.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Chatbots can answer factual questions about programs and careers, but no deployed product reliably performs the full advising task at production scale in educational institutions; implementations remain narrow tools for information delivery rather than genuine advisory systems.
Technical feasibility today	claude-sonnet-5	2/5	AI chatbots and advising tools exist in some universities for FAQs and degree audits, but reliable, trusted personalized academic/career advising by AI is not yet deployed at scale.

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

CI 25–32 · exposure 25 · augmentation 75 · importance 4.3/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Academic institutions are adopting AI writing and analysis tools as assistants, but the core research and publication task remains human-driven with slow adoption of any automation that claims to reduce researcher involvement. Sector digitization is moderate and institutional resistance to algorithmic research is high.
Sector adoption velocity	claude-sonnet-5	3/5	Academia has seen notable uptake of AI for literature review, writing assistance, and data analysis, but full-scale institutional adoption for core research tasks remains uneven and cautious.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI significantly augments researcher productivity through literature search, data visualization, writing assistance, and statistical computation, allowing researchers to focus on conceptualization and interpretation. These tools are widely adopted in academic workflows to enhance human capability while the researcher remains the primary decision-maker.
Augmentation potential	claude-sonnet-5	4/5	AI substantially aids literature synthesis, drafting, editing, and data analysis, meaningfully boosting researcher productivity while the scholar retains intellectual control.
Task automatability	claude-haiku-4-5-20251001	2/5	AI can assist with literature review, data analysis, and manuscript drafting, but original research conception, novel experimental design, and the creative synthesis required for publishable findings remain dependent on human expertise and domain insight. Automating the full research lifecycle end-to-end with 50% time savings at equal quality is not feasible with current systems.
Task automatability	claude-sonnet-5	2/5	AI can assist with literature review, drafting, and data analysis but cannot independently design novel research programs, generate original insights, or conduct empirical/experimental work end-to-end at equal quality.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Academic publishing, institutional advancement, and professional credibility require human authorship and accountability; journals demand human intellectual contribution and responsibility for findings. Regulatory and professional norms strongly protect this task—falsifying or misrepresenting authorship carries legal and career consequences.
Adoption barriers	claude-sonnet-5	3/5	Academic norms, peer review, authorship ethics, and institutional requirements for original human-driven scholarship create moderate friction, though not a hard licensing barrier.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI tools (GPT subscriptions, research databases, writing assistants) cost substantially less than a researcher's salary, but they do not replace the core intellectual work; the human researcher remains the primary cost driver and must validate all outputs, making the all-in cost comparable to or exceeding manual research.
Cost vs. human wage	claude-sonnet-5	2/5	AI tools are cheap for subtasks like summarization, but the overall research process still requires substantial expert human time for design, experimentation, and validation, keeping cost savings modest.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	While AI tools exist for literature summarization and writing assistance, no deployed system can autonomously conduct novel research, design experiments, interpret results, or produce publication-ready work that meets peer-review standards. Products support narrow aspects (citation management, writing) but not the complete task reliably.
Technical feasibility today	claude-sonnet-5	2/5	Products like AI writing assistants and literature-search tools are used but no deployed system performs full research conception-to-publication reliably; hallucination and citation errors remain common issues.

Participate in student recruitment, registration, and placement activities.

CI 25–30 · exposure 25 · augmentation 50 · importance 3.8/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Higher education adoption of AI in recruitment and placement remains in pilot phase; most institutions rely on traditional staff-driven processes with incremental tool adoption (CRM, email automation), not deep AI agent deployment.
Sector adoption velocity	claude-sonnet-5	2/5	Higher education adopts AI unevenly; admissions offices use some automation but faculty-level recruitment/placement participation remains largely traditional and slow to change.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can assist with candidate screening, data organization, and outreach coordination, improving faculty efficiency in these administrative parts; however, the human educator must remain central to relationship-building and placement decisions, limiting transformative potential.
Augmentation potential	claude-sonnet-5	3/5	AI can help draft recruitment content, analyze applicant data, and suggest placement matches, giving useful support even though final decisions and interactions remain human-led.
Task automatability	claude-haiku-4-5-20251001	2/5	Only narrow parts of recruitment and placement can be automated today (email screening, scheduling logistics). The task fundamentally requires human relationship-building, contextual judgment about student fit, and persuasion—core elements that current AI systems cannot reliably handle end-to-end at the ≥50% time-saving threshold.
Task automatability	claude-sonnet-5	2/5	Parts like drafting recruitment materials or answering FAQs can be automated, but personal advising, interviews, and judgment-based placement decisions resist full automation.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Institutional policy, accreditation bodies, and student expectation for human guidance create strong friction; many institutions legally require faculty sign-off on placement and enrollment decisions. Liability and regulatory oversight of admissions further protect this task from full automation.
Adoption barriers	claude-sonnet-5	3/5	No licensing requirement for this specific task, but institutional policy, accreditation standards, and student trust favor human involvement in admissions and placement decisions.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI tools for recruitment and registration are modest in cost-savings relative to loaded faculty wages; integration with existing student systems, compliance checks, and the need for human oversight keep all-in costs from dropping significantly below equivalent human effort.
Cost vs. human wage	claude-sonnet-5	2/5	AI can cheaply handle mass communications, but the faculty-specific advising and placement judgment portions still require paid human time, keeping overall cost savings modest.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Some AI tools exist for student matching and CRM assistance in higher ed, but no production systems reliably automate the full recruitment-registration-placement pipeline. Deployed products handle fragments (scheduling, basic communication) but lack the judgment and rapport-building that define these activities.
Technical feasibility today	claude-sonnet-5	2/5	CRM tools and chatbots assist with outreach and initial screening in admissions offices, but faculty participation in recruitment/placement still relies heavily on human judgment and relationship-building not replicated by deployed products.

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

CI 16–38 · exposure 17 · augmentation 75 · importance 4.2/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Academia has been slow to adopt AI agents for core professional responsibilities; while some institutions pilot AI literature-monitoring tools, the field still emphasizes human agency in professional development as a non-delegable responsibility.
Sector adoption velocity	claude-sonnet-5	3/5	Higher education and academic professionals are moderate adopters of AI research tools (e.g., literature review assistants), though full professional development practices remain human-centered.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI can significantly assist professors by filtering and summarizing recent literature, flagging relevant conferences, and synthesizing key developments from thousands of sources—substantially raising the efficiency of staying current while the faculty member retains judgment and active professional engagement.
Augmentation potential	claude-sonnet-5	4/5	AI literature summarization, alerting services, and conference note synthesis tools meaningfully speed up how educators track developments in their field.
Task automatability	claude-haiku-4-5-20251001	1/5	This task fundamentally requires human judgment about which developments matter, professional network engagement, and critical assessment of field trends—activities that depend on human expertise and discretion that AI cannot replicate end-to-end.
Task automatability	claude-sonnet-5	2/5	AI can help summarize and surface literature but the task inherently requires human judgment, networking, and conference participation that cannot be fully offloaded to AI end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Professional development and scholarly engagement are intrinsic to academic roles and faculty responsibility; institutional expectations, tenure/promotion criteria, and the professional norm that faculty must demonstrate active field engagement create strong organizational and cultural barriers to full automation.
Adoption barriers	claude-sonnet-5	2/5	No licensing barrier, but professional norms, tenure/promotion expectations, and value of human networking create moderate friction against pure AI substitution.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI tools for literature summarization cost relatively little, but they do not replace the time investment required for meaningful professional engagement, and human oversight of AI summaries adds cost rather than reducing it.
Cost vs. human wage	claude-sonnet-5	2/5	AI summarization tools are cheap, but since the task cannot be substituted wholesale, the cost comparison to a human performing the full task (networking, conference attendance) is not favorable to AI replacement.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	While AI can summarize literature and highlight conference announcements, it cannot authentically participate in colleague conversations or selectively attend conferences based on relevance to one's specific research agenda—the core elements of the task.
Technical feasibility today	claude-sonnet-5	2/5	Tools like AI research summarizers and literature-alerting services exist but are narrow aids, not products that autonomously 'keep abreast' across a discipline for a professional.

Prepare and deliver lectures to undergraduate or graduate students on topics such as children's literature, learning and development, and reading instruction.

CI 25–25 · exposure 25 · augmentation 75 · importance 4.1/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Higher education has been slow to adopt AI for core instruction delivery; most adoption is limited to administrative tasks or optional supplementary tools. Faculty governance and credentialing gatekeeping slow displacement in this sector.
Sector adoption velocity	claude-sonnet-5	2/5	Higher education has been slow to adopt AI for core instructional delivery, though use of AI-assisted materials creation is growing among faculty.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI assistants can meaningfully augment lecture preparation (drafting outlines, generating examples, creating visuals) and enhance content delivery (generating real-time summaries, adaptive quizzes). These tools raise instructor productivity while human expertise and presence remain central.
Augmentation potential	claude-sonnet-5	4/5	AI is widely useful for lecture prep, generating slides, examples, quizzes, and summarizing research, significantly boosting instructor productivity while they retain the teaching role.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can generate lecture outlines and draft slides on educational topics, the interactive delivery, real-time responsiveness to student confusion, and adaptive teaching required in live lectures fundamentally depend on human presence and judgment. Current AI systems cannot replicate the full end-to-end task with 50% time savings at equal quality.
Task automatability	claude-sonnet-5	2/5	AI can help draft lecture content and slides, but live delivery, in-person interaction, Q&A, and adapting to student needs require human presence and judgment that current systems cannot fully replace end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Institutional accreditation standards, professional licensure requirements for degree-granting educators, and regulatory expectations that students receive instruction from qualified faculty create strong legal and organizational barriers to full automation.
Adoption barriers	claude-sonnet-5	4/5	Accreditation standards, degree-granting requirements, and expectations of qualified faculty create strong institutional and regulatory barriers to full automation of instruction.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI lecture-generation tools cost substantially less per unit than instructor salaries, but the task also requires live delivery, real-time student interaction, and credentialing—human instructors remain necessary. The cost advantage is eroded by retained human overhead.
Cost vs. human wage	claude-sonnet-5	2/5	While AI content generation is cheap, replicating credentialed, interactive lecture delivery at scale still requires human oversight, so overall cost savings are limited.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	AI can assist with content generation and note synthesis, but no deployed system reliably performs full lecture delivery end-to-end in production academic settings. Prototype systems exist for content drafting, but classroom delivery requires human facilitation and remains the standard.
Technical feasibility today	claude-sonnet-5	2/5	AI tools (e.g., video-generation, virtual tutors) exist for content delivery but are not widely deployed as substitutes for live postsecondary lecturing in accredited programs.

Advise and instruct teachers employed in school systems by providing activities, such as in-service seminars.

CI 16–30 · exposure 17 · augmentation 63 · importance 2.9/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Education sectors have been slow to adopt AI-driven professional development at scale; pilots exist but most in-service seminars remain human-led, and districts are cautious about substituting AI for trusted expert instruction.
Sector adoption velocity	claude-sonnet-5	2/5	Higher education and K-12 professional development are slow-adopting sectors for AI-driven instruction, with pilots more common than production deployment for teacher training.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can assist human instructors by drafting seminar agendas, generating example activities, and personalizing content recommendations for teachers, raising preparation efficiency; however, the human expert remains essential for delivery and adaptation.
Augmentation potential	claude-sonnet-5	4/5	AI can meaningfully help postsecondary educators draft seminar content, create instructional materials, and personalize training modules, improving efficiency while the human retains the advisory role.
Task automatability	claude-haiku-4-5-20251001	1/5	This task requires designing and delivering customized professional development, understanding teacher pedagogical needs, and facilitating interactive learning—activities that demand human judgment, contextual awareness, and real-time adaptation that current AI systems cannot reliably perform end-to-end.
Task automatability	claude-sonnet-5	2/5	Designing and delivering interactive in-service seminars requires relational credibility, adapting to a live room of teachers, and contextual judgment that current AI cannot replicate end-to-end, though content prep can be partly automated.
Adoption barriers	claude-haiku-4-5-20251001	4/5	School districts and educator organizations typically require that professional development be delivered by credentialed instructors with pedagogical expertise; legal and contractual obligations to employ qualified human instructors create substantial barriers to pure automation.
Adoption barriers	claude-sonnet-5	3/5	No formal licensing barrier prevents AI-assisted content, but institutional norms favor experienced human instructors for peer credibility and interactive mentorship in professional development.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI-generated content can reduce preparation costs, but delivering professional development seminars requires experienced instructional experts commanding significant loaded wages; AI does not yet offer an order-of-magnitude cost advantage given the need for human facilitation and oversight.
Cost vs. human wage	claude-sonnet-5	2/5	Human-led seminars carry significant preparation and delivery costs, but AI still requires substantial human oversight and facilitation, limiting cost savings for the full task.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	While AI can generate seminar outlines or draft training materials, no deployed product reliably performs the full task of advising and instructing teachers in live seminars with the credibility, responsiveness, and domain expertise required; implementations remain narrow and supplementary.
Technical feasibility today	claude-sonnet-5	2/5	AI tools exist to generate training materials or slides, but no deployed product independently advises or instructs practicing teachers in professional development settings at scale.

Initiate, facilitate, and moderate classroom discussions.

CI 13–30 · exposure 20 · augmentation 50 · importance 4.1/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	1/5	Postsecondary institutions remain heavily dependent on in-person or synchronous instruction with human instructors. There is minimal adoption of AI systems to replace live classroom facilitation; discussions are a core function tied to the instructor role and institutional identity.
Sector adoption velocity	claude-sonnet-5	2/5	Higher education is adopting AI for content creation and grading assistance, but live classroom facilitation tools remain niche and mostly limited to online/asynchronous formats.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can provide modest augmentation: generating discussion prompts, summarizing contributions in real time, or flagging student questions for follow-up. However, the assistance is limited to preparatory or post-hoc support; AI cannot meaningfully augment the live, in-the-moment facilitation and moderation that defines the task.
Augmentation potential	claude-sonnet-5	3/5	AI can help generate discussion questions, summarize contributions, or moderate online forums, giving meaningful but partial support to instructors running discussions.
Task automatability	claude-haiku-4-5-20251001	2/5	AI cannot reliably initiate, facilitate, or moderate live classroom discussions today. While AI can generate discussion prompts or summarize points, the task requires real-time responsiveness to student contributions, managing group dynamics, reading social cues, and making pedagogical adjustments—capabilities that current systems lack in live, unscripted settings. Modest parts (prompt generation) are automatable, but the core task remains fundamentally human.
Task automatability	claude-sonnet-5	2/5	Live, in-person facilitation requires reading a room, adapting to unpredictable student input, managing dynamics, and building rapport, which current AI cannot do end-to-end in a classroom setting.dependencies remain human-driven.
Adoption barriers	claude-haiku-4-5-20251001	5/5	Postsecondary teaching is heavily regulated; instructors are licensed professionals and institutions have accreditation requirements that mandate human instructors lead pedagogy and direct student engagement. Liability and duty-of-care obligations further require human presence and judgment in classroom discussion moderation.
Adoption barriers	claude-sonnet-5	3/5	No formal licensing barrier prevents AI-assisted discussion tools, but strong institutional norms, accreditation expectations, and student/parent preference for human instructors create moderate friction.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Even where AI could assist with partial tasks (prompt generation, summary), the marginal cost savings are small compared to the instructor's loaded wage, and human oversight is still required. The full task cannot be cost-displaced by AI infrastructure.
Cost vs. human wage	claude-sonnet-5	2/5	A human instructor's time is already bundled into teaching duties; deploying AI moderation tools for discussions adds cost without clearly reducing instructor time, especially for synchronous in-person settings.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed product reliably performs live classroom discussion facilitation and moderation. Chatbots can respond to individual queries, but orchestrating and moderating multi-participant classroom dialogue in real time—with contextual awareness and pedagogical judgment—remains research-stage only. Production systems do not exist for this task.
Technical feasibility today	claude-sonnet-5	2/5	AI chatbots can moderate asynchronous online discussion boards or generate discussion prompts, but no deployed product reliably facilitates live in-person classroom discussion at scale.

Provide professional consulting services to government or industry.

CI 11–28 · exposure 13 · augmentation 75 · importance 2.3/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Consulting firms are piloting AI-assisted research and drafting, but production-level consulting automation remains rare. Adoption is slow because consulting revenue depends on client perception of expertise and personal relationships, which AI cannot replicate. Displacement is minimal to date.
Sector adoption velocity	claude-sonnet-5	3/5	Academia and consulting-adjacent professional services show moderate AI tool adoption for research and writing support, but the core consulting engagement remains human-delivered with adoption mostly at the assistive stage.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI can meaningfully assist consultants by generating background research, summarizing industry reports, drafting analysis sections, and organizing data—tasks that occupy significant time. A human consultant using AI tools for preparation and synthesis can serve more clients at higher quality, making augmentation substantial while human judgment remains central.
Augmentation potential	claude-sonnet-5	4/5	AI substantially aids postsecondary educators in research synthesis, data analysis, and drafting reports or recommendations, meaningfully boosting productivity while the professional retains responsibility for advice.
Task automatability	claude-haiku-4-5-20251001	2/5	Professional consulting requires deep domain expertise, relationship-building, and nuanced judgment about organizational context. While AI can assist with research and report generation, the core consulting task—advising clients on strategy, implementation, and organizational fit—remains heavily dependent on human credibility, experience, and accountability.
Task automatability	claude-sonnet-5	1/5	This task involves original expert judgment, relationship management, and contextualized advice-giving that current AI cannot perform end-to-end at equal quality with major time savings.ateurs.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Regulatory and professional standards often require licensed or credentialed experts to provide certain consulting advice, particularly in regulated sectors (engineering, law, finance). Clients demand accountability and sign-off from named experts, creating legal and reputational barriers to full automation. Liability asymmetry is significant.
Adoption barriers	claude-sonnet-5	4/5	Professional consulting often involves credentialed expertise, contractual liability, confidentiality, and client expectations of accountable human advisors, creating strong barriers to full AI substitution.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	A postdoctoral consultant's billing rate is high ($150–300+/hour loaded), and AI inference cost is negligible. However, integration requires human oversight and liability management, and consulting output quality is judged by client outcomes rather than throughput. The human remains the cost driver for credible consulting.
Cost vs. human wage	claude-sonnet-5	2/5	AI can cheaply produce background research or drafts, but the actual consulting deliverable requires expert human time, oversight, and liability-bearing judgment, keeping overall costs comparable to human-led work.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	No mature AI product reliably performs professional consulting end-to-end. AI tools can draft reports or analyze data, but deployed systems cannot independently conduct client interviews, negotiate scope, or take responsibility for advice that affects organizational outcomes.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product independently performs professional consulting engagements for government or industry; AI tools at best support research and drafting within a human-led process.

Collaborate with colleagues to address teaching and research issues.

CI 7–30 · exposure 13 · augmentation 50 · importance 4.0/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Postsecondary education is a laggard sector in AI adoption, with strong institutional conservatism and human-centric norms. While some universities pilot AI writing assistance, little evidence exists of AI-driven adoption for collaborative problem-solving among faculty.
Sector adoption velocity	claude-sonnet-5	2/5	Higher education is a slower-adopting sector for AI in core professional/collegial activities, though administrative and research-support tools are being piloted.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can usefully assist by drafting meeting agendas, synthesizing prior discussions, or generating background summaries of research issues, helping faculty collaborate more efficiently. However, the core negotiation and consensus-building remains human-led, limiting augmentation to preparation and documentation support.
Augmentation potential	claude-sonnet-5	3/5	AI can assist by summarizing research, drafting meeting notes, or suggesting talking points, but the core collaborative interaction remains human-driven.
Task automatability	claude-haiku-4-5-20251001	2/5	AI can assist with drafting collaborative documents and summarizing research, but genuine professional collaboration requires human judgment, negotiation, and interpersonal nuance that AI cannot autonomously execute. The task intrinsically involves resolving disagreements and building consensus, which remain fundamentally human activities.
Task automatability	claude-sonnet-5	1/5	This task is inherently interpersonal collaboration among colleagues requiring shared judgment, negotiation, and relationship-building that AI cannot substitute for end-to-end.'
Adoption barriers	claude-haiku-4-5-20251001	4/5	Collaboration on pedagogical and research matters requires professional trust, accountability, and direct human communication. Institutional culture and the need for genuine peer engagement create strong barriers to AI substitution; colleagues expect human-to-human discussion on substantive academic issues.
Adoption barriers	claude-sonnet-5	3/5	No formal licensing requirement blocks AI involvement, but organizational and academic culture strongly favor human collegial deliberation, creating moderate friction.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Current AI assistance (email drafting, summarization) costs are modest but pale against the loaded cost of a postsecondary educator's time. Full autonomy is not feasible, so the cost advantage of partial assistance is minimal relative to faculty hourly rates.
Cost vs. human wage	claude-sonnet-5	1/5	There is no AI product that replaces this task, so cost comparison is moot; humans remain the only means of performing it.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	While AI tools exist for document collaboration and communication support, no deployed product autonomously conducts professional collaboration on teaching and research issues. Existing systems can draft emails or summarize positions, but cannot replace the iterative dialogue and relationship-building required.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product performs collegial collaboration on teaching/research issues; AI tools at best support communication logistics, not the collaborative act itself.

Supervise students' fieldwork, internship, and research work.

CI 0–16 · exposure 8 · augmentation 50 · importance 4.2/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	1/5	Adoption of AI for fieldwork supervision is negligible; educational institutions are slow-moving on labor substitution and face strong cultural and legal pressures to maintain human faculty oversight of student learning and safety.
Sector adoption velocity	claude-sonnet-5	2/5	Higher education adopts AI tools for administrative and content tasks but supervisory/mentorship functions in fieldwork and research show minimal displacement.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can assist supervisors by automating progress tracking, generating reports, flagging scheduling conflicts, or analyzing submitted work, which raises faculty productivity in documentation and monitoring tasks while humans retain decision-making authority.
Augmentation potential	claude-sonnet-5	3/5	AI can help with scheduling, tracking progress, summarizing reports, or providing feedback drafts, but the core supervisory relationship remains human-led.
Task automatability	claude-haiku-4-5-20251001	2/5	AI cannot meaningfully supervise fieldwork or internships end-to-end, as these require real-time observation, judgment about student safety and learning outcomes, and adaptive feedback based on in-situ performance. While AI could assist with scheduling, documentation, or automated progress reports, human supervision remains essential for the core responsibilities.
Task automatability	claude-sonnet-5	1/5	Supervising fieldwork, internships, and research requires in-person or relational oversight, mentorship, and real-time judgment calls that current AI cannot perform end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	5/5	Educational institutions face hard legal and regulatory barriers: faculty members bear duty-of-care obligations, institutional liability for student safety, and accreditation standards requiring qualified human supervision of fieldwork and internships. Licensing and professional responsibility requirements make automated substitution infeasible.
Adoption barriers	claude-sonnet-5	4/5	Accreditation, institutional policy, and liability considerations typically require a qualified faculty member to supervise and sign off on student fieldwork and research.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	The cost of AI systems with sufficient real-time monitoring, safety assurance, and liability coverage would exceed the cost of human faculty supervision, particularly given legal and institutional accountability requirements.
Cost vs. human wage	claude-sonnet-5	1/5	There is no viable AI substitute performing this supervisory role, so cost comparison favors the human by default.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed products reliably perform real-time supervision of students in fieldwork or internship settings. AI systems lack the situational awareness, accountability, and duty-of-care capabilities needed for this high-stakes educational and safety function.
Technical feasibility today	claude-sonnet-5	1/5	No deployed products act as substitutes for a faculty supervisor overseeing student fieldwork or research placements; this remains a human relational role.

Maintain regularly scheduled office hours to advise and assist students.

CI 4–11 · exposure 0 · augmentation 38 · importance 3.9/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	1/5	Higher-education institutions have shown minimal adoption of AI to replace office hours. The sector remains conservative on human-contact substitution, especially for advisory roles central to the educational mission.
Sector adoption velocity	claude-sonnet-5	2/5	Higher education adopts AI advising tools cautiously and unevenly; office hours remain a traditional, slow-changing institutional practice.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI could marginally assist a professor by preparing summaries of prior advising notes or suggesting common resources, but the core task—listening, personalizing advice, and building trust—depends on the human faculty member.
Augmentation potential	claude-sonnet-5	3/5	AI can help faculty prepare materials, answer routine student questions beforehand, or provide scheduling/FAQ support, moderately easing the advising workload without replacing the core interaction.
Task automatability	claude-haiku-4-5-20251001	1/5	This task requires real-time, synchronous human interaction and personalized mentorship. Students seek advisory relationships with faculty that depend on human judgment, empathy, and accountability—elements current AI systems cannot replicate in a legally or pedagogically acceptable way.
Task automatability	claude-sonnet-5	1/5	This task fundamentally requires a designated human presence and relationship-based mentoring; AI cannot substitute for the institutional role of faculty office hours even if chatbots can answer some factual questions.
Adoption barriers	claude-haiku-4-5-20251001	5/5	Strong regulatory and institutional barriers protect this task. Faculty advising is often a contractual and accreditation requirement; institutions and accreditors expect qualified humans to sign off on student guidance and developmental decisions.
Adoption barriers	claude-sonnet-5	4/5	Office hours are typically a contractual/institutional obligation tied to faculty role and accreditation expectations, and students often expect direct human mentorship, creating strong organizational and normative barriers.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Even if feasibility were higher, the cost would be unfavorable: deploying AI to replace office hours would still require human oversight, moderation, and liability coverage, while the human professor's salary would remain an organizational fixed cost.
Cost vs. human wage	claude-sonnet-5	2/5	While AI chat costs are low, the task is defined by institutional requirement for human presence, so cost comparison is largely moot; any AI substitute would still require human backstop for institutional compliance.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed product can substitute for faculty office hours today. While chatbots can answer routine questions, they cannot conduct genuine advising, build relationships, or take responsibility for student guidance as institutions require.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product replaces the faculty office-hours function itself; AI advising tools exist as supplements but not as substitutes for the scheduled human availability requirement.

Perform administrative duties, such as serving as department head.

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

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	1/5	Educational institutions have been slow to adopt AI for leadership roles; administrative automation remains limited to scheduling and document handling, with human judgment and accountability remaining central to departmental governance.
Sector adoption velocity	claude-sonnet-5	2/5	Higher education administration adopts AI slowly for governance and leadership functions, with most AI use limited to scheduling or communication support rather than headship duties.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI can assist with routine administrative tasks such as scheduling, data aggregation, and correspondence drafting, but offers limited augmentation for the core leadership judgment, stakeholder negotiation, and strategic planning that define the role.
Augmentation potential	claude-sonnet-5	3/5	AI can assist department heads with drafting reports, summarizing data, scheduling, and administrative correspondence, meaningfully aiding but not replacing the leadership function.
Task automatability	claude-haiku-4-5-20251001	1/5	Serving as department head requires strategic decision-making, personnel management, budget oversight, and stakeholder engagement that depend heavily on contextual judgment, organizational politics, and human relationships. Current AI cannot perform these complex leadership functions end-to-end.
Task automatability	claude-sonnet-5	1/5	Serving as department head involves leadership, personnel decisions, strategic planning, and interpersonal negotiation that current AI cannot execute end-to-end.dimension of judgment and authority makes this infeasible to automate.
Adoption barriers	claude-haiku-4-5-20251001	5/5	Institutional governance, fiduciary duty, personnel management, and budget authority create hard legal and organizational barriers. A human must legally hold the department head role and sign off on hiring, budgets, and policy decisions.
Adoption barriers	claude-sonnet-5	5/5	Department head roles require institutional authority, accountability, tenure/rank status, and formal appointment by the institution, which cannot legally or organizationally be delegated to AI.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	Department head duties are deeply tied to human accountability and institutional governance. The cost of deploying AI oversight infrastructure plus human verification would exceed the salary burden, and AI cannot legally assume fiduciary responsibility.
Cost vs. human wage	claude-sonnet-5	1/5	There is no AI substitute performing this role, so cost comparison favors the human entirely; AI cannot deliver the output at all.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	While AI can assist with scheduling, document management, and routine reporting, no deployed product reliably handles the full scope of departmental leadership tasks. Products exist for isolated components but cannot substitute for the human department head's judgment and accountability.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product performs departmental leadership or administrative headship roles; this remains firmly outside current AI product capability.

Serve as a liaison between the university and other governmental and educational agencies.

CI 0–7 · exposure 0 · augmentation 38 · importance 3.3/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	1/5	Universities are not automating liaison roles; these positions remain tied to formal governance structures and require human accountability, showing zero displacement trend.
Sector adoption velocity	claude-sonnet-5	2/5	Higher education administration adopts AI tools slowly for relationship-based and governance functions, with most current use limited to scheduling or communication support rather than liaison roles themselves.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI could modestly assist with scheduling, background research on partner agencies, or drafting correspondence, but the core work of relationship maintenance and negotiation requires human presence and judgment.
Augmentation potential	claude-sonnet-5	3/5	AI can help draft communications, summarize meetings, track correspondence, and manage scheduling with partner agencies, meaningfully supporting but not replacing the liaison function.
Task automatability	claude-haiku-4-5-20251001	1/5	This task requires sustained relationship-building, negotiation, and judgment about institutional interests that depend on human authority and trust. Current AI cannot autonomously represent a university's interests in formal liaison roles or make binding commitments.
Task automatability	claude-sonnet-5	1/5	This task requires ongoing relationship management, negotiation, and representation of institutional interests across organizations, which is fundamentally interpersonal and political in nature and cannot be executed end-to-end by AI.
Adoption barriers	claude-haiku-4-5-20251001	5/5	This role is legally and institutionally restricted to humans; agencies require human counterparts with official authority, accountability, and legal standing to negotiate agreements and represent institutional interests.
Adoption barriers	claude-sonnet-5	4/5	Institutional representation typically requires designated authority, trust, and accountability vested in a specific person, creating strong organizational and reputational barriers to automation.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	The task fundamentally requires a salaried human professional; AI could assist with document preparation or scheduling, but cannot replace the liaison function itself, making cost comparison unfavorable for AI substitution.
Cost vs. human wage	claude-sonnet-5	1/5	There is no viable AI substitute for the human relationship-building and representation involved, so cost comparison favors the human by default.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed product can serve as a university's official liaison to government or educational agencies; such roles require human decision-making authority and legal standing that AI systems do not possess.
Technical feasibility today	claude-sonnet-5	1/5	No deployed AI product performs the role of institutional liaison; this remains squarely a human relationship and trust-based function.

Act as advisers to student organizations.

CI 0–7 · exposure 0 · augmentation 38 · importance 2.5/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	1/5	Adoption of AI for student advising is minimal and unlikely to accelerate significantly. Higher education institutions are conservative around student welfare and duty of care; there is no sector-wide movement to automate this advising role.
Sector adoption velocity	claude-sonnet-5	2/5	Higher education is generally slow to adopt AI for interpersonal mentorship and student affairs roles, with pilots mostly focused on administrative tasks rather than advisory relationships.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI could marginally assist an advisor by drafting communications or organizing information about student group policies, but the core advising relationship cannot be substantially augmented by current AI without the human remaining the primary decision-maker and relationship holder.
Augmentation potential	claude-sonnet-5	3/5	AI can help advisors with scheduling, drafting communications, budget tracking, or event planning support, but the core advisory and mentorship relationship remains human-driven.
Task automatability	claude-haiku-4-5-20251001	1/5	This task requires relationship-building, mentorship, and nuanced judgment about student welfare and organizational dynamics—core human competencies. Current AI cannot replicate the trust, contextual awareness, and emotional intelligence needed to advise student groups on complex interpersonal and institutional matters.
Task automatability	claude-sonnet-5	1/5	Advising student organizations requires ongoing relationship-building, mentorship, event supervision, and situational judgment that AI cannot perform end-to-end.9 No off-the-shelf system can substitute for the interpersonal and institutional role of a faculty advisor.
Adoption barriers	claude-haiku-4-5-20251001	5/5	Strong legal and institutional barriers exist: faculty advisers typically hold formal roles with fiduciary duties, institutions require human accountability for student organizations, and liability for failures (policy violations, student harm) rests on designated humans. Regulations around student life and institutional governance require authorized human oversight.
Adoption barriers	claude-sonnet-5	4/5	Institutional policy typically requires a human faculty/staff advisor to be officially designated, often for liability, safety, and accreditation reasons, creating a strong organizational and quasi-regulatory barrier.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	The cost of an AI system capable of maintaining ongoing relationships with student organizations, combined with necessary human oversight and liability coverage, would exceed the cost of a faculty advisor providing this service as part of their normal duties.
Cost vs. human wage	claude-sonnet-5	1/5	Since AI cannot perform this role at all, there is no viable cost comparison—human faculty involvement is required and AI offers no substitute cost savings.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed product reliably performs student advising with the judgment and accountability required. While chatbots can provide generic information, they cannot substitute for a faculty advisor who knows students individually and bears institutional responsibility for their welfare and compliance.
Technical feasibility today	claude-sonnet-5	1/5	There are no deployed AI products that act as advisors to student clubs or organizations; this remains outside current product capability.

Participate in campus and community events.

CI 0–5 · exposure 0 · augmentation 25 · importance 2.9/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	1/5	This task cannot be automated and has seen no AI adoption because it is inherently non-automatable; sectors show zero displacement signals.
Sector adoption velocity	claude-sonnet-5	1/5	Higher education is generally slow to adopt AI for interpersonal, community-facing roles, and this task is not a target of current automation trends.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI might marginally help with event logistics (scheduling, reminder systems, or communication drafting), but it does not meaningfully augment the core task of participating and engaging in events.
Augmentation potential	claude-sonnet-5	2/5	AI could help with scheduling, event promotion materials, or follow-up communications, but offers little assistance to the actual act of attending and engaging at events.
Task automatability	claude-haiku-4-5-20251001	1/5	Participating in campus and community events requires physical presence, social interaction, and contextual responsiveness that current AI systems cannot replicate. This is inherently a human relational task where an AI cannot substitute.
Task automatability	claude-sonnet-5	1/5	Physical presence, social engagement, and relationship-building at events cannot be performed by AI systems; this is an inherently human, in-person activity.'
Adoption barriers	claude-haiku-4-5-20251001	5/5	The task fundamentally requires a human presence and social participation; events require actual human attendance and interpersonal engagement, creating an absolute barrier to automation.
Adoption barriers	claude-sonnet-5	4/5	Institutional expectations, faculty visibility, networking, and community relations require a human representative; no regulatory license but strong organizational/social norms mandate physical human presence.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	There is no AI system to deploy for this task, so cost comparison is not applicable; automation would require a physical presence that AI cannot provide.
Cost vs. human wage	claude-sonnet-5	1/5	There is no AI equivalent that produces this output, so cost comparison favors the human by default.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed AI product can autonomously participate in or attend events with meaningful social engagement. This falls entirely outside the scope of current AI capabilities.
Technical feasibility today	claude-sonnet-5	1/5	No deployed AI product substitutes for a person's participation in live campus or community events.

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

CI 0–0 · exposure 0 · augmentation 38 · importance 3.4/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	1/5	Educational institutions have not adopted AI for committee service; governance structures legally and culturally mandate human faculty participation. Adoption is near-zero and unlikely to change substantively.
Sector adoption velocity	claude-sonnet-5	1/5	Higher education governance and committee work is a low-digitization, relationship-driven process with essentially no AI displacement occurring in this specific function.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI could assist with drafting agendas, summarizing policy documents, or preparing background materials for committee members, but the deliberation and voting are inherently human activities. Assistance is marginal compared to the core value of informed human discussion.
Augmentation potential	claude-sonnet-5	3/5	AI can help summarize meeting materials, draft policy language, or synthesize prior committee documents, providing moderate assistance while the human still attends and deliberates.
Task automatability	claude-haiku-4-5-20251001	1/5	Committee service requires human judgment on institutional policies, nuanced deliberation among stakeholders, and accountability for decisions affecting academic governance. Current AI cannot meaningfully participate in or replace this interactive, consensus-building process.
Task automatability	claude-sonnet-5	1/5	Committee service requires deliberation, negotiation, institutional judgment, and representing stakeholder interests in real-time meetings, which AI cannot perform end-to-end today.dependencies on relationships and authority.
Adoption barriers	claude-haiku-4-5-20251001	5/5	Academic governance requires human faculty members with institutional affiliation, voting rights, and legal standing. Regulations and institutional bylaws explicitly require human participation and decision-making authority in committee roles.
Adoption barriers	claude-sonnet-5	5/5	Committee membership typically requires formal institutional appointment, faculty governance status, and accountability that only a human employee with standing can hold.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	Committee service is intrinsic to faculty duties and governance; it cannot be cost-replaced by AI systems, which cannot assume institutional responsibility or voting authority.
Cost vs. human wage	claude-sonnet-5	1/5	Since AI cannot perform this role, there is no viable AI cost comparison; the human is required and AI offers no substitution to compare cost against.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed product can autonomously serve on institutional committees or make binding decisions on departmental matters. This requires legal standing, fiduciary responsibility, and human authority that AI systems do not possess.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product substitutes for a human sitting on and voting/deliberating within academic committees; this remains entirely research-stage or non-existent.

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.