Economics Teachers, Postsecondary

25-1063.00

Median wage $123,920/yr11,560 employed (US)Rank #251 of 923 scored · top 27% by substitution

Teach courses in economics. 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

Substitution36

Exposure33

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

22 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

14%

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%35

panel mean rating 2.4/5 → substitution pressure 35/100

Technical feasibility todayw 20%32

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

Cost vs. human wagew 15%40

panel mean rating 2.6/5 → substitution pressure 40/100

Adoption barriersw 20%inverted — strong barriers lower the score42

panel mean rating 3.3/5 (barrier strength) → substitution pressure 42/100

Sector adoption velocityw 10%31

panel mean rating 2.3/5 → substitution pressure 31/100

Task breakdown (22 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 92–95 · exposure 100 · augmentation 63 · 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	4/5	Postsecondary institutions have nearly universal adoption of automated student information systems; virtually all accredited colleges and universities delegate this task to software rather than manual entry.
Sector adoption velocity	claude-sonnet-5	5/5	Higher education has near-universal adoption of digital gradebooks and LMS-based attendance/grade tracking already embedded in standard workflows.
Augmentation potential	claude-haiku-4-5-20251001	3/5	Automated systems assist faculty by providing dashboards, exception alerts, and grade export tools that enhance visibility and reduce administrative overhead, though the core task is replacement rather than human-AI teaming.
Augmentation potential	claude-sonnet-5	4/5	Automated systems significantly reduce faculty administrative burden for recordkeeping while instructors retain oversight over grade accuracy and final submission.
Task automatability	claude-haiku-4-5-20251001	5/5	Maintaining attendance records, grades, and student records is almost entirely routine data entry and management tasks that modern educational information systems (learning management systems, student information systems, and administrative software) automate end-to-end, often with ≥90% time savings.
Task automatability	claude-sonnet-5	5/5	Recording and tallying attendance and grades is a structured data-entry and computation task fully handled by existing LMS software and gradebook automation, meeting the 50% time-saving bar easily.
Adoption barriers	claude-haiku-4-5-20251001	2/5	While some institutions require human sign-off on final grade submission (a lightweight governance requirement), the underlying record-keeping itself faces minimal legal or regulatory barriers; most adoption friction is organizational inertia rather than hard compliance mandates.
Adoption barriers	claude-sonnet-5	2/5	Some institutional policy requires instructor sign-off on final grades, but the recordkeeping mechanics themselves face little regulatory or licensing barrier.
Cost vs. human wage	claude-haiku-4-5-20251001	5/5	Cloud-based SIS and LMS systems cost cents per student per year for this function, and AI-assisted data entry (OCR, form processing) further reduces marginal cost to a fraction of an hour of faculty time.
Cost vs. human wage	claude-sonnet-5	5/5	Digital gradebooks and attendance systems cost a fraction of a cent per student-record compared to manual clerical time, an order-of-magnitude or greater savings.
Technical feasibility today	claude-haiku-4-5-20251001	5/5	Mature, deployed products (Canvas, Blackboard, Banner, Workday for higher education) reliably perform this task at scale across thousands of institutions worldwide, with robust audit trails and compliance-ready outputs.
Technical feasibility today	claude-sonnet-5	5/5	Learning management systems (Canvas, Blackboard, Moodle) already perform automated attendance tracking, grade calculation, and recordkeeping reliably in production at most universities.

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

CI 76–92 · exposure 87 · augmentation 100 · importance 4.5/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	4/5	Adoption is rapid and visible in higher education, a digitized, information-sector domain. Surveys and public reporting show widespread faculty use of AI for syllabus and assignment drafting, with integration into learning management systems accelerating. Not yet universal, but well past pilot phase in many institutions.
Sector adoption velocity	claude-sonnet-5	3/5	Higher education is adopting AI tools for course prep at a moderate pace, with growing informal use by faculty but no widespread institutional mandate or standardized deployment yet.
Augmentation potential	claude-haiku-4-5-20251001	5/5	AI dramatically augments faculty productivity: it drafts templates, generates assignment variations, adapts materials for different student levels, and helps iterate on clarity—all while the instructor retains judgment on learning objectives and course design. Faculty stay in the loop and make final decisions, while AI transforms the efficiency of material creation.
Augmentation potential	claude-sonnet-5	5/5	AI significantly speeds up drafting of syllabi, assignments, and handouts, letting instructors focus on customizing content, aligning with learning objectives, and reviewing quality.
Task automatability	claude-haiku-4-5-20251001	5/5	Current AI systems (GPT-4, Claude) can generate complete syllabi, homework assignments, and handouts from course descriptions in minutes, meeting the 50% time-saving threshold. A faculty member can prompt an AI to create materials, review/edit output, and deploy—a workflow that saves substantially more than 50% of the time traditionally spent drafting these materials from scratch.
Task automatability	claude-sonnet-5	4/5	LLMs can draft syllabi, homework problem sets, and handouts from a course outline or textbook with substantial time savings, needing mainly instructor review and customization to specific course policies and student needs.
Adoption barriers	claude-haiku-4-5-20251001	2/5	Few hard regulatory or legal barriers prevent faculty from using AI to draft course materials. However, institutional policies on academic integrity, intellectual property, and instructor attribution; faculty cultural resistance; and perceived quality concerns create modest friction. No licensed requirement or human sign-off mandate exists.
Adoption barriers	claude-sonnet-5	2/5	No licensing requirement governs drafting course materials, though institutional policies, accreditation standards, and academic integrity norms create some review friction before materials are finalized.
Cost vs. human wage	claude-haiku-4-5-20251001	5/5	LLM inference and integration costs are negligible (dollars per term), while faculty time at loaded wages (salary + benefits) is hundreds of dollars per hour. The cost ratio heavily favors AI—at least an order of magnitude cheaper than human labor for equivalent output.
Cost vs. human wage	claude-sonnet-5	5/5	Generating draft syllabi and assignments via an LLM costs a few cents to dollars in compute versus hours of a postsecondary instructor's paid time, an order-of-magnitude cost advantage.
Technical feasibility today	claude-haiku-4-5-20251001	5/5	Large language models deployed in production (ChatGPT, Claude, institutional LLM integrations) reliably generate course materials that are educationally sound and ready to use with minimal revision. Many universities and faculty already use these tools operationally to draft syllabi and assignments.
Technical feasibility today	claude-sonnet-5	4/5	Widely used tools like ChatGPT, Copilot, and course-design assistants are already routinely used by instructors to generate drafts of these materials in production settings, though final editing by the instructor is standard practice.

Compile bibliographies of specialized materials for outside reading assignments.

CI 72–81 · exposure 75 · augmentation 100 · 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	3/5	Higher education institutions have begun piloting AI research and writing assistants, but adoption remains patchy; many faculty retain control over bibliography curation as pedagogically meaningful. Adoption is faster in information-rich sectors but slower in academia's traditionally conservative teaching practices.
Sector adoption velocity	claude-sonnet-5	3/5	Higher education is adopting AI tools for research assistance at a moderate pace, with pilots and individual faculty use common but institutional-scale deployment still emerging.
Augmentation potential	claude-haiku-4-5-20251001	5/5	AI dramatically accelerates the research and collation phases while faculty review, validate, and customize selections—preserving human judgment on pedagogical relevance. This is a canonical augmentation use case: the professor's productivity on assignment design and curation is substantially raised while they remain fully in control.
Augmentation potential	claude-sonnet-5	5/5	AI substantially speeds up literature discovery and organization while the instructor retains judgment over final reading selections, making it a strong augmentation case.
Task automatability	claude-haiku-4-5-20251001	4/5	AI can compile comprehensive bibliographies from academic databases, extract citations, and format them reliably with minimal human intervention. Current systems (ChatGPT, Claude, specialized tools) can achieve substantial time savings (>50%) on this largely mechanical task of research, collation, and formatting, though human judgment on selection and scope may still be needed.
Task automatability	claude-sonnet-5	4/5	AI language models can search literature, identify relevant readings, and compile formatted bibliographies quickly, meeting the time-saving bar though a professor still needs to verify relevance and quality.
Adoption barriers	claude-haiku-4-5-20251001	2/5	Few regulatory or legal barriers exist; this is not a licensed profession and no human authorization is legally mandated. However, academic norms and instructor preference for curated (human-selected) readings create moderate organizational friction around full automation.
Adoption barriers	claude-sonnet-5	1/5	No licensing, legal, or safety barriers restrict who compiles a reading list; it's a routine administrative/academic task with no required credential.
Cost vs. human wage	claude-haiku-4-5-20251001	4/5	AI-powered bibliography compilation costs pennies to dollars per assignment versus hours of faculty labor at $50–100+/hour loaded cost. The cost differential is substantial, though institutional library resources and faculty time investment in validation add some friction to the pure inference cost.
Cost vs. human wage	claude-sonnet-5	5/5	Generating a bibliography via AI tools costs a few cents to a few dollars in compute versus substantial faculty time at academic hourly rates, an order-of-magnitude cost advantage.
Technical feasibility today	claude-haiku-4-5-20251001	4/5	Multiple deployed products (academic citation managers with AI integration, LLM-based research assistants, library database tools) perform bibliography compilation reliably today. Production use exists in academic settings, though some manual oversight for accuracy and relevance filtering remains common practice.
Technical feasibility today	claude-sonnet-5	4/5	Products like citation managers integrated with AI search (e.g., Semantic Scholar, Elicit, ChatGPT with browsing) reliably generate topic bibliographies today, though occasional citation errors or hallucinated sources require checking.

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

CI 62–76 · exposure 70 · augmentation 88 · 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	3/5	Postsecondary institutions have widely adopted learning management systems with auto-grading; however, many faculty still manually grade essays and exams. Adoption of AI-driven exam generation and subjective-answer grading is still in pilot/early mainstream phase rather than normalized practice.
Sector adoption velocity	claude-sonnet-5	3/5	Higher education is adopting AI grading and exam tools unevenly; pilots and partial deployments are common but full-scale reliance is still limited by policy and trust concerns.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI substantially assists faculty by auto-generating quiz and exam banks tailored to learning objectives, flagging potential ambiguities, and providing preliminary grade distributions and item analysis. These capabilities meaningfully reduce preparation and grading workload while keeping the instructor in control of assessment design and standards.
Augmentation potential	claude-sonnet-5	5/5	AI substantially helps instructors draft questions, create answer keys, and pre-grade or flag responses, meaningfully speeding up the entire exam workflow while the instructor retains final judgment.
Task automatability	claude-haiku-4-5-20251001	4/5	Current AI systems can automatically generate exam questions, administer online tests, and grade most objective answers (multiple choice, short calculation problems) with high accuracy. Grading essays and open-ended responses remains harder but semi-automatable with rubric-based scoring and human review. End-to-end automation with ~50% time savings is achievable for most exam workflows with existing LLM and test-platform tools.
Task automatability	claude-sonnet-5	4/5	AI can draft exam questions, generate rubrics, and grade many question types (especially short-answer, multiple-choice, and even essay-style economics questions) with substantial time savings, though final review still typically occurs.'
Adoption barriers	claude-haiku-4-5-20251001	3/5	Institutions have institutional policies and accreditation expectations around academic integrity, grade fairness, and human oversight of summative assessment. While AI is permitted for exam drafting and grading support, faculty autonomy and perceived need for human judgment in grades create moderate friction to full replacement.
Adoption barriers	claude-sonnet-5	2/5	No licensing requirement blocks AI grading, but academic integrity policies, institutional norms, and instructor accountability for grades create moderate friction.
Cost vs. human wage	claude-haiku-4-5-20251001	5/5	AI-based exam administration and objective grading costs pennies per student per exam via API or SaaS platforms, compared to faculty labor at $40–80/hour per grading batch. The cost advantage is at least 10–100× for large classes once systems are set up.
Cost vs. human wage	claude-sonnet-5	4/5	AI-based exam compilation and grading is far cheaper per assessment than faculty or TA time once workflows are set up, though initial integration and oversight add some cost.
Technical feasibility today	claude-haiku-4-5-20251001	4/5	Mature products (Canvas, Blackboard, Gradescope, ChatGPT/Claude via integrations) reliably handle exam creation, delivery, and grading of objective items at scale in educational institutions. Subjective grading assistance is deployed but typically requires human final review, limiting fully unattended feasibility to structured assessments.
Technical feasibility today	claude-sonnet-5	3/5	Products like AI-assisted grading tools (Gradescope AI features, LLM-based graders) exist and are used in some universities, but reliability on nuanced economics reasoning/essays still requires instructor spot-checking.

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

CI 32–84 · exposure 58 · augmentation 88 · importance 4.4/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	4/5	Academic and research institutions are rapidly adopting AI for literature review, data synthesis, and manuscript drafting; adoption is visibly accelerating in information-intensive sectors (academia, finance, pharma) with LLM-based research tools already in common use by 2024.
Sector adoption velocity	claude-sonnet-5	3/5	Academia has moderate AI adoption for literature review and writing assistance, but norms around research integrity and publication slow deeper adoption compared to fast-moving tech/finance sectors.
Augmentation potential	claude-haiku-4-5-20251001	5/5	AI substantially augments researcher productivity by automating literature synthesis, co-drafting, statistical analysis suggestions, and formatting, allowing researchers to focus on novel theory, experiment design, and critical judgment while AI handles routine cognitive work.
Augmentation potential	claude-sonnet-5	4/5	AI significantly aids literature reviews, data analysis, coding, and manuscript drafting, meaningfully boosting researcher productivity while the scholar retains responsibility for direction and validity.
Task automatability	claude-haiku-4-5-20251001	5/5	AI can significantly automate research task components—literature review, data analysis, statistical modeling, manuscript drafting, and formatting—achieving well over 50% time savings. However, the requirement for novel theoretical contribution, human judgment on research direction, and peer validation means full end-to-end automation without human involvement remains limited; nonetheless, the automation threshold of ≥50% time saving at equal quality is clearly met.
Task automatability	claude-sonnet-5	2/5	AI can assist with literature review, data analysis, and drafting, but original economic research requiring novel hypotheses, rigorous methodology, and defensible conclusions still requires substantial human intellectual contribution and cannot be fully automated end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	2/5	Few hard regulatory or licensing barriers exist for AI-assisted or AI-automated research; authorship conventions and institutional reputation concerns create modest friction, but neither prevents adoption of AI research tools in practice.
Adoption barriers	claude-sonnet-5	3/5	No licensing requirement blocks AI use, but academic norms, authorship rules, journal policies on AI-generated content, and reputational/tenure incentives create meaningful friction against full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	4/5	AI inference and integration costs for literature review, data analysis, and drafting are orders of magnitude cheaper than the salary cost of a researcher's time on these tasks; a loaded academic salary easily exceeds the per-task AI service cost by 10–100×.
Cost vs. human wage	claude-sonnet-5	2/5	AI tools reduce time on literature search and drafting, but the overall cost of producing publishable original research still requires significant paid faculty time for design, analysis validation, and peer-review response, keeping cost savings modest.
Technical feasibility today	claude-haiku-4-5-20251001	4/5	Deployed AI systems (LLMs, research tools, statistical packages) demonstrably perform literature synthesis, data analysis, and manuscript drafting in production; however, the creative and novel aspects of original research contribution, peer review, and publication gatekeeping still require human oversight, limiting fully autonomous deployment to narrower scopes.
Technical feasibility today	claude-sonnet-5	2/5	Products like AI research assistants and writing tools exist and are used by academics, but no deployed system reliably conducts independent economics research and publishes peer-reviewed findings without heavy human direction.

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

CI 51–54 · exposure 50 · 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	3/5	Higher education is piloting AI grading tools, especially for large lecture courses, but adoption remains uneven and cautious. Some universities are integrating automated grading into learning platforms; however, widespread production deployment is still emerging rather than entrenched.
Sector adoption velocity	claude-sonnet-5	2/5	Higher education is a comparatively slow-adopting sector for grading automation due to institutional inertia, faculty autonomy, and academic integrity concerns, despite pilots in some departments.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI significantly assists instructors by auto-scoring objective items, flagging common errors, and generating preliminary feedback on essays, allowing instructors to focus review on borderline cases and nuanced feedback. This raises instructor productivity substantially while maintaining human oversight of final grades.
Augmentation potential	claude-sonnet-5	4/5	AI substantially speeds up grading of routine assignments, generates feedback drafts, and flags plagiarism, meaningfully boosting instructor productivity while they retain final judgment.
Task automatability	claude-haiku-4-5-20251001	3/5	AI can automatically grade objective or multiple-choice work and provide preliminary assessments of essays/papers using rubrics, but nuanced evaluation of complex economic arguments, original thinking, and contextual understanding typically requires human judgment for fair grading. Current systems handle roughly half the grading workflow reliably.
Task automatability	claude-sonnet-5	3/5	AI can draft grades and feedback for essays and problem sets with significant time savings, but nuanced economics reasoning, originality assessment, and fairness concerns limit full end-to-end automation without human review.
Adoption barriers	claude-haiku-4-5-20251001	3/5	Institutional policies on academic integrity, faculty governance, and faculty union agreements often restrict automated grading without human review. Student and faculty resistance to perceived unfairness, plus liability concerns around grade disputes, create significant friction even where technical capability exists.
Adoption barriers	claude-sonnet-5	3/5	No licensing requirement blocks AI grading, but academic integrity policies, institutional accreditation standards, and student appeals processes create moderate friction requiring instructor sign-off.
Cost vs. human wage	claude-haiku-4-5-20251001	4/5	AI inference and integration into LMS platforms costs far less than instructor time for full end-to-end grading, especially at scale. An instructor grading 100+ papers is expensive; automated pre-grading or objective-question grading is typically an order of magnitude cheaper per task unit.
Cost vs. human wage	claude-sonnet-5	4/5	Once integrated, AI can grade large volumes of assignments far cheaper than paying a professor or TA per assignment, though setup and oversight add some cost.
Technical feasibility today	claude-haiku-4-5-20251001	3/5	Deployed LLM-based grading tools and learning management system integrations exist and are used in some institutions, but they still produce material errors in essay evaluation and struggle with inconsistent rubric application across varied student work. Production use is growing but remains inconsistent in reliability.
Technical feasibility today	claude-sonnet-5	3/5	AI grading tools and LLM-based rubrics are used in some courses, but adoption for postsecondary economics grading is uneven and often supplementary rather than authoritative.

Select and obtain materials and supplies, such as textbooks.

CI 25–72 · exposure 45 · augmentation 75 · 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	2/5	Higher education procurement remains largely traditional with slow digitization of purchasing workflows. Adoption of AI-assisted material sourcing in this context is minimal, with most institutions still relying on manual processes.
Sector adoption velocity	claude-sonnet-5	3/5	Higher education has moderate AI adoption for administrative and research support tasks, with pilots for course planning tools but not yet widespread systematic use for material selection specifically.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can meaningfully assist by aggregating textbook options, comparing prices across vendors, and summarizing content reviews, helping faculty make better-informed selection decisions more quickly.
Augmentation potential	claude-sonnet-5	5/5	AI significantly assists instructors by quickly surfacing textbook options, summarizing content, comparing editions/costs, and drafting supply lists, while the instructor retains final decision authority.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can help search for and compare textbooks and materials, the actual procurement process requires human judgment about institutional needs, budget constraints, and approval workflows. AI cannot autonomously execute purchases or navigate institutional procurement systems end-to-end.
Task automatability	claude-sonnet-5	4/5	AI can readily search, compare, and recommend textbooks and course materials based on syllabus goals, and can even summarize reviews or generate reading lists, saving substantial time versus manual research.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Institutional purchasing requires authorization, approval workflows, and accountability that legally and administratively rest with human decision-makers. Budget authority and vendor relationships are typically locked to specific personnel.
Adoption barriers	claude-sonnet-5	1/5	No licensing or regulatory requirement mandates a human perform this selection process; it's largely an administrative/academic judgment task with low liability risk.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI assistance in material selection and research is relatively cheap, but this task is already low-cost for humans (administrative staff or faculty) and requires minimal labor. The human cost per task is already low, limiting savings.
Cost vs. human wage	claude-sonnet-5	4/5	Using an AI assistant to shortlist textbooks and supplies costs a fraction of the faculty time spent manually researching options, though final purchasing still requires some human administrative cost.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	No deployed product reliably performs full procurement end-to-end; existing systems only assist with research and comparison phases. Integration with institutional purchasing systems and approval hierarchies remains manual.
Technical feasibility today	claude-sonnet-5	3/5	General-purpose AI assistants and library/course-management tools can suggest materials, but the final procurement/ordering process still involves institutional systems and human approval not fully integrated with AI.

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

CI 38–50 · exposure 30 · 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	3/5	Academic institutions are adopting literature recommendation tools and AI-assisted research monitoring in research contexts, but uptake for active professional development remains moderate; many academics still prefer manual review and conference attendance despite growing tooling.
Sector adoption velocity	claude-sonnet-5	3/5	Academia has moderate AI adoption for literature review and research assistance, though conference/networking aspects remain largely traditional.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI strongly augments this task by rapidly filtering and summarizing thousands of articles, flagging conference papers, and personalizing recommendations, allowing faculty to stay current more efficiently while maintaining human judgment on what is most relevant to their teaching and research.
Augmentation potential	claude-sonnet-5	4/5	AI literature summarization, alerting services, and research assistants significantly speed up staying current with the field, even though full engagement still requires human participation.
Task automatability	claude-haiku-4-5-20251001	2/5	AI can read and summarize current literature and conference abstracts, but the nuanced collegial discussion and selective filtering of developments relevant to one's specific teaching context requires human judgment and relationship-building that current systems cannot fully replicate.
Task automatability	claude-sonnet-5	2/5	AI can help surface and summarize literature, but the core task of building expertise through reading, discourse, and conference participation requires genuine human engagement and judgment that cannot be fully offloaded.
Adoption barriers	claude-haiku-4-5-20251001	2/5	There are weak barriers to adoption: no licensing requirement to read literature, no legal mandate for human review, and professional communities increasingly use AI-assisted discovery. However, collegial relationships and conference networking retain informal social value that resists full automation.
Adoption barriers	claude-sonnet-5	2/5	No formal licensing barrier, but professional norms and tacit knowledge-sharing (colleague discussions, conferences) create moderate friction against full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	4/5	AI literature monitoring and summarization tools are inexpensive to deploy at scale (subscription services, open models), making the marginal cost per educator far lower than the time a human would spend manually reading journals and conference proceedings.
Cost vs. human wage	claude-sonnet-5	2/5	AI tools for literature summarization are cheap, but the task also includes networking and conference attendance which have no AI substitute, keeping overall cost comparable to human time.
Technical feasibility today	claude-haiku-4-5-20251001	3/5	AI-powered literature summarization and conference alert systems exist (e.g., paper recommenders, RSS aggregators), but they produce material error rates in filtering relevance and lack the conversational depth of genuine peer discussion that the task statement emphasizes.
Technical feasibility today	claude-sonnet-5	2/5	AI research assistants and summarization tools exist and are used to track literature, but they don't reliably replace conference networking or nuanced scholarly discussion.

Write grant proposals to procure external research funding.

CI 25–43 · exposure 30 · augmentation 75 · importance 3.0/5 · 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 are slow to adopt AI for research administration; most use remains experimental or supplemental. Postsecondary sectors lag far behind tech and finance in systematic AI agent deployment for knowledge work.
Sector adoption velocity	claude-sonnet-5	2/5	Academic research administration adopts AI writing tools cautiously and unevenly, with concerns about originality and grant integrity slowing deeper deployment despite general AI hype in academia.
Augmentation potential	claude-haiku-4-5-20251001	4/5	Current AI tools substantially assist grant writers by generating drafts, improving clarity, checking budget language, and brainstorming framing—raising productivity while the faculty member retains full control over research vision and proposal strategy.
Augmentation potential	claude-sonnet-5	4/5	AI is genuinely useful for drafting narrative sections, formatting, literature summaries, and refining language, meaningfully speeding up the proposal-writing process while the researcher retains ownership of content and strategy.
Task automatability	claude-haiku-4-5-20251001	2/5	Grant proposal writing requires domain expertise, narrative coherence, and alignment with funder priorities that vary significantly. While AI can draft sections and improve text, the strategic positioning, original research justification, and institutional context demand substantial human judgment and revision, falling short of the 50% time-saving bar for end-to-end execution.
Task automatability	claude-sonnet-5	2/5	AI can draft sections and generate boilerplate text, but crafting a competitive proposal requires original research framing, budget justification tied to specific institutional realities, and strategic knowledge of funder priorities that current AI cannot fully replicate end-to-end.'
Adoption barriers	claude-haiku-4-5-20251001	4/5	Grant writing is typically performed by faculty themselves or by dedicated grant-writing staff; there are no licensing barriers, but organizational norms, faculty autonomy over research direction, and institutional accountability for proposal quality create meaningful friction against substitution.
Adoption barriers	claude-sonnet-5	3/5	No licensing requirement, but funders and institutions expect the PI's original scholarly voice and accountability, and error costs (rejected funding, integrity issues) create real friction against full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	4/5	AI assistance costs (inference plus integration) are minimal compared to the loaded cost of a professor's time spent writing proposals, making automation assistance economically favorable even if only partial.
Cost vs. human wage	claude-sonnet-5	2/5	While drafting assistance is cheap, the extensive expert review, fact-checking, and strategic tailoring required to make a proposal competitive keeps the effective all-in cost close to or above human-only effort.
Technical feasibility today	claude-haiku-4-5-20251001	3/5	AI writing assistants (GPT, Claude) exist and are used in academic settings for proposal drafting and editing, but no deployed product reliably generates competitive, funding-ready proposals end-to-end without expert human oversight and substantial revisions.
Technical feasibility today	claude-sonnet-5	2/5	Writing assistants and LLMs are used to draft proposal text, but no deployed product reliably produces fundable, submission-ready grant proposals without heavy human revision and domain expertise.

Prepare and deliver lectures to undergraduate or graduate students on topics such as econometrics, price theory, and macroeconomics.

CI 25–34 · exposure 25 · augmentation 75 · importance 4.5/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Higher education is a laggard sector in AI adoption; while some faculty use AI drafting tools informally, institutional deployment is slow and cautious. No major university has publicly replaced instructors with AI lecturers, and faculty resistance remains high.
Sector adoption velocity	claude-sonnet-5	2/5	Higher education adopts AI tools slowly for core instructional delivery; while some experimentation exists, most postsecondary institutions still rely on human-delivered lectures with minimal production-scale AI substitution.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI assistants (summarization, outline generation, answer drafting, slide design) can significantly boost an instructor's preparation efficiency and help structure explanations, while the instructor retains pedagogical judgment and live teaching responsibility.
Augmentation potential	claude-sonnet-5	4/5	AI can meaningfully help instructors draft lecture notes, generate examples/problem sets, create visualizations, and prepare supplementary materials, boosting prep productivity while the instructor still delivers.
Task automatability	claude-haiku-4-5-20251001	2/5	AI can draft lecture content and generate slides on economic topics, but delivering engaging lectures requires real-time student interaction, responsiveness to classroom dynamics, and the ability to answer unpredictable questions—capabilities current systems cannot reliably handle end-to-end at equal quality. Preparation is automatable; delivery remains deeply human-dependent.
Task automatability	claude-sonnet-5	2/5	AI can help draft lecture content and slides, but live delivery, classroom interaction, adapting to student questions, and pedagogical presence still require a human instructor, so full end-to-end automation with equal quality isn't achieved today.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Academic accreditation, faculty employment law, and institutional governance create strong barriers: courses must be taught by qualified, credentialed faculty, and student/regulatory bodies expect human instruction. Full automation faces legal and contractual friction.
Adoption barriers	claude-sonnet-5	3/5	There is no strict licensing requirement to lecture, but institutional accreditation, tenure structures, and student/institutional expectations of a credentialed instructor create moderate friction against full replacement.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI lecture preparation tools are cheap per use, but the human instructor's wage remains the dominant cost driver in academic labor. Automation of drafting does not meaningfully reduce the institution's per-student teaching cost when the instructor must still deliver in person.
Cost vs. human wage	claude-sonnet-5	3/5	AI can cheaply assist in content generation, but actual delivery still requires paid faculty time, oversight, and course integration, making the all-in cost roughly comparable rather than dramatically cheaper.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	AI can generate lecture notes and outlines at scale, but no deployed product reliably delivers complete lectures (live teaching with Q&A, adaptive pacing, emotional presence) in a way that substitutes for the human instructor. ChatGPT can draft content; it cannot replace classroom instruction.
Technical feasibility today	claude-sonnet-5	2/5	Tools exist for generating lecture outlines, slides, and even AI-narrated video lectures, but no mature product reliably delivers full interactive university-level econ lectures in production at scale replacing instructors.

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

CI 25–34 · exposure 25 · augmentation 63 · 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	2/5	Higher education has been slow to adopt AI for core advising functions; most institutions still rely on human advisors and use AI only as a supplementary tool. Pilots exist but production-scale replacement is rare.
Sector adoption velocity	claude-sonnet-5	2/5	Higher education adopts AI slowly for personalized human-facing functions like advising, with most current use confined to pilot programs and administrative chatbots rather than academic mentoring itself.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can assist human advisors by quickly synthesizing curriculum information, suggesting career pathways, and highlighting relevant opportunities, improving the advisor's efficiency and breadth of input. However, the transformative potential is limited by the advice-giving task's core dependence on human judgment and relationship.
Augmentation potential	claude-sonnet-5	4/5	AI can help faculty advisors quickly look up program requirements, summarize course options, draft messages, and research career pathways, meaningfully speeding up their advising prep work.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can provide basic information about curricula and careers, advising requires understanding individual student circumstances, aspirations, and constraints—nuanced judgment that AI systems struggle to replicate reliably. Meaningful parts (e.g., curriculum information retrieval, career pathway data) could be automated, but the holistic advisory role requires human interpretation and does not meet the 50% time-saving threshold at equal quality.
Task automatability	claude-sonnet-5	2/5	Advising requires understanding a specific student's history, goals, institutional requirements, and building rapport, which current AI cannot fully replicate end-to-end despite being able to answer generic curriculum questions.'
Adoption barriers	claude-haiku-4-5-20251001	4/5	Institutions typically require that qualified human advisors (often faculty or trained staff) provide official academic and career guidance; student support expectations, accreditation norms, and liability concerns create strong organizational and regulatory friction against full AI substitution.
Adoption barriers	claude-sonnet-5	3/5	No formal licensing requirement blocks AI from giving advice, but institutional norms, liability for bad career guidance, and expectations of faculty mentorship create real friction.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Building and maintaining an AI advising system, including integration with student information systems and oversight mechanisms, is expensive relative to the wage cost of a single advising session. Marginal benefit does not yet justify replacement or substantial cost savings.
Cost vs. human wage	claude-sonnet-5	3/5	AI-based advising tools are cheap to run per query, but the professor's advising role is a small part of a bundled salary, so per-task savings are modest once oversight and integration are factored in.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Some standalone products (chatbots, career-matching tools) exist, but no deployed system reliably performs full academic and vocational advising that meets institutional standards and handles the diversity of student situations. Products that exist tend to have narrow scope or require heavy human oversight.
Technical feasibility today	claude-sonnet-5	2/5	Chatbot advising tools exist at some universities but are typically limited to FAQ-style navigation of degree requirements rather than substantive career or academic counseling matching human advisors.

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

CI 25–30 · exposure 25 · 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	Postsecondary institutions are conservative adopters of curriculum automation; adoption is limited to supplemental drafting aids rather than replacement of faculty curriculum authority. Governance structures and academic tradition slow deployment.
Sector adoption velocity	claude-sonnet-5	2/5	Higher education is a relatively slow-adopting sector for core academic functions like curriculum design, with pilots more common than production deployment.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI can meaningfully assist by generating draft syllabi, suggesting pedagogical approaches, analyzing course materials, and helping revise content—substantially raising faculty productivity in curriculum planning while they retain decision authority.
Augmentation potential	claude-sonnet-5	4/5	AI tools can meaningfully assist by generating draft syllabi, suggesting readings, creating practice problems, and summarizing student feedback, substantially speeding up parts of the curriculum development process while the instructor retains control.
Task automatability	claude-haiku-4-5-20251001	2/5	AI can assist with generating course outlines, drafting materials, and analyzing pedagogical methods, but curriculum planning requires expertise in learning outcomes, institutional constraints, and disciplinary depth that AI cannot reliably synthesize end-to-end. Human judgment on student needs and institutional context remains essential.
Task automatability	claude-sonnet-5	2/5	AI can draft or suggest curriculum elements and materials, but the actual planning, evaluation, and revision requires human judgment about pedagogical goals, institutional standards, and student needs, so it doesn't meet the 50% end-to-end automation bar today.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Curriculum design is a core faculty responsibility often governed by shared governance, accreditation standards, and institutional approval processes. Department committees and faculty senate approval create strong organizational and procedural barriers to full automation.
Adoption barriers	claude-sonnet-5	3/5	Curriculum design is generally not licensed work, but institutional accreditation, departmental governance, and academic freedom norms create meaningful organizational friction against full AI substitution.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	While AI drafting tools are cheap per use, the human oversight required to evaluate, revise, and implement curricula means total cost remains comparable to or exceeds having a faculty member do the work directly.
Cost vs. human wage	claude-sonnet-5	2/5	While AI drafting tools are cheap per output, the human oversight, subject-matter validation, and institutional approval processes needed keep overall costs comparable to or only modestly below faculty time costs.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Tools like ChatGPT can draft syllabi and lesson plans, but no deployed system reliably handles the full cycle of curriculum evaluation and revision at the depth required for economics instruction. Products lack domain expertise and cannot autonomously assess pedagogical effectiveness.
Technical feasibility today	claude-sonnet-5	2/5	Some ed-tech products offer AI-assisted lesson planning or content generation, but no deployed system reliably performs full curriculum evaluation and revision at scale in postsecondary economics departments.

Participate in student recruitment, registration, and placement activities.

CI 25–30 · exposure 25 · augmentation 50 · 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 recruitment and advising at scale; most institutions still rely on human admissions staff and career advisors, with AI adoption limited to supplementary tools like chatbots rather than end-to-end replacement.
Sector adoption velocity	claude-sonnet-5	2/5	Higher education administrative functions are adopting AI tools slowly, with pilots in chatbots and CRM systems but limited large-scale deployment for recruitment/placement tasks specifically.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can meaningfully assist recruitment staff by automating email distribution, flagging candidate fit, and managing registration scheduling, but the core relationship-building and placement counseling remain human-centered activities where AI provides useful but bounded support.
Augmentation potential	claude-sonnet-5	3/5	AI-powered CRM systems, application screening, and communication drafting tools can meaningfully assist staff in outreach and registration tasks, improving efficiency while humans retain decision-making roles.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can assist with email outreach, initial screening, and scheduling, the core recruitment and placement work requires human judgment about fit, relationship-building, and personalized outreach. The task fundamentally involves human judgment and interpersonal trust, preventing full end-to-end automation at 50% time savings.
Task automatability	claude-sonnet-5	2/5	This task involves interpersonal outreach, admissions decisions, and advising that require human judgment and relationship-building, though some administrative sub-steps (scheduling, communications) could be automated.'
Adoption barriers	claude-haiku-4-5-20251001	4/5	Higher education institutions have institutional commitments to human advising and recruitment roles; liability and reputational risk from poor placements, regulatory requirements around student records, and strong cultural preference for human contact in admissions create significant friction against full automation.
Adoption barriers	claude-sonnet-5	3/5	No strict licensing requirement, but institutional policy, accreditation expectations, and student preference for personal interaction create moderate friction against full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI tooling for recruitment (CRM, email automation, scheduling) requires infrastructure setup and ongoing oversight; total cost remains comparable to or sometimes exceeds hiring dedicated staff for these interpersonal tasks.
Cost vs. human wage	claude-sonnet-5	2/5	AI tools can cheaply handle routine outreach messaging, but the substantive human elements of counseling and placement decisions still require costly faculty/staff time, keeping overall cost comparable to human labor.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Limited deployed products handle this holistically; existing systems (CRM platforms, chatbots) handle narrow components like scheduling or basic inquiries but lack the contextual judgment needed for meaningful recruitment and placement counseling.
Technical feasibility today	claude-sonnet-5	2/5	Some CRM and chatbot tools assist with recruitment communications and registration logistics, but no deployed product handles the full recruitment/placement counseling role reliably.

Provide professional consulting services to government or industry.

CI 20–30 · exposure 20 · augmentation 75 · 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	Consulting sectors are conservative; pilots exist for AI-assisted research, but actual client-facing consulting remains human-led. Adoption of autonomous or near-autonomous AI consulting is slow due to liability, regulatory, and reputational risk aversion in these sectors.
Sector adoption velocity	claude-sonnet-5	2/5	Academic consulting and government/industry advisory work adopt AI tools slowly for drafting and research, but the core advisory relationship remains largely human-driven with limited production-scale AI deployment.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI augmentation is already substantial: literature reviews, data synthesis, scenario modeling, and report drafting are meaningfully assisted by AI tools. Economists use these to work faster, though the core judgment and client relationship remain human-centered.
Augmentation potential	claude-sonnet-5	4/5	AI can meaningfully speed up literature review, data analysis, report drafting, and scenario modeling, significantly boosting a consulting economist's productivity while they retain final judgment and client interaction.
Task automatability	claude-haiku-4-5-20251001	2/5	Consulting requires deep contextual analysis, stakeholder engagement, and bespoke recommendations tailored to specific institutions and policy constraints. While AI can draft reports or synthesize economic data, the interpersonal negotiation, trust-building, and accountability inherent in professional consulting cannot be meaningfully automated end-to-end at quality parity.
Task automatability	claude-sonnet-5	2/5	Consulting involves synthesizing client-specific context, judgment, negotiation, and relationship management that AI can assist but not fully replace end-to-end at equal quality today.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Government and industry consulting often requires explicit credentials, professional liability insurance, regulatory compliance (SEC, OMB), and contractual accountability that legally bind a named human expert. Clients demand human judgment and sign-off, creating strong adoption barriers.
Adoption barriers	claude-sonnet-5	3/5	While not formally licensed like law or medicine, consulting engagements often require credentialed academic expertise, institutional reputation, and accountability that create real friction to full AI substitution.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI infrastructure (data, compute, oversight) for generating credible economic analysis is non-trivial, but even if cheaper per unit, the liability and reputation costs of autonomous consulting are prohibitive, keeping human oversight costs high and limiting cost advantage.
Cost vs. human wage	claude-sonnet-5	2/5	AI can cut research and drafting time cheaply, but the full consulting engagement still requires substantial human expert oversight, client interaction, and liability-bearing judgment, keeping costs comparable to human-led work.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed product reliably performs professional economic consulting for government or industry at scale. AI tools can assist with research and drafting, but client-facing consulting requires licensed or credentialed economists accountable for advice, which remains a human-led practice in production.
Technical feasibility today	claude-sonnet-5	2/5	AI tools can generate analysis, reports, and data summaries, but no deployed product independently delivers professional economic consulting engagements to government or industry clients reliably.

Initiate, facilitate, and moderate classroom discussions.

CI 9–20 · exposure 16 · 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	1/5	Universities have shown minimal adoption of AI-driven classroom moderation, and institutional culture strongly favors human faculty presence in teaching. Adoption remains negligible even in highly digitized sectors of higher education.
Sector adoption velocity	claude-sonnet-5	2/5	Higher education has been slow to adopt AI for live teaching functions, with most current use confined to grading, content creation, and administrative support rather than real-time classroom facilitation.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI tools can assist faculty by generating discussion prompts, summarizing discussion threads asynchronously, or identifying quiet participants—useful productivity aids. However, these are ancillary to the core task of live moderation, which remains human-centric.
Augmentation potential	claude-sonnet-5	3/5	AI can help instructors prepare discussion questions, summarize readings, or suggest discussion prompts, offering moderate productivity gains in preparation even though it doesn't participate in live facilitation.
Task automatability	claude-haiku-4-5-20251001	2/5	AI cannot meaningfully replicate the dynamic, real-time facilitation and judgment required to guide a live classroom discussion—sensing participant engagement, responding to unexpected questions, managing group dynamics, and making pedagogical adjustments on the fly. While AI could draft discussion prompts or synthesize themes post-hoc, these represent only small portions of the full task.
Task automatability	claude-sonnet-5	2/5	Facilitating live, adaptive classroom discussion requires real-time reading of student engagement, improvisation, and social presence that current AI cannot replicate end-to-end. Some scripted prompts can be generated, but that is a small fraction of the live-moderation task.
Adoption barriers	claude-haiku-4-5-20251001	5/5	Accreditation and institutional policy typically require faculty presence and responsibility for classroom facilitation and student assessment. Liability, duty of care toward students, and the requirement that a credentialed human sign off on academic discussion and student learning are hard legal and regulatory barriers.
Adoption barriers	claude-sonnet-5	4/5	Accreditation, in-person instructional norms, and expectations of faculty presence in the classroom create strong institutional and pedagogical barriers to replacing a live human moderator.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	The cost of deploying a system to monitor, respond to, and moderate a live classroom discussion (including infrastructure, moderation oversight, and failure recovery) exceeds the cost of a faculty member leading the session, especially at postsecondary institutions with smaller classes.
Cost vs. human wage	claude-sonnet-5	2/5	Since AI cannot substitute for the live facilitation role, there is no cost savings from replacement; at best AI tools add a small cost on top of the instructor's time for prep, not a substitute for the labor itself.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed product reliably moderates live university-level classroom discussions end-to-end. Chatbots can respond to individual queries, but cannot simultaneously manage multiple participants, read the room, enforce psychological safety, or maintain coherent pedagogical flow in real time.
Technical feasibility today	claude-sonnet-5	1/5	There are no deployed products that reliably run live in-person classroom discussions in place of an instructor; AI is used at most for supplementary online discussion boards or chatbots, not as a moderator of a live seminar.

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

CI 0–16 · exposure 8 · 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	1/5	Academic institutions move slowly on automation and have strong cultural and regulatory attachment to human faculty mentorship; displacement of supervisory roles in higher education is minimal.
Sector adoption velocity	claude-sonnet-5	2/5	Higher education adopts AI tools slowly for supervisory and mentorship roles, though some pilot tools assist with feedback or scheduling.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI could usefully assist with logistics (scheduling, grade aggregation, progress tracking) and provide writing feedback drafts, but the core mentoring and research guidance remain heavily human-dependent tasks where augmentation is helpful but not transformative.
Augmentation potential	claude-sonnet-5	3/5	AI can help by drafting feedback on student work, tracking progress, or suggesting research resources, moderately aiding but not replacing the supervisory relationship.
Task automatability	claude-haiku-4-5-20251001	2/5	Supervising teaching, internship, and research involves substantial human judgment, mentoring, and relationship-building that AI cannot replicate end-to-end. While AI could assist with grading, scheduling, and documentation, the core supervisory and evaluative functions require human presence and accountability.
Task automatability	claude-sonnet-5	1/5	Supervising students' research, internships, and teaching requires relationship-based mentorship, judgment calls, and institutional accountability that current AI cannot execute end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	5/5	Institutional policy, accreditation standards, and faculty labor agreements typically require a licensed human faculty member to formally supervise student work, internships, and research; this is a hard legal and professional requirement.
Adoption barriers	claude-sonnet-5	4/5	Academic institutions require faculty of record to supervise theses, internships, and teaching for accreditation, evaluation, and legal/liability reasons, creating strong institutional and credentialing barriers.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	The all-in cost of an AI system to handle supervision (infrastructure, fine-tuning, human oversight, liability management) would far exceed the cost of a faculty member already employed for teaching and research duties.
Cost vs. human wage	claude-sonnet-5	1/5	There is no AI system replacing this function, so cost comparison favors the human entirely; any AI use is supplementary, not substitutive.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed product reliably supervises academic work, internships, or research at production scale. The task inherently requires human authority, responsibility for student learning outcomes, and institutional accountability that current AI systems cannot assume.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product performs student supervision autonomously; this remains a human faculty responsibility with no production substitute.

Maintain regularly scheduled office hours to advise and assist students.

CI 0–16 · exposure 0 · augmentation 38 · 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	1/5	Higher education has shown little momentum toward automating faculty office hours; institutions continue to mandate them and view them as core to teaching quality and student retention, with no evidence of AI displacement.
Sector adoption velocity	claude-sonnet-5	2/5	Higher education adopts AI tools slowly for direct student advising due to relationship-based norms, though usage of chatbots for basic Q&A is growing incrementally.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI could assist with minor administrative tasks (scheduling, sending reminders, providing canned answers to common questions), but the core advising and mentoring work requires the faculty member's judgment and presence, limiting the transformative potential.
Augmentation potential	claude-sonnet-5	3/5	AI can help by answering routine student questions, scheduling, and providing supplementary resources, freeing office hours for more substantive discussion, though it doesn't replace the interaction itself.
Task automatability	claude-haiku-4-5-20251001	1/5	This task fundamentally requires synchronous, real-time human interaction and interpersonal judgment—advising students on academic/career matters, responding to their specific concerns, and building mentoring relationships. Current AI cannot replace the dynamic, context-aware counseling and accountability that office hours provide.
Task automatability	claude-sonnet-5	1/5	This task requires synchronous, personalized human presence and mentorship; AI cannot maintain or substitute for a scheduled human office-hour presence and relational advising role.
Adoption barriers	claude-haiku-4-5-20251001	5/5	Accreditation standards and faculty employment contracts legally require scheduled office hours staffed by the instructor. Institutions and students expect direct faculty availability; removing human office hours would violate educational norms and regulatory expectations around student support.
Adoption barriers	claude-sonnet-5	3/5	No formal licensing barrier, but strong institutional/organizational norms and student expectations for human availability and personalized mentorship create real friction against replacement.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	The cost of AI systems (infrastructure, integration, moderation) would need to be weighed against faculty time savings, but faculty must maintain office hours as contractual and accreditation obligations—the human cost is already sunk and mandated.
Cost vs. human wage	claude-sonnet-5	2/5	While AI chat support is cheap, it cannot fully replace the task, so any cost comparison is for a partial substitute rather than the full task itself, keeping realized savings low.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed product reliably handles the full scope of postsecondary student advising and assistance in real office-hour settings. Chatbots can answer FAQs, but they cannot substitute for the nuanced guidance, course-planning decisions, and student relationship-building that characterize actual office hours.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product replaces the act of a professor holding office hours; chatbots exist for FAQs but not as a substitute for scheduled personal advising sessions.

Collaborate with colleagues to address teaching and research issues.

CI 0–13 · exposure 0 · 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	Academic institutions have shown minimal adoption of AI for faculty collaboration and governance roles; cultural and structural commitments to collegial process remain deeply entrenched in higher education.
Sector adoption velocity	claude-sonnet-5	2/5	Higher education is a relatively slow-adopting sector for AI in core faculty collaborative work, with pilots mostly in administrative or content-generation areas rather than collegial interaction itself.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI might assist with drafting agendas, summarizing prior discussions, or collating research literature for a collaboration meeting, but it cannot meaningfully augment the core deliberative work that defines this task.
Augmentation potential	claude-sonnet-5	3/5	AI can help draft shared documents, summarize research, prepare meeting agendas, or suggest curriculum ideas, usefully supporting but not replacing the collaborative process.
Task automatability	claude-haiku-4-5-20251001	1/5	Meaningful collaboration on teaching and research issues requires interpersonal negotiation, shared problem-solving, and contextual judgment about pedagogical and scholarly direction that AI cannot perform end-to-end. Current AI systems cannot reliably replace the deliberation and consensus-building central to this task.
Task automatability	claude-sonnet-5	1/5	Collaborative dialogue with colleagues on pedagogy and research direction requires genuine interpersonal judgment, relationship-building, and institutional context that current AI cannot replicate end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	5/5	Strong barriers exist: academic governance structures legally require faculty input and decision-making, institutional norms embed human deliberation in curriculum and research decisions, and organizational culture treats collaboration as a protected faculty prerogative.
Adoption barriers	claude-sonnet-5	3/5	No formal licensing barrier, but strong organizational and social norms mean colleague collaboration is expected to be human-led, creating moderate friction against automation.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	This task requires human judgment and relationship continuity; AI has no meaningful cost advantage because the task inherently demands human-to-human engagement. Attempting to substitute AI would eliminate the collaboration itself.
Cost vs. human wage	claude-sonnet-5	1/5	There is no AI substitute performing this task, so cost comparison favors the human doing the actual collaboration.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed product demonstrates reliable, independent performance of cross-colleague collaboration on academic issues. AI cannot authentically participate in departmental deliberations or represent a colleague's research interests and values.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product performs this collaborative, relational task on behalf of a faculty member; AI tools at best support preparation for such interactions.

Act as advisers to student organizations.

CI 0–11 · exposure 0 · augmentation 25 · importance 2.6/5 · 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	Higher education remains among the slowest sectors to adopt AI automation for instructional and mentoring roles. Student advising relationships are deeply valued in the academic mission and universities have not pursued automation of these advisory functions.
Sector adoption velocity	claude-sonnet-5	2/5	Higher education adopts AI unevenly and this specific relational/administrative task sees essentially no adoption momentum toward automation.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI could assist with administrative tasks like scheduling, document drafting, or information retrieval, but the core work—providing counsel, building trust, exercising judgment on student matters—offers limited scope for augmentation without displacing the human adviser's role.
Augmentation potential	claude-sonnet-5	2/5	AI can help with logistics like scheduling, drafting communications, or budget tracking for the organization, but it does not meaningfully enhance the core advisory/mentorship function.
Task automatability	claude-haiku-4-5-20251001	1/5	Advising student organizations requires nuanced relationship-building, understanding individual student circumstances, navigating organizational politics, and providing mentorship that fundamentally depends on human judgment and interpersonal trust. AI cannot meaningfully replace the core advisory function.
Task automatability	claude-sonnet-5	1/5	Advising a student organization requires ongoing relational trust, in-person mentorship, and real-time judgment calls that current AI cannot replicate end-to-end.4 It's inherently a human relationship role, not a discrete information task.
Adoption barriers	claude-haiku-4-5-20251001	5/5	Institutional policy, university liability, and accreditation standards typically require a credentialed faculty member to serve as official adviser and sign-off authority for student organizations. Legal and regulatory frameworks explicitly mandate human oversight.
Adoption barriers	claude-sonnet-5	4/5	Institutions typically require a designated faculty/staff advisor for liability, accreditation, and institutional governance reasons, creating strong organizational and quasi-regulatory barriers to substitution.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Even if AI could assist with logistical tasks (scheduling, documentation), the core advisory and mentoring function requires a human faculty member as the responsible party, making full cost displacement impossible. The marginal cost of AI assistance is low but human salary is unavoidable.
Cost vs. human wage	claude-sonnet-5	1/5	There is no AI substitute delivering this service, so cost comparison favors the human by default since AI cannot produce the output at all.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed product reliably performs the role of faculty adviser to student organizations, which requires sustained engagement, institutional authority, and accountability. This task is not addressable by current AI systems in any production capacity.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product performs the role of a faculty organizational advisor; this is not an area where research or commercial AI has attempted deployment.

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

CI 0–5 · exposure 0 · augmentation 38 · 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	1/5	Academic governance and committee structures are deeply entrenched in institutional tradition and legal/compliance frameworks. Adoption of AI for committee work itself is negligible; institutions have not moved toward algorithmic governance of these functions.
Sector adoption velocity	claude-sonnet-5	1/5	Higher education governance structures are slow-moving and largely unaffected by AI adoption trends seen in other sectors.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI could provide limited assistance by summarizing meeting materials, analyzing policy language, or drafting background documents for committee members to review. However, the core deliberative and voting functions resist augmentation, and such tools remain niche and underdeployed in academic settings.
Augmentation potential	claude-sonnet-5	3/5	AI can help prepare briefing materials, summarize policy documents, draft meeting minutes, or analyze data for committee discussions, offering moderate productivity assistance.
Task automatability	claude-haiku-4-5-20251001	1/5	Committee work requires nuanced judgment on institutional policies, stakeholder negotiation, and discretionary decision-making that cannot be automated end-to-end. While AI could assist with analysis or documentation, the deliberative and interpersonal core—voting, advocating positions, building consensus—is inherently human and unautomatable.
Task automatability	claude-sonnet-5	1/5	Committee service requires interpersonal negotiation, institutional judgment, and representing stakeholder interests in real-time deliberation, which current AI cannot perform end-to-end.'
Adoption barriers	claude-haiku-4-5-20251001	5/5	Committee membership is typically a formal institutional responsibility requiring a human faculty member with voting rights, fiduciary duties, and accountability to colleagues and the institution. Institutional bylaws, accreditation standards, and governance norms require human participation and decision-making authority.
Adoption barriers	claude-sonnet-5	4/5	Governance participation typically requires faculty status, institutional authorization, and shared governance rules that legally or contractually reserve this role for human faculty members.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	Committee service is a salaried responsibility bundled into faculty compensation; there is no discrete 'per-task' cost to displace. AI tools supporting committee work (e.g., document preparation) would need to be procured separately, making the all-in cost unlikely to be lower than the embedded labor already paid.
Cost vs. human wage	claude-sonnet-5	1/5	There is no viable AI substitute performing this role, so cost comparison favors the human by default.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed products perform committee membership, governance deliberation, or institutional decision-making as a standalone task. Committee roles require legal authority, institutional accountability, and human judgment that AI systems are not designed or authorized to exercise.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product substitutes for a human serving as a voting/participating committee member on institutional governance matters.

Participate in campus and community events.

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

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	1/5	This task is inherently bound to human presence and interpersonal dynamics; adoption of AI for event participation is not occurring and is not feasible in any sector.
Sector adoption velocity	claude-sonnet-5	1/5	Higher education institutional culture around community engagement changes slowly and there is no push toward automating physical presence at events.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI can minimally assist by suggesting relevant events, managing calendars, or drafting reflection notes post-event, but these are peripheral to the core task of actual participation and relationship-building.
Augmentation potential	claude-sonnet-5	2/5	AI could help with scheduling, event planning, or preparing talking points, but offers minimal assistance for the core act of attending and participating.
Task automatability	claude-haiku-4-5-20251001	1/5	Participating in campus and community events requires physical presence, social interaction, judgment about which events matter, and relationship-building—tasks that fall outside current AI capabilities. AI cannot attend events or meaningfully represent an institution in social contexts.
Task automatability	claude-sonnet-5	1/5	Physical presence and social participation in events cannot be performed by AI systems; this requires embodied human attendance and interaction.'
Adoption barriers	claude-haiku-4-5-20251001	5/5	Institutional and social norms strongly require human faculty presence at campus events to represent the department and build community relationships. There is no substitution pathway without fundamentally changing institutional expectations.
Adoption barriers	claude-sonnet-5	4/5	Community and institutional expectations require genuine human presence and relationship-building, an implicit organizational/social barrier even without formal licensing.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	AI cannot perform this task at all, making cost comparisons inapplicable. The task requires human presence and cannot be replaced by computational means.
Cost vs. human wage	claude-sonnet-5	1/5	There is no AI substitute providing this output, so no meaningful cost comparison exists; the human must be present.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed AI system can autonomously attend and participate in in-person events. While AI can help schedule or analyze event data, actual participation is beyond current technology.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product substitutes for a human's physical/social participation in campus or community events.

Perform administrative duties, such as serving as department head.

CI 0–0 · exposure 0 · 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	Academic institutions are conservative in governance; department head roles remain firmly human-controlled positions with minimal AI adoption because they require institutional legitimacy and legal accountability.
Sector adoption velocity	claude-sonnet-5	1/5	Higher education administrative leadership roles show negligible AI adoption for the core leadership function itself, despite AI tools being used for supporting tasks.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI can assist department heads with data analysis, email drafting, or scheduling logistics, but these are peripheral to the core leadership role, which remains human-centric and judgment-heavy.
Augmentation potential	claude-sonnet-5	3/5	AI can help with scheduling, drafting reports, summarizing data, and other supporting administrative subtasks, offering moderate productivity gains for the person holding the position.
Task automatability	claude-haiku-4-5-20251001	1/5	Serving as a department head involves strategic leadership, personnel management, budget decisions, and stakeholder negotiations that require human judgment, accountability, and institutional authority. Current AI cannot autonomously manage these responsibilities or legally represent a department.
Task automatability	claude-sonnet-5	1/5	Serving as department head involves complex interpersonal leadership, personnel decisions, budget authority, and institutional politics that AI cannot execute end-to-end today.'
Adoption barriers	claude-haiku-4-5-20251001	5/5	Academic department leadership is legally and institutionally bound to human faculty—university governance structures, accreditation bodies, and employment law require a human department head with contractual authority and accountability.
Adoption barriers	claude-sonnet-5	5/5	Department head roles typically require formal institutional appointment, tenure/faculty governance structures, and accountability that legally and organizationally require a human in the role.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	Department head roles are senior positions requiring experienced faculty with institutional credibility and legal authority. AI cannot substitute for this role at any cost; the functions demand human responsibility.
Cost vs. human wage	claude-sonnet-5	1/5	There is no AI system performing this administrative leadership role, so no meaningful cost comparison exists; a human must be paid regardless.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed AI product can reliably perform department head duties end-to-end. While AI can assist with scheduling or data analysis, it cannot make binding administrative decisions, hire staff, or hold formal leadership accountability.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product performs the role of department head; this remains a human leadership position with no automated substitute in production.

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.