Chemistry Teachers, Postsecondary

25-1052.00

Median wage $93,250/yr19,980 employed (US)Rank #484 of 923 scored · top 52% by substitution

Teach courses pertaining to the chemical and physical properties and compositional changes of substances. Work may include providing instruction in the methods of qualitative and quantitative chemical analysis. 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

Substitution28

Exposure25

Augmentation55

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

28 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

11%

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%26

panel mean rating 2.0/5 → substitution pressure 26/100

Technical feasibility todayw 20%24

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

Cost vs. human wagew 15%31

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

Adoption barriersw 20%inverted — strong barriers lower the score35

panel mean rating 3.6/5 (barrier strength) → substitution pressure 35/100

Sector adoption velocityw 10%23

panel mean rating 1.9/5 → substitution pressure 23/100

Task breakdown (28 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 4.4/5 · 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	Postsecondary institutions have been automating attendance and grade recording through LMS platforms for 15+ years; automation is now standard practice. Newer AI-assisted approaches (auto-attendance from engagement logs, predictive interventions) are being piloted broadly across the sector.
Sector adoption velocity	claude-sonnet-5	5/5	Higher education has near-universal adoption of digital LMS and gradebook systems for this exact administrative function already.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can assist instructors by auto-populating records from course activity (participation tracking, assignment submission timestamps) and flagging anomalies (grade outliers, missing records), reducing manual verification burden while the instructor retains oversight.
Augmentation potential	claude-sonnet-5	4/5	Automated systems significantly reduce faculty burden by calculating grades, flagging attendance issues, and generating reports, though instructors still review and finalize records.
Task automatability	claude-haiku-4-5-20251001	5/5	Attendance tracking, grade recording, and record-keeping are highly structured, rule-based tasks that integrate seamlessly with existing student information systems (Canvas, Blackboard, Banner, etc.). Current AI can fully automate these tasks with substantial time savings through API integration and automated data entry.
Task automatability	claude-sonnet-5	5/5	Recording and tallying attendance and grades is a structured data-entry task fully handled by existing LMS/gradebook software with automated calculations, meeting the ≥50% time savings bar easily.
Adoption barriers	claude-haiku-4-5-20251001	2/5	While FERPA compliance and institutional data governance policies require oversight, these are organizational friction points rather than legal prohibitions on automation. Most universities already delegate significant record-keeping to administrative staff and systems, so automation faces no hard regulatory barrier to the task itself.
Adoption barriers	claude-sonnet-5	2/5	Some institutional policy requires instructor sign-off on final grades, but the underlying record maintenance itself faces minimal regulatory or licensing barriers.
Cost vs. human wage	claude-haiku-4-5-20251001	5/5	The marginal cost of automating record-keeping via existing LMS or AI integration is negligible (cents per semester per student), whereas manual entry by an instructor costs hours of paid labor per course per term, making AI orders of magnitude cheaper per task-equivalent.
Cost vs. human wage	claude-sonnet-5	5/5	Software-based record-keeping costs a small fraction of the faculty time it would take to manually track attendance and grades, especially amortized across large enrollments.
Technical feasibility today	claude-haiku-4-5-20251001	5/5	Deployed products for student record management are mature and widely used in production across higher education institutions globally. Learning management systems with automated grade-book functions and attendance logging are standard infrastructure in postsecondary environments.
Technical feasibility today	claude-sonnet-5	5/5	Learning management systems (Canvas, Blackboard, PowerSchool, etc.) already perform automated attendance and gradebook management reliably at scale in production across universities.

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

CI 71–81 · exposure 70 · augmentation 100 · 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	3/5	Higher education is experiencing pilot adoption of AI-assisted course material generation, but deployment remains inconsistent. Many institutions lack clear policies; early adopters exist in tech-forward departments, but mainstream production use is still emerging rather than widespread.
Sector adoption velocity	claude-sonnet-5	3/5	Higher education is adopting generative AI for course prep at a moderate pace, with growing but uneven usage across departments and institutions.
Augmentation potential	claude-haiku-4-5-20251001	5/5	AI powerfully augments course material preparation by drafting initial templates, suggesting problem sets, and auto-formatting handouts, allowing instructors to focus on pedagogical refinement and customization rather than starting from blank pages.
Augmentation potential	claude-sonnet-5	5/5	AI substantially speeds up drafting of syllabi, assignments, and handouts, letting instructors focus on refining chemistry content and pedagogy rather than starting from scratch.
Task automatability	claude-haiku-4-5-20251001	4/5	AI can generate syllabi, homework assignments, and handouts at scale with high quality using prompts based on course objectives, textbooks, and learning outcomes. The output requires minimal human intervention and easily achieves >50% time savings while maintaining teaching standards.
Task automatability	claude-sonnet-5	4/5	LLMs can draft syllabi, homework problem sets, and handouts from a course outline or textbook chapter with substantial time savings, though chemistry-specific accuracy (formulas, reaction mechanisms) requires instructor review.
Adoption barriers	claude-haiku-4-5-20251001	2/5	Few hard legal or regulatory barriers exist; course materials do not require licensed sign-off. Adoption friction comes mainly from instructor preference to retain pedagogical control and institutional inertia, not from legal prohibition or liability asymmetry.
Adoption barriers	claude-sonnet-5	1/5	No licensing or legal requirement mandates a human author these documents; academic norms allow AI-assisted drafting with instructor oversight.
Cost vs. human wage	claude-haiku-4-5-20251001	5/5	API costs for generating course materials via AI are negligible (pennies per document) compared to an instructor's hourly wage for creating the same materials from scratch, making AI at least 10–100× cheaper all-in.
Cost vs. human wage	claude-sonnet-5	5/5	Generating a draft syllabus or problem set via AI costs cents in compute versus hours of faculty/TA time, making it dramatically cheaper even after review.
Technical feasibility today	claude-haiku-4-5-20251001	4/5	Mature LLM and document-generation tools are already deployed in educational settings to assist with course material creation. Products like ChatGPT, Claude, and specialized ed-tech platforms reliably produce usable syllabi and assignments, though instructors typically review and customize outputs.
Technical feasibility today	claude-sonnet-5	3/5	Products like ChatGPT, Claude, and course-authoring tools are widely used by instructors to draft materials, but reliability for technical chemistry content (equations, structures) is inconsistent and requires verification.

Compile bibliographies of specialized materials for outside reading assignments.

CI 71–81 · exposure 70 · augmentation 88 · 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 is adopting AI tools for course preparation, but adoption varies widely by discipline and institution. Many chemistry departments are piloting AI bibliography tools, but production adoption remains inconsistent.
Sector adoption velocity	claude-sonnet-5	3/5	Higher education is adopting AI research assistants and literature tools at a moderate pace, with pilots and individual faculty use common but not yet institutionalized broadly.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI excels at rapidly expanding bibliography scope and suggesting cross-disciplinary sources, augmenting instructor productivity significantly while the faculty member retains final curation and selection authority.
Augmentation potential	claude-sonnet-5	5/5	AI substantially speeds up literature discovery and organization, letting instructors focus on curating and evaluating quality rather than manual searching, a clear productivity boost while the human remains in control.
Task automatability	claude-haiku-4-5-20251001	4/5	AI systems can reliably search academic databases, identify peer-reviewed sources, and format citations automatically, achieving significant time savings. The task requires minimal domain-specific judgment beyond keyword matching and relevance filtering, which current tools handle well.
Task automatability	claude-sonnet-5	4/5	AI systems can quickly search literature, identify relevant papers/textbooks, and compile organized bibliographies on a chemistry topic, saving substantial time versus manual compilation.rating reflects need for final verification.
Adoption barriers	claude-haiku-4-5-20251001	2/5	Minimal legal or regulatory barriers exist; academic institutions are adopting these tools for administrative tasks. The main friction is instructor preference for human curation and institutional policies around AI use in course preparation.
Adoption barriers	claude-sonnet-5	1/5	There is no licensing, regulatory, or liability barrier to using AI for compiling reading lists; it's a low-stakes administrative/academic support task.
Cost vs. human wage	claude-haiku-4-5-20251001	5/5	API costs for AI literature search and citation formatting are negligible compared to a faculty member's hourly wage; a single query costs cents while manual compilation takes hours.
Cost vs. human wage	claude-sonnet-5	5/5	Generating a bibliography via AI tools costs a small fraction of a cent to a few dollars in compute versus a faculty member's hourly wage spent manually searching and compiling references.
Technical feasibility today	claude-haiku-4-5-20251001	4/5	Multiple deployed products (ChatGPT, Claude, specialized academic search tools) can generate formatted bibliographies and identify relevant chemistry literature with high reliability. Some human review is typical, but the core capability is mature and widely available.
Technical feasibility today	claude-sonnet-5	3/5	Tools like AI-assisted literature search (e.g., Semantic Scholar, Elicit, or LLMs with citation tools) are deployed and used by academics, but citation accuracy and relevance still require human review, limiting full reliability.

Prepare and submit required reports related to instruction.

CI 46–65 · exposure 58 · 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	While higher-education institutions are digitizing, adoption of AI-driven report automation specifically lags behind other sectors. Most colleges and universities still rely on manual or semi-automated (template-based) report processes with limited autonomous AI deployment.
Sector adoption velocity	claude-sonnet-5	2/5	Higher education administrative processes adopt AI tools slowly relative to tech/finance sectors, with LMS and reporting systems typically legacy and conservatively updated.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI can meaningfully assist instructors by auto-populating data fields, drafting narrative sections, formatting institutional templates, and flagging missing or inconsistent information, significantly reducing the clerical burden of report preparation while the instructor retains full oversight and sign-off authority.
Augmentation potential	claude-sonnet-5	4/5	AI can significantly speed up drafting narrative reports, summarizing data, and formatting submissions, letting instructors focus on verification rather than composition.
Task automatability	claude-haiku-4-5-20251001	3/5	Current AI systems can assist with drafting, formatting, and data compilation for routine instructional reports (attendance, grades, course assessments), but typically require human review for accuracy, context, and institutional compliance before submission. Full end-to-end automation would need reliable subject matter understanding and institutional knowledge.
Task automatability	claude-sonnet-5	4/5	Generating structured administrative reports (grades summaries, syllabi compliance, course outcome reports) from provided data is well within current LLM capability, especially with templates and prior examples.report generation and formatting is largely automatable with human review.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Academic institutions often have strict compliance requirements, accreditation standards, and institutional review procedures that mandate instructor review and signature on instructional reports. Faculty autonomy and liability concerns also create organizational friction against full automation.
Adoption barriers	claude-sonnet-5	2/5	No licensing requirement for report writing itself, but faculty are formally responsible for accuracy of grades/instructional records, creating light oversight friction.
Cost vs. human wage	claude-haiku-4-5-20251001	4/5	LMS and reporting automation tools cost institutions relatively little per report generated, and inference costs for AI drafting assistance are minimal compared to instructor salary per-report overhead. However, integration and compliance checks still require some human time.
Cost vs. human wage	claude-sonnet-5	4/5	Drafting reports via an LLM costs a small fraction of the faculty member's or administrator's time, though some data-gathering and final review still requires human labor.
Technical feasibility today	claude-haiku-4-5-20251001	3/5	Educational institutions use Learning Management Systems and administrative tools that can auto-generate portions of reports, but most deployed systems require manual data entry, review, and sign-off by instructors. Fully autonomous report generation without human oversight remains rare in production higher-ed environments.
Technical feasibility today	claude-sonnet-5	3/5	AI writing assistants and LMS-integrated tools can draft reports today, but few universities have deployed fully automated reporting pipelines that pull grade/attendance data and generate submission-ready reports without human assembly.

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

CI 35–59 · exposure 42 · augmentation 75 · 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	2/5	Higher education has been slow to adopt AI for assessment at scale; while individual instructors experiment with AI-aided question writing, systemic adoption in grading and exam administration remains limited due to concerns about plagiarism, fairness, and institutional risk.
Sector adoption velocity	claude-sonnet-5	3/5	Higher education has adopted AI grading tools and quiz generators in pilots and some production use (e.g., Gradescope, LMS-integrated tools), but broad deep adoption specifically for chemistry exam compilation/grading remains uneven.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI meaningfully assists faculty by drafting exam questions, generating rubrics, and auto-scoring objective questions, significantly reducing the time spent on routine administrative grading while instructors retain oversight of learning objectives and subjective evaluation.
Augmentation potential	claude-sonnet-5	4/5	AI substantially helps instructors draft question banks, generate rubrics, provide first-pass grading suggestions, and flag inconsistent grading, meaningfully speeding up the overall workflow while the instructor retains final control.
Task automatability	claude-haiku-4-5-20251001	2/5	AI can automate parts of this task—generating exam questions, creating answer keys, and basic grading of objective questions—but compiling exams still requires judgment about learning objectives and course-specific content. Administering exams in person is not automatable, and grading subjective responses (essays, problem-solving) requires nuanced human evaluation, limiting the time saving to well under 50% for the full end-to-end task.
Task automatability	claude-sonnet-5	3/5	AI can draft exam questions and grade objective or short-answer responses effectively, but compiling exams aligned to course-specific learning objectives and grading complex chemistry problem-solving (with partial credit, work shown) still requires human judgment for full end-to-end automation.
Adoption barriers	claude-haiku-4-5-20251001	3/5	Institutions typically expect faculty to maintain direct control over assessment design and integrity; there is no licensing barrier, but organizational norms around academic integrity and the desire to preserve faculty autonomy in exam composition create moderate friction against full automation.
Adoption barriers	claude-sonnet-5	2/5	No licensing requirement mandates a human grade exams, though institutional academic integrity policies and faculty accountability for grades create moderate friction against full delegation to AI.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Automating portions (question generation, objective grading) is cheaper than human labor, but the overhead of prompt engineering, review, and oversight for each exam, plus the continued need for human proctoring and subjective grading, makes the all-in cost comparable to or only modestly below a TA or instructor's time.
Cost vs. human wage	claude-sonnet-5	4/5	AI-assisted question generation and automated grading of multiple-choice/short-answer items is dramatically cheaper per exam than faculty or TA time, though complex free-response grading still needs human cost layered in.
Technical feasibility today	claude-haiku-4-5-20251001	3/5	Deployed products (e.g., LLMs for question generation, automated test platforms) exist and perform parts of this task, but they produce inconsistent quality for course-aligned exams and require significant human oversight. No mature system reliably handles the entire compilation-administration-grading pipeline without material error rates or missing context.
Technical feasibility today	claude-sonnet-5	3/5	Products like Gradescope and AI-assisted quiz generators are deployed in real classrooms today, but reliable grading of open-ended chemistry work (structures, mechanisms, calculations) still has meaningful error rates requiring instructor review.

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

CI 25–59 · exposure 38 · augmentation 63 · importance 4.6/5 · 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, especially STEM, has historically lagged in automation adoption and maintains strong norms around faculty autonomy in assessment. While some departments pilot AI-assisted grading, widespread production use for chemistry lab and paper evaluation remains rare and localized.
Sector adoption velocity	claude-sonnet-5	3/5	Higher education is adopting AI grading tools unevenly—common in large intro courses and MOOCs, but slower in specialized upper-level science courses with lab components.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can meaningfully assist instructors by drafting initial feedback on written assignments, flagging common errors, or organizing objective rubric scoring, which could speed up the grading process while instructors retain full judgment. However, the task's emphasis on laboratory performance evaluation limits the augmentation potential.
Augmentation potential	claude-sonnet-5	4/5	AI substantially speeds up grading of quizzes, lab reports, and written work by pre-scoring and flagging issues, letting instructors focus review time on edge cases and feedback quality.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can assist with objective components like grading multiple-choice or formula-based lab calculations, evaluating laboratory performance and assessing conceptual understanding in papers requires nuanced human judgment that AI systems cannot reliably replicate end-to-end. Current systems lack sufficient context about individual student progress and safety protocols to replace the full grading task with ≥50% time savings at equal quality.
Task automatability	claude-sonnet-5	3/5	AI can grade written assignments, quizzes, and structured lab reports reasonably well, but nuanced grading of lab technique, original scientific reasoning, and partial credit judgment still requires human oversight, so only partial time savings are achievable end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Institutional and accreditation policies typically require faculty sign-off on all grades; moreover, liability and student-grievance concerns create strong friction against delegating evaluation to AI without substantial human review, effectively preserving the instructor's gatekeeping role.
Adoption barriers	claude-sonnet-5	2/5	No licensing requirement mandates a human grade coursework, though institutional academic integrity policies and grade appeal processes create moderate friction requiring instructor accountability for final grades.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	An instructor's hourly wage for grading is relatively modest, and the AI infrastructure cost (API calls, platform subscriptions, human review/correction overhead) does not yet undercut instructor time significantly, especially when accounting for the need for human oversight of grades.
Cost vs. human wage	claude-sonnet-5	4/5	AI grading tools cost a fraction of instructor/TA hourly wages for high-volume assignments, though human spot-checking is still needed, moderating the full cost advantage.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Deployed AI tools can assist with objective rubric scoring and basic written feedback, but no production system reliably evaluates the full spectrum of postsecondary chemistry assessment (lab safety, experimental technique, conceptual reasoning in papers) at the quality required for institutional grading. Most commercial tools remain narrow and require heavy human oversight.
Technical feasibility today	claude-sonnet-5	3/5	Products like Gradescope, Turnitin, and LLM-based grading assistants are deployed in real courses for grading essays and structured problems, but reliability drops for open-ended lab work and instructors still review/adjust scores.

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

CI 31–50 · exposure 30 · augmentation 75 · 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	3/5	Academic institutions are adopting literature-review tools and alert systems at moderate pace, but human participation in conferences and collegial discussion remains deeply embedded in academic practice and hiring/promotion criteria.
Sector adoption velocity	claude-sonnet-5	3/5	Academic and research sectors are moderately adopting AI tools for literature review and summarization, though conference attendance remains a slower-to-change norm.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI tools that summarize recent papers, highlight key findings, and alert to relevant conferences substantially augment a chemist's ability to stay current without replacing the need for critical reading and professional judgment.
Augmentation potential	claude-sonnet-5	4/5	AI tools like literature summarizers, alert systems, and research assistants substantially help chemistry teachers stay current more efficiently.
Task automatability	claude-haiku-4-5-20251001	2/5	AI can assist with literature scanning and summarization, but staying current requires judgment about significance, relevance to one's research/teaching, and contextual understanding that varies by institution and student population. The human must ultimately curate and interpret.
Task automatability	claude-sonnet-5	2/5	AI can help summarize and surface literature, but 'keeping abreast' involves ongoing human judgment, networking, and synthesis that isn't fully substitutable end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Professional development and staying current in one's field are largely intrinsic expectations of academic employment; there is no licensing barrier, but organizational culture and disciplinary norms create strong implicit requirements for human engagement in the scholarly community.
Adoption barriers	claude-sonnet-5	2/5	No licensing barrier, but professional norms, tenure expectations, and the social/networking value of conferences create some friction against pure automation.
Cost vs. human wage	claude-haiku-4-5-20251001	4/5	Literature summarization and content filtering via AI is cheap compared to a professor's time spent manually scanning journals and conference programs, though human curation and interaction costs remain.
Cost vs. human wage	claude-sonnet-5	3/5	AI-assisted literature review is cheap, but professional conferences and colleague interaction have costs not easily replaced by AI at lower cost.
Technical feasibility today	claude-haiku-4-5-20251001	3/5	Products like literature-review tools and conference-alert systems exist and can scan publications reliably, but they cannot fully replace the exploratory reading, collegial discussion, and networking judgment that constitute staying 'abreast' in an active field.
Technical feasibility today	claude-sonnet-5	2/5	Tools like AI literature summarizers and research digest apps exist but are not widely deployed as complete replacements for professional engagement and conference participation.

Participate in student recruitment, registration, and placement activities.

CI 30–34 · 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 is generally a laggard in AI adoption outside administrative systems. While some institutions pilot chatbots for registration support, deep automation of recruitment and placement remains rare; most institutions rely on faculty and admissions staff for these strategic functions.
Sector adoption velocity	claude-sonnet-5	2/5	Higher education is a relatively slow-adopting sector for AI in interpersonal advising and recruitment tasks compared to finance or tech, though administrative chatbots are spreading.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can meaningfully assist with data management, matching students to opportunities, scheduling, and preliminary communication filtering. However, the core tasks—relationship-building, counseling, and personalized placement advice—remain human-led, with AI providing useful but non-transformative support.
Augmentation potential	claude-sonnet-5	3/5	AI can assist with drafting recruitment materials, scheduling, tracking applicant data, and generating placement analytics, meaningfully aiding faculty who remain in the loop.
Task automatability	claude-haiku-4-5-20251001	2/5	Student recruitment, registration, and placement require significant human judgment, relationship-building, and understanding of individual student circumstances. Current AI can assist with administrative tasks like sending bulk emails or organizing registration data, but cannot autonomously perform the relationship-driven, personalized outreach and decision-making that characterizes effective recruitment and placement.
Task automatability	claude-sonnet-5	2/5	This task involves relational, evaluative, and institutional coordination work (interviewing prospects, advising on placement, representing the program) that AI can support but not perform end-to-end at equal quality.rat
Adoption barriers	claude-haiku-4-5-20251001	3/5	Postsecondary institutions have established preference for faculty involvement in student recruitment and placement to maintain institutional relationships and credibility. There are no hard legal barriers, but organizational friction and stakeholder expectations (students, employers, accreditors) protect human roles here.
Adoption barriers	claude-sonnet-5	3/5	No licensing requirement, but institutional norms, personal relationships with prospective students, and faculty governance roles create moderate friction against full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	3/5	AI-driven CRM and registration systems can reduce administrative overhead, bringing costs closer to parity with human labor. However, relationship-intensive recruitment and placement counseling still requires human expertise, keeping overall cost savings modest and aligned with human wage costs.
Cost vs. human wage	claude-sonnet-5	2/5	Some cost savings from automating communications and scheduling, but the substantive recruitment/placement judgment still requires paid faculty time, limiting overall savings.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	While AI systems exist for basic student communication (chatbots, automated emails) and data management, no deployed product reliably performs end-to-end recruitment and placement with the personalization, institutional knowledge, and judgment these tasks demand. Chatbots handle narrow inquiries, but cannot strategically recruit or place students.
Technical feasibility today	claude-sonnet-5	2/5	CRM and chatbot tools exist for initial outreach and FAQ handling in admissions, but the faculty-specific recruitment/placement judgment work is not handled by deployed AI products today.

Write grant proposals to procure external research funding.

CI 25–39 · exposure 33 · 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	2/5	Adoption in academic research institutions remains limited to assistive drafting tools; production replacement of grant writing is rare. Academic culture privileges researcher ownership and credibility, and risk-averse funding environments discourage delegation to AI-authored proposals.
Sector adoption velocity	claude-sonnet-5	2/5	Academic research culture adopts AI writing tools cautiously and unevenly, with many funders and universities issuing restrictive AI-use policies for grant writing.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI tools can substantially assist in literature summarization, outline generation, formatting, and editing of grant proposals, meaningfully raising a researcher's drafting speed while they retain full control over novelty claims, research direction, and strategic framing.
Augmentation potential	claude-sonnet-5	4/5	AI is widely useful for brainstorming, literature summaries, editing prose, and formatting, meaningfully speeding up the drafting process while the PI retains ownership of ideas and final content.
Task automatability	claude-haiku-4-5-20251001	2/5	Grant proposal writing requires substantive research narrative, original framing, and strategic alignment with funder priorities. While AI can assist with sections (literature synthesis, formatting), human judgment on novelty, significance claims, and competitive positioning is essential, and AI alone cannot reliably meet the >50% time-saving threshold at equal quality for full proposals.
Task automatability	claude-sonnet-5	3/5	AI can draft substantial portions of grant proposals (background, literature synthesis, structure) but the novel research ideas, budget justification, and institutional specifics still require significant human input, so only partial time-saving is realistic today.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Grant proposal quality and researcher credibility are legally and reputationally tied to the PI's signature and institutional standing. Institutions and funders hold the submitting researcher liable for proposal content, creating a hard barrier to full automation of research claims and strategy.
Adoption barriers	claude-sonnet-5	3/5	No formal licensing requirement, but funders and institutions expect PI authorship, accountability, and expertise, and misrepresentation or plagiarism risk creates real institutional friction.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI inference cost is low, but integration requires expert oversight to validate research claims, ensure funder alignment, and revise output. Total cost (tool + human review) remains comparable to or exceeds having a faculty member or grant writer draft directly.
Cost vs. human wage	claude-sonnet-5	2/5	AI can reduce drafting time cheaply, but the overall cost is still dominated by expert oversight, fact-checking, and iteration to meet funder-specific requirements, so total savings versus a PI's time are modest.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	AI writing tools exist and can generate proposal text, but deployed systems lack the domain-specific research credibility assessment and funder-match judgment needed to produce fundable proposals. Products generate drafts only; production use still requires substantial expert human revision.
Technical feasibility today	claude-sonnet-5	2/5	Some faculty use LLMs to draft sections of proposals, but no mature product reliably produces fundable, compliant, and accurate grant proposals end-to-end in production.

Write letters of recommendation for students.

CI 4–54 · 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	1/5	Academic institutions remain heavily reliant on human-authored letters and have shown strong resistance to AI-generated recommendations due to authenticity and liability concerns; adoption of AI for generating these letters in production remains negligible.
Sector adoption velocity	claude-sonnet-5	3/5	Academic professionals are increasingly using generative AI for administrative writing tasks like this, but adoption is uneven and often informal rather than tracked or institutionally sanctioned.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI might assist with drafting structure or grammar checking, but cannot augment the core value of the letter—assessing the student's actual merits—since the teacher must still perform the essential judgment work independent of AI scaffolding.
Augmentation potential	claude-sonnet-5	5/5	AI substantially speeds up drafting, structuring, and polishing recommendation letters when the professor supplies key facts, making it a strong productivity tool while the human retains final judgment and signs their name.
Task automatability	claude-haiku-4-5-20251001	1/5	Writing letters of recommendation requires deep knowledge of a specific student's capabilities, character, academic performance, and potential that only a teacher who has directly observed and interacted with that student can credibly provide. AI cannot authentically assess individual student merit or generate personalized endorsements that meet institutional standards.
Task automatability	claude-sonnet-5	3/5	AI can draft strong letter templates from provided details about a student's performance, but genuine personal knowledge, specific anecdotes, and the instructor's authoritative voice still require substantial human input and review, limiting full end-to-end automation.
Adoption barriers	claude-haiku-4-5-20251001	5/5	Letters of recommendation carry significant liability and legal/ethical requirements: the recommender must personally vouch for claims, institutions expect human judgment and accountability, and many accreditation bodies and hiring panels treat AI-authored letters with deep skepticism or explicit rejection policies.
Adoption barriers	claude-sonnet-5	3/5	Recommendation letters carry an implicit expectation that the named recommender personally knows and vouches for the student, creating an integrity/authenticity norm that discourages full automation even though no formal license is required.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	While AI inference cost is low, the letter must still be written by the teacher (who bears liability for its content and accuracy), meaning AI cannot reduce the human labor cost below current levels—the teacher cannot outsource their professional judgment and signature.
Cost vs. human wage	claude-sonnet-5	4/5	Drafting a letter with an LLM costs pennies in compute versus the substantial time a professor would spend writing from scratch, making AI drafting assistance far cheaper even after review time.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed AI product can generate genuine, institution-accepted letters of recommendation that claim personal knowledge of a specific student's abilities and character. Such letters require human authentication and would face immediate rejection if transparently AI-generated.
Technical feasibility today	claude-sonnet-5	3/5	Writing assistants like ChatGPT are commonly used by faculty to draft or polish recommendation letters, but reliability depends heavily on the human supplying accurate, specific details, so it's an assistive product rather than an autonomous one.

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

CI 25–30 · exposure 25 · augmentation 63 · 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	Postsecondary education remains relatively slow in AI adoption for core academic functions; while some institutions pilot AI for content drafting, curriculum planning remains faculty-driven and resistant to automation. Adoption is nascent, mostly in pilot form, with little displacement of curriculum work.
Sector adoption velocity	claude-sonnet-5	2/5	Higher education is a slower-adopting sector for AI in core academic functions like curriculum design, with pilots more common than widespread production use.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can meaningfully assist faculty by generating draft materials, suggesting instructional approaches, and synthesizing research on pedagogy, thereby raising efficiency in content creation and material revision. However, the core tasks of evaluating curricula against learning outcomes and making pedagogical trade-offs remain human-centered, limiting transformative upside.
Augmentation potential	claude-sonnet-5	4/5	AI tools can meaningfully assist in brainstorming course content, generating practice problems, summarizing research, and drafting materials, significantly aiding instructors who retain final judgment.
Task automatability	claude-haiku-4-5-20251001	2/5	AI can assist with generating draft course materials, outlines, and instructional methods, but curriculum planning and evaluation require deep understanding of pedagogical goals, student outcomes, institutional context, and subject-matter expertise. End-to-end automation with 50% time savings at equal quality is not achievable today; human judgment and revision remain essential.
Task automatability	claude-sonnet-5	2/5	AI can draft syllabi or suggest topics, but genuinely planning, evaluating, and revising a chemistry curriculum requires disciplinary judgment, institutional alignment, and pedagogical expertise that current tools cannot fully replace end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Postsecondary faculty retain professional authority and institutional governance over curriculum; accreditation bodies, department chairs, and disciplinary standards require human faculty judgment and sign-off on curricula. Academic freedom and professional norms create strong organizational friction against full automation of curriculum design.
Adoption barriers	claude-sonnet-5	3/5	No licensing barrier prevents AI assistance, but institutional accreditation, faculty governance, and academic freedom norms create meaningful organizational friction against full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	LLM-based content generation is inexpensive per token, but the integration cost, human review time, and oversight required to ensure pedagogical soundness and curricular coherence mean the total cost of AI-assisted planning approaches the loaded wage of a faculty member doing the work directly.
Cost vs. human wage	claude-sonnet-5	2/5	While AI can cheaply generate draft content, the human oversight, subject-matter validation, and institutional review needed still make the all-in cost comparable to or only modestly cheaper than faculty time.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	AI systems can generate course outlines and teaching material drafts (via LLMs), but no deployed product reliably evaluates curricula or revises instruction methods in ways that meet accreditation and learning-outcome standards autonomously. Products exist for supplementary content generation but not for the full planning-evaluation-revision cycle.
Technical feasibility today	claude-sonnet-5	2/5	Some AI-assisted curriculum design tools exist but are not widely deployed as reliable, standalone solutions for postsecondary chemistry curriculum revision in production settings.

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

CI 25–30 · exposure 25 · augmentation 50 · 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	2/5	Higher education adoption of AI for core advising functions remains in early stages (pilots, chatbots for triage); most institutions maintain human advisors as the primary point of contact. Displacement is minimal and velocity is slow, constrained by regulatory and quality concerns.
Sector adoption velocity	claude-sonnet-5	2/5	Higher education adoption of AI for advising is still in pilot stages at most institutions, lagging behind fast-moving sectors like finance or tech despite growing interest in chatbot-based student services.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can usefully assist advisors by surfacing relevant program data, synthesizing job market information, and drafting templates or talking points, raising advisor productivity on information-intensive parts of the task while the human retains the relationship and judgment role.
Augmentation potential	claude-sonnet-5	3/5	AI can help faculty advisors by summarizing degree requirements, drafting career resources, or answering routine questions, freeing time for higher-value personalized conversations.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can generate generic career information and curriculum suggestions, advising on academic and vocational pathways requires understanding individual student goals, aptitudes, constraints, and institutional options—nuanced judgment that current AI systems cannot reliably perform end-to-end. Meaningful portions could be automated (info retrieval, initial drafting), but not to the 50% time-savings threshold with equal quality.
Task automatability	claude-sonnet-5	2/5	Advising involves personalized judgment, knowledge of institutional requirements, and relational trust-building that current AI cannot fully replicate end-to-end, though chatbots can handle basic FAQ-style guidance.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Postsecondary institutions typically require that academic and career advising be performed or directly supervised by a credentialed faculty member or advisor due to accreditation, duty-of-care, and fiduciary expectations; there is also significant liability risk in misadvising students on career and curriculum. These create meaningful legal and professional barriers to full automation.
Adoption barriers	claude-sonnet-5	3/5	No strict licensing requirement mandates a human advisor, but institutional norms, liability concerns around inaccurate guidance, and student preference for human mentorship create moderate friction.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Even when limited to information retrieval and generic guidance, the oversight and customization required to make AI output actionable for students remains labor-intensive, making the all-in cost (inference + integration + validation) comparable to or higher than employing a part-time advisor.
Cost vs. human wage	claude-sonnet-5	2/5	While AI chat tools are cheap per interaction, effective advising requires integration, escalation paths, and human oversight for edge cases, keeping total cost comparable to or only modestly below a professor's time for this task.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	No deployed product reliably performs personalized academic and career advising at the level expected of a postsecondary educator. Chatbots can provide generic advice, but production systems do not yet trustworthily assess individual fit, navigate institutional policies, or integrate institution-specific curriculum knowledge in ways that would satisfy institutional needs.
Technical feasibility today	claude-sonnet-5	2/5	Some universities deploy AI advising chatbots for scheduling or basic curriculum questions, but nuanced career and academic advising in specialized fields like chemistry remains largely human-driven with narrow-scope tools only.

Select, order, and maintain materials and supplies for teaching and research, such as textbooks, chemicals, and laboratory equipment.

CI 25–30 · exposure 25 · augmentation 50 · 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	2/5	While universities use general procurement systems, chemistry departments remain cautious about automating material selection and maintenance due to safety-critical nature; adoption of AI for this specific task is minimal and limited to routine reordering of standard items.
Sector adoption velocity	claude-sonnet-5	2/5	Higher education administrative and lab management functions have been slow to adopt AI-driven procurement tools compared to sectors like finance or software.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI could assist by recommending suppliers, tracking inventory levels, or flagging expired materials, but the human instructor must remain central to selecting appropriate chemicals and validating equipment specifications for pedagogical and safety reasons.
Augmentation potential	claude-sonnet-5	3/5	AI can assist with drafting supply lists, tracking usage patterns, and suggesting reorder timing, providing moderate productivity gains while humans retain control of ordering and compliance.
Task automatability	claude-haiku-4-5-20251001	2/5	Ordering materials can be partially automated (inventory tracking, purchase requisitions), but selecting appropriate chemicals and equipment requires domain expertise to match course/research needs, and maintaining supplies involves physical inspection and hands-on assessment that current AI cannot reliably perform end-to-end.
Task automatability	claude-sonnet-5	2/5	AI can help identify and compile lists of chemicals, textbooks, and equipment, but ordering involves budget approval, vendor coordination, safety compliance, and physical inventory checks that require human judgment and action.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Chemical safety regulations (OSHA, EPA), institutional procurement policies, liability for hazardous material handling, and requirement for qualified personnel sign-off on equipment purchases create substantial legal and organizational barriers to full automation.
Adoption barriers	claude-sonnet-5	3/5	Chemical procurement involves safety regulations, budget authorization, and institutional purchasing policies, creating moderate friction against full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Integration of procurement automation and inventory systems has upfront costs; the task involves low-volume, specialized purchases that don't achieve order-of-magnitude cost savings compared to a part-time staff member managing supplies.
Cost vs. human wage	claude-sonnet-5	2/5	AI assistance can speed up list generation but the overall workflow (approvals, physical handling, safety checks) still requires human labor, so total cost savings are modest.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Some e-procurement systems can automate ordering workflows, but no deployed AI system reliably handles the full task of selecting specialized chemistry materials, verifying safety compliance, and managing physical inventory in lab environments at production scale.
Technical feasibility today	claude-sonnet-5	2/5	No deployed product autonomously manages lab procurement and inventory for a chemistry department; existing lab management software still requires substantial human input and oversight.

Prepare and deliver lectures to undergraduate or graduate students on topics such as organic chemistry, analytical chemistry, and chemical separation.

CI 14–34 · exposure 25 · augmentation 63 · importance 4.8/5 · 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	Despite recent investment in educational technology, postsecondary institutions remain slow to adopt AI for core instruction; lectures continue to be delivered by human faculty, and adoption of AI lecturing systems in production is negligible across institutions.
Sector adoption velocity	claude-sonnet-5	2/5	Higher education adopts AI tools for content support and some course design, but classroom lecture delivery itself remains a slow-adopting, human-centered practice.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can usefully assist instructors by generating practice problems, automatically grading assignments, providing slide templates, or offering real-time explanation support—moderately boosting preparation and grading efficiency while the instructor remains the primary deliverer of live instruction.
Augmentation potential	claude-sonnet-5	4/5	AI substantially helps in preparing lecture content, generating examples, explanations, quizzes, and visual aids, meaningfully boosting instructor productivity while they retain teaching responsibilities.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can generate lecture outlines and draft explanations of chemical concepts, delivering effective lectures requires real-time interaction, responsiveness to student confusion, pedagogical judgment about pacing and emphasis, and live demonstration coordination—tasks that demand human presence and adaptability. Current systems cannot reliably replicate the full lecture experience with 50% time savings at equal educational quality.
Task automatability	claude-sonnet-5	2/5	AI can help draft lecture content and slides, but live delivery, adapting to student questions, and pedagogical judgment in a classroom setting remain largely human-performed activities not meeting the 50% end-to-end substitution bar.
Adoption barriers	claude-haiku-4-5-20251001	5/5	Postsecondary teaching is legally and institutionally bound to credentialed faculty; accreditation bodies, tenure systems, and institutional governance require human instructors. Liability, quality assurance, and regulatory standards in higher education create hard barriers to AI substitution.
Adoption barriers	claude-sonnet-5	3/5	While no license is required to use AI-generated content, universities require credentialed faculty for instruction, accreditation standards, and student expectations of human interaction create meaningful organizational friction.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	The cost of developing, fine-tuning, and continuously updating an AI lecture system, combined with oversight and technical infrastructure, significantly exceeds the direct cost of paying an adjunct or tenure-track instructor for their teaching hours when amortized.
Cost vs. human wage	claude-sonnet-5	3/5	AI can cheaply generate lecture materials and explanations, but the overall task still requires an instructor's salary for delivery, oversight, and interaction, keeping costs roughly comparable when the full task is considered.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	AI can produce lecture notes, slides, and pre-recorded video segments, but no deployed product reliably substitutes for a live instructor managing a classroom, fielding questions, and adapting explanations. Educational AI tools exist but fill supportive rather than replacement roles in actual postsecondary teaching environments.
Technical feasibility today	claude-sonnet-5	2/5	AI tools (e.g., ChatGPT, lecture-generation software) exist for content creation, but no deployed product reliably delivers full postsecondary chemistry lectures in real classrooms at scale.

Clean laboratory facilities.

CI 10–28 · exposure 8 · augmentation 13 · 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	Academic laboratory facilities are traditional, low-digitization environments with small budgets and slow infrastructure change cycles. Adoption of cleaning automation in this sector has been minimal, and most institutions continue relying on human custodial staff.
Sector adoption velocity	claude-sonnet-5	1/5	Physical facility cleaning in academic labs sees essentially no AI/robotic adoption currently; this is a low-digitization, low-priority automation target.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI/robotic assistance for lab cleaning is limited today; there exist no widely deployed systems that augment human cleaning productivity in academic labs. Scheduling and inventory tools offer minor support, but core cleaning work remains purely human.
Augmentation potential	claude-sonnet-5	1/5	AI offers essentially no meaningful assistance for the physical act of cleaning lab facilities, though it might help generate cleaning checklists or safety protocols as a minor adjacent benefit.
Task automatability	claude-haiku-4-5-20251001	2/5	Cleaning laboratory facilities involves navigating complex, hazardous environments with fragile equipment, specialized chemical handling, and variable contamination contexts. While robotic systems exist for some industrial cleaning, they lack the dexterity, safety awareness, and adaptability required for typical lab settings, and would require extensive customization per facility.
Task automatability	claude-sonnet-5	1/5	Cleaning laboratory facilities is a physical manipulation task requiring handling of chemical residues, glassware, and equipment in varied configurations; no current AI system can perform this end-to-end without robotic hardware that doesn't exist in typical labs.
Adoption barriers	claude-haiku-4-5-20251001	3/5	Safety regulations and liability concerns around chemical exposure and equipment damage create meaningful friction; institutions may prefer human oversight for quality assurance and chemical-handling accountability. However, no legal requirement mandates human cleaning in all cases.
Adoption barriers	claude-sonnet-5	2/5	No licensing requirement exists for cleaning per se, though chemical safety and hazardous waste handling protocols create some procedural friction; it's mostly a low-status task without strict barriers to substitution.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	A custom or semi-autonomous cleaning robot system would require significant capital investment, integration, and ongoing maintenance—easily exceeding the annual wage of a laboratory technician tasked with cleaning. Setup and oversight costs are high relative to human labor savings.
Cost vs. human wage	claude-sonnet-5	1/5	There is no viable AI-driven automation solution for this physical task, so AI cost is effectively infinite relative to a human janitor or lab assistant performing the work.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed products reliably perform full laboratory cleaning autonomously in production today. Specialized lab-cleaning robots remain research prototypes or highly constrained proof-of-concepts; general-purpose robotic systems lack the chemical safety knowledge and fine-motor control needed.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product performs general laboratory cleaning; this remains a manual custodial/lab-safety task performed by humans, with robotics limited to narrow research demos.

Collaborate with colleagues to address teaching and research issues.

CI 5–30 · exposure 13 · 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	1/5	Postsecondary education remains a slow-adopter sector with strong professional norms against outsourcing collaborative and deliberative functions. Faculty retention and autonomy over research direction are culturally entrenched, limiting AI tool deployment in this domain.
Sector adoption velocity	claude-sonnet-5	2/5	Higher education is a relatively slow-adopting sector for AI in core faculty governance and collaborative functions, with pilots for administrative tasks but little for interpersonal collaboration.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can assist by synthesizing literature, drafting meeting agendas, or organizing discussion points, providing useful structure to colleague conversations. However, the augmentation is limited to preparatory and organizational work rather than transforming the core collaborative judgment itself.
Augmentation potential	claude-sonnet-5	4/5	AI tools can meaningfully assist by summarizing research literature, drafting meeting agendas, analyzing curriculum data, or facilitating communication, enhancing the productivity of collaborative discussions.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can draft collaboration agendas or summarize discussion points, the core task requires interpersonal negotiation, nuanced judgment about research priorities, and consensus-building among domain experts—activities that resist full automation. AI might support 20-30% of the work, but significant human deliberation and decision-making remain essential.
Task automatability	claude-sonnet-5	1/5	Collaboration on teaching and research issues is a fundamentally interpersonal, judgment-based activity involving relationship building, negotiation, and shared decision-making that current AI cannot perform end-to-end.assistant
Adoption barriers	claude-haiku-4-5-20251001	4/5	Institutional and professional norms strongly favor human-to-human collaboration on teaching and research matters; academic governance structures, departmental decision-making processes, and collective faculty expertise requirements create significant organizational and cultural friction against substitution with automated systems.
Adoption barriers	claude-sonnet-5	3/5	No formal licensing requirement blocks AI involvement, but institutional norms, tenure/promotion processes, and collegial trust create moderate organizational friction against replacing human collaboration.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	The labor cost of faculty time spent in collaboration is high, and current AI assistance tools (if used) only marginally reduce that cost while still requiring human oversight and decision-making, making the all-in cost ratio unfavorable compared to the human wage.
Cost vs. human wage	claude-sonnet-5	1/5	There is no viable AI substitute performing this task independently, so cost comparison favors the human doing the actual collaborative work.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	No production AI system reliably performs collaborative problem-solving with colleagues at the level required for genuine research and teaching decisions. Some generative AI can assist with meeting notes or idea synthesis, but no deployed product independently conducts colleague collaboration.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product substitutes for human faculty collaboration on curriculum design or research strategy; this remains a human-to-human interaction.

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

CI 5–28 · exposure 13 · augmentation 63 · 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	Research in chemistry is performed by specialized, relatively small cohorts (academic labs, R&D teams) with high human involvement and slow organizational change. These sectors have not adopted AI agents for autonomous research generation.
Sector adoption velocity	claude-sonnet-5	3/5	Academia and research increasingly use AI tools for literature review, writing assistance, and data analysis, though wet-lab experimentation and grant-funded human-led research remain the norm with slow structural change.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI assists researchers with literature searching, data analysis, result visualization, and manuscript drafting, materially improving productivity in those workflow segments. However, it does not augment the core creative and empirical research tasks themselves.
Augmentation potential	claude-sonnet-5	4/5	AI substantially helps researchers with literature synthesis, drafting manuscripts, statistical analysis, and generating hypotheses, meaningfully increasing research productivity while humans retain control of experimental design and interpretation.
Task automatability	claude-haiku-4-5-20251001	1/5	Conducting original research requires hypothesis formation, experimental design, iterative troubleshooting, and novel insight generation—activities that demand human creativity and domain expertise. Current AI cannot independently conceive meaningful research questions or execute novel experimental protocols without human direction.
Task automatability	claude-sonnet-5	2/5	Original chemistry research requires designing experiments, physical labwork, novel hypothesis generation, and critical judgment that current AI cannot perform end-to-end; AI can assist with literature review, drafting, and data analysis but not conduct the full research program.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Peer review, authorship attribution, institutional accountability, and research ethics requirements create strong structural barriers. Publishing and funding institutions legally and professionally require human researchers to take responsibility for findings and integrity.
Adoption barriers	claude-sonnet-5	4/5	Academic publishing requires named human authorship, accountability for research integrity, and institutional/ethical review processes that create strong barriers to full AI substitution, though not a licensing requirement per se.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	Research-conducting requires expensive lab infrastructure, equipment, human oversight, and domain expertise; AI inference costs are negligible relative to the total cost structure. The human researcher's loaded wage is far smaller than the full operational cost of research.
Cost vs. human wage	claude-sonnet-5	2/5	Lab equipment, wet-chemistry experiments, and peer review still require human expertise and physical infrastructure, so AI only offsets a fraction of costs (e.g., writing, data crunching) rather than replacing the overall cost structure.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed product independently conducts original chemistry research from conception through publication-ready findings. AI tools assist with literature review and data analysis, but cannot replace the core research process that produces novel, publishable chemistry findings.
Technical feasibility today	claude-sonnet-5	2/5	AI writing and literature-analysis tools are deployed, but no product independently designs and executes original chemistry experiments and publishes peer-reviewed findings; human PIs remain essential throughout.

Establish, teach, and monitor students' compliance with safety rules for handling chemicals, equipment, and other hazardous materials.

CI 9–23 · exposure 20 · augmentation 50 · importance 4.7/5 · 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 and lab safety are among the slowest sectors to adopt AI for core instructional and supervisory tasks; safety liability concerns, regulatory inertia, and the physical immediacy required make displacement minimal in practice.
Sector adoption velocity	claude-sonnet-5	1/5	Physical lab safety supervision in postsecondary education is a low-digitization, high-liability domain with essentially no AI adoption trend for this specific function.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI could usefully assist by generating safety checklists, summarizing compliance logs, flagging anomalous patterns in student behavior data, or providing adaptive safety content; however, the human instructor's real-time judgment remains central.
Augmentation potential	claude-sonnet-5	3/5	AI can help draft safety rules, quizzes, training materials, and compliance checklists, usefully supporting the instructional design portion of this task even though it can't handle live monitoring.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI could draft safety protocols and monitor written work, the core task requires real-time observation of student behavior in labs, judgment about when to intervene, and adaptive instruction based on individual student comprehension—all requiring human presence. Limited automation is achievable for documentation and compliance tracking only.
Task automatability	claude-sonnet-5	2/5	AI can generate safety protocols and educational content, but establishing and physically monitoring student compliance in a live lab setting requires embodied presence and real-time judgment that current AI cannot perform.6
Adoption barriers	claude-haiku-4-5-20251001	4/5	Strong legal and institutional barriers exist: institutions are liable for student safety, regulatory frameworks (OSHA, lab safety standards) implicitly require qualified human oversight, and accreditation bodies expect direct human instruction in hazardous environments. Human sign-off on safety compliance is practically mandatory.
Adoption barriers	claude-sonnet-5	5/5	Lab safety compliance involves legal liability, institutional accreditation requirements, and physical risk to students, requiring a qualified, present, responsible instructor by regulation and institutional policy.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	A comprehensive AI system for real-time lab monitoring, safety rule establishment, and student compliance tracking would require substantial infrastructure (cameras, sensors, software integration) that would exceed the cost of a postsecondary instructor's loaded wage for this specific task.
Cost vs. human wage	claude-sonnet-5	1/5	There is no viable AI substitute for in-person safety monitoring, so cost comparison favors the human entirely; any AI attempt would need to be paired with human oversight anyway.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	No deployed product reliably monitors lab safety compliance in real time or establishes safety rules autonomously. Vision-based monitoring systems exist in research but not in production for educational settings; human instructors remain essential in practice.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product monitors students' physical handling of chemicals and hazardous equipment in real labs; this remains firmly in the domain of human supervision.

Initiate, facilitate, and moderate classroom discussions.

CI 5–25 · exposure 13 · augmentation 38 · 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	Postsecondary education has historically lagged in AI adoption, and classroom discussion facilitation—a core pedagogical activity—has seen minimal production AI deployment due to its human-intensive, relationship-dependent nature.
Sector adoption velocity	claude-sonnet-5	2/5	Higher education adoption of AI for live classroom facilitation is minimal; most AI use in academia is confined to content creation, grading assistance, or online discussion boards rather than real-time classroom moderation.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI can offer limited assistance by generating discussion prompts beforehand or helping prepare summaries after class, but does not significantly augment a faculty member's ability to actively moderate and facilitate an ongoing live discussion.
Augmentation potential	claude-sonnet-5	3/5	AI can help instructors prepare discussion questions, summarize readings, or analyze online discussion threads, providing moderate support even though it cannot run the live discussion itself.
Task automatability	claude-haiku-4-5-20251001	1/5	Classroom discussion facilitation requires real-time social dynamics, emotional intelligence, and the ability to respond to unexpected student contributions and manage interpersonal group dynamics—capabilities that current AI systems cannot reliably perform end-to-end in a live classroom setting.
Task automatability	claude-sonnet-5	2/5	Live classroom discussion requires real-time social presence, reading the room, and dynamic responsiveness that current AI cannot replicate for in-person postsecondary teaching, though some scripted prompts could be AI-generated in advance.dw
Adoption barriers	claude-haiku-4-5-20251001	4/5	Strong pedagogical, institutional, and accreditation expectations that postsecondary instruction involve direct faculty-student interaction; student and family expectations of human instruction; and potential liability concerns create substantial friction against full automation.
Adoption barriers	claude-sonnet-5	4/5	Accreditation and institutional norms require a qualified instructor to lead and be accountable for classroom instruction, and human presence is expected by students and administrators, creating strong organizational and quasi-regulatory barriers.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	The infrastructure, integration, and continuous human oversight required to deploy AI as a classroom discussion facilitator would exceed the cost of a postsecondary instructor's hourly labor for this activity.
Cost vs. human wage	claude-sonnet-5	2/5	Even where AI-assisted discussion prompts or chatbot moderation exist for online forums, the human instructor's presence remains required for in-person classes, so cost savings are minimal relative to the wage.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	While AI can generate discussion prompts or summarize conversations offline, no deployed product reliably moderates live classroom discussions in real time with the pedagogical judgment and adaptive responsiveness that this task demands.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product autonomously runs or moderates live in-person classroom discussions in higher education; AI tools exist only for asynchronous online discussion boards with limited moderation capability.

Maintain regularly scheduled office hours to advise and assist students.

CI 9–16 · exposure 5 · 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	Higher education has shown minimal adoption of AI for displacing office hours, despite decades of chatbot development. The sector prioritizes human mentoring and faculty-student relationships, and institutional inertia is substantial; isolated pilots exist but near-zero production displacement.
Sector adoption velocity	claude-sonnet-5	2/5	Higher education adopts AI tools slowly for personal advising functions; while some AI tutoring pilots exist, replacing office hours is not a common trend yet.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can assist professors by preparing summaries of common student questions, drafting responses to routine homework inquiries, or suggesting teaching resources, modestly improving office hour efficiency. However, the core task—live advising and mentoring—remains human-centric, so augmentation is partial rather than transformative.
Augmentation potential	claude-sonnet-5	3/5	AI chatbots and tutoring tools can supplement office hours by answering routine questions or providing supplementary explanations, freeing the instructor for more complex student interactions.
Task automatability	claude-haiku-4-5-20251001	1/5	This task requires real-time, contextual interaction with individual students addressing their specific academic concerns, personal circumstances, and emotional needs. Current AI systems cannot authentically replicate the mentoring, judgment, and interpersonal relationship-building that office hours fundamentally entail.
Task automatability	claude-sonnet-5	1/5	This task requires physical or scheduled real-time presence and personal relationship-building with students, which AI cannot substitute for as an end-to-end replacement of the role.rating
Adoption barriers	claude-haiku-4-5-20251001	4/5	Institutional norms and student expectations strongly favor human faculty contact for advising; many universities explicitly mandate faculty office hours as part of the employment contract. Educational and duty-of-care principles create organizational and regulatory friction against full substitution by automated systems.
Adoption barriers	claude-sonnet-5	4/5	Institutional norms, accreditation expectations, and student advising relationships create strong organizational and relational barriers to full substitution, though not strict licensure requirements.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Automating office hours would require sophisticated conversational AI, persistent student data management, and integration with scheduling systems. The infrastructure and oversight costs approach or exceed the wage cost of a teaching assistant or professor hour, especially when accounting for error handling and liability.
Cost vs. human wage	claude-sonnet-5	2/5	While AI chat assistance is cheap, the task as defined (scheduled human office hours) still requires the professor's paid time, so cost savings are minimal unless the task itself is redefined.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	While chatbots can answer routine chemistry questions or provide generic study advice, no deployed system reliably performs the full scope of office hours—scheduling coordination, individualized problem-solving, research mentoring, and adaptive support based on student state. Narrow capabilities exist but fall far short of production-grade office hour replacement.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product performs the function of a professor holding office hours to advise students; chatbots can supplement but do not replace this human institutional role.

Provide professional consulting services to government or industry.

CI 9–11 · exposure 0 · augmentation 63 · 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	Consulting services are relationship-driven and high-stakes; adoption of AI as primary consultant is minimal. Firms use AI tools to support consultants' research, but human consultants remain the irreplaceable front-end and decision authority, reflecting slow substitution in this sector.
Sector adoption velocity	claude-sonnet-5	2/5	Academic and consulting sectors are experimenting with AI tools for research support, but formal consulting engagements still rely heavily on human experts with slow structural change.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can assist by rapidly synthesizing literature, suggesting experimental protocols, analyzing datasets, and generating preliminary technical reports—raising consultant productivity. However, the human expert must interpret findings, engage with clients, and validate conclusions, so augmentation is meaningful but bounded.
Augmentation potential	claude-sonnet-5	4/5	AI can meaningfully assist with literature review, data analysis, and report drafting, enhancing the consultant's productivity while the human retains final judgment and responsibility.
Task automatability	claude-haiku-4-5-20251001	1/5	Consulting services require deep domain expertise, client relationship management, negotiation, and judgment about complex technical and business tradeoffs—none of which current AI can perform end-to-end at the professional level required. While AI can assist with literature review or data analysis, the consulting engagement itself demands human credibility and accountability.
Task automatability	claude-sonnet-5	1/5	Consulting requires original expert judgment, contextual problem-solving, and accountability that current AI cannot deliver end-to-end; at best AI supports research and drafting within the process.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Consulting relationships demand professional licensure, legal liability for recommendations, and client trust in a named human expert. Government and industry contracts typically require certified professionals with professional indemnity insurance, creating substantial legal and regulatory barriers to AI substitution.
Adoption barriers	claude-sonnet-5	4/5	Government/industry consulting often requires credentialed expertise, professional liability, and named accountability, creating strong barriers to full AI substitution.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Even though AI inference is cheap, the oversight and integration burden for a consulting engagement is high. A human consultant's loaded wage is amortized across many billable hours; AI cannot yet replace that productivity on complex advisory work, making the all-in cost comparable or unfavorable.
Cost vs. human wage	claude-sonnet-5	2/5	AI tools can cut research/drafting time cheaply, but the actual deliverable—credentialed expert advice—still requires the human consultant, so overall cost savings are limited.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed product can reliably replace a professional chemistry consultant for government or industry. AI systems lack the credentials, legal standing, and contextual judgment to serve as primary consultants; they cannot independently scope problems or be held liable for advice.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product independently provides expert scientific consulting services to government or industry; this remains firmly in the human-expert domain.

Supervise students' laboratory work.

CI 0–9 · exposure 8 · augmentation 38 · 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	1/5	Educational institutions remain conservative adopters of automation in high-liability activities like lab supervision. No measurable adoption of AI-based lab supervision replacements exists in postsecondary settings.
Sector adoption velocity	claude-sonnet-5	1/5	Physical, hands-on lab supervision in academic settings sees minimal AI adoption due to safety, liability, and the fundamentally in-person nature of the task.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI could assist by flagging safety violations in recorded footage, providing pre-lab procedure reminders to students, or generating lab reports from observations, but these are supplementary rather than transformative to the core supervisory relationship.
Augmentation potential	claude-sonnet-5	2/5	AI can help with pre-lab instructions, safety checklists, or answering procedural questions, but offers little assistance during the actual real-time physical supervision.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI could assist in monitoring safety compliance or reviewing lab procedures through video, supervising real-time laboratory work requires immediate intervention for safety hazards and one-on-one guidance—tasks that demand human presence and judgment. Significant safety-critical components cannot be automated effectively.
Task automatability	claude-sonnet-5	1/5	Supervising live laboratory work requires physical presence to monitor chemical safety, hazardous materials handling, and real-time student technique correction, which current AI cannot perform end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	5/5	Educational institutions have legal liability for student safety during lab work, and many jurisdictions require credentialed instructors or supervisory personnel to oversee experiments. Institutional accreditation and insurance policies typically mandate human supervision.
Adoption barriers	claude-sonnet-5	5/5	Safety regulations, liability for chemical accidents, and institutional requirements mandate a qualified human instructor be physically present during lab sessions.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	The cost of an AI system capable of reliable lab supervision (cameras, sensors, real-time analysis, integration) plus human oversight would exceed the loaded wage of a teaching assistant or instructor, making it economically unviable today.
Cost vs. human wage	claude-sonnet-5	1/5	There is no viable AI substitute for physical lab supervision, so cost comparison favors the human by default since the AI alternative doesn't functionally exist.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed product reliably supervises live laboratory work end-to-end. AI monitoring systems exist for limited safety detection, but comprehensive supervision—ensuring proper technique, safety compliance, and student understanding simultaneously—remains research-stage.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product provides autonomous physical supervision of student lab work; safety-critical in-person oversight remains entirely human-performed.

Serve on committees or in professional societies.

CI 0–7 · 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	This governance and professional service task is foundational to academic and professional institutions and shows no displacement by AI; the task remains exclusively human-performed.
Sector adoption velocity	claude-sonnet-5	2/5	Higher education is a moderately slow-adopting sector for AI in governance and interpersonal representation roles, with little evidence of AI displacing committee participation.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI might assist in preparing briefing materials or analyzing committee data, but the core governance and relational work remains strictly human; augmentation is marginal to the task's essential nature.
Augmentation potential	claude-sonnet-5	3/5	AI can help draft meeting agendas, summarize minutes, prepare reports, or research policy issues, providing moderate productivity support for the human committee member.
Task automatability	claude-haiku-4-5-20251001	1/5	Committee and professional society participation requires sustained interpersonal judgment, relationship-building, consensus navigation, and domain-contextual decision-making that current AI cannot perform autonomously. No deployed AI system can meaningfully represent a human in governance, voting, or collegial deliberation.
Task automatability	claude-sonnet-5	1/5	Serving on committees requires human judgment, relationship-building, deliberation, and representation of institutional/professional interests that AI cannot perform end-to-end.'
Adoption barriers	claude-haiku-4-5-20251001	5/5	Serving on committees and professional societies requires human appointment, voting rights, fiduciary responsibility, and professional accountability—hard legal and organizational barriers prevent AI substitution.
Adoption barriers	claude-sonnet-5	4/5	Institutional governance structures, faculty bylaws, and professional society rules typically require human membership, voting rights, and accountability, creating strong structural barriers to automation.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	The task is inherently human-performed and valued for the individual's expertise and judgment; there is no meaningful AI cost comparison, as substitution is not feasible.
Cost vs. human wage	claude-sonnet-5	1/5	There is no AI system that can replace a human's presence and voice in committee work, so no meaningful cost comparison for full substitution exists.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No commercial product performs this task end-to-end; it fundamentally requires a human agent with authority, accountability, and professional standing to serve in a representative capacity.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product substitutes for a human's active membership and participation in committees or professional societies; this remains inherently a human social/institutional role.

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

CI 0–5 · exposure 0 · augmentation 38 · importance 3.5/5 · 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	There is no adoption of AI to perform committee service because the task is fundamentally non-automatable at the governance level. Academic institutions continue to require human faculty committees by accreditation and institutional structure.
Sector adoption velocity	claude-sonnet-5	1/5	Committee governance in higher education is a low-digitization, human-relational process with essentially no AI displacement occurring or anticipated soon.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI could assist with minor administrative tasks like preparing agenda documents or summarizing background materials, but the core deliberation, voting, and policy judgment must remain with human committee members. Assistance is marginal and not transformative.
Augmentation potential	claude-sonnet-5	3/5	AI can help draft agendas, summarize meeting notes, analyze policy documents, or prepare reports to support committee members, offering moderate productivity assistance.
Task automatability	claude-haiku-4-5-20251001	1/5	Committee work requires human judgment on institutional policies, stakeholder consensus-building, and nuanced deliberation on academic matters—core functions that demand accountability and authority that cannot be delegated to AI systems. No meaningful part of this task can be automated end-to-end today.
Task automatability	claude-sonnet-5	1/5	Committee service requires human presence, deliberation, negotiation, and institutional judgment that AI cannot substitute for; no end-to-end automation is plausible.
Adoption barriers	claude-haiku-4-5-20251001	5/5	Hard regulatory and institutional barriers exist: committee members must be credentialed humans with legal standing to vote on departmental and institutional policies. Academic governance structures legally require human faculty participation and decision-making authority.
Adoption barriers	claude-sonnet-5	4/5	Faculty governance roles typically require institutional membership, tenure/rank status, and shared-governance norms that legally and organizationally require a human faculty member to serve.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	This is a governance and accountability function that cannot be performed by AI; the task is not substitutable, so cost comparison is not applicable. The task has no meaningful AI alternative today.
Cost vs. human wage	claude-sonnet-5	1/5	There is no AI-driven substitute output to compare cost against; the task requires a human role holder, so AI cost is not applicable/comparable.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed AI product reliably performs institutional committee work, which requires legal standing, fiduciary responsibility, voting authority, and the ability to represent colleagues in governance contexts. This remains entirely a human function.
Technical feasibility today	claude-sonnet-5	1/5	No deployed AI product actually sits on or substitutes for a human on academic/administrative committees; this remains outside current product scope.

Participate in campus and community events.

CI 0–5 · exposure 0 · augmentation 13 · 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	There is no measurable adoption of AI for participating in campus events because the task is fundamentally incompatible with AI capabilities and institutional expectations for faculty presence.
Sector adoption velocity	claude-sonnet-5	1/5	Higher education is a slow-adopting sector for physical/social presence tasks, and this specific activity has no adoption trajectory since it's not automatable.
Augmentation potential	claude-haiku-4-5-20251001	1/5	AI offers minimal assistance for human participation in events; at most, it might help with scheduling or logistics, but does not meaningfully augment the core act of participation itself.
Augmentation potential	claude-sonnet-5	2/5	AI can help with scheduling, event planning materials, or communications around events, but offers minimal assistance to the actual participation itself.
Task automatability	claude-haiku-4-5-20251001	1/5	Participation in campus and community events is inherently a human-presence task requiring social interaction, relationship-building, and real-time responsiveness that cannot be meaningfully automated today. No AI system can substitute for an instructor's physical and social engagement at events.
Task automatability	claude-sonnet-5	1/5	This task requires physical presence, social interaction, and institutional representation that AI cannot perform end-to-end today.
Adoption barriers	claude-haiku-4-5-20251001	5/5	Institutional and social expectations, combined with the requirement for authentic human presence and relationship-building, create near-total barriers to automation. Campus and community engagement depends on the instructor's actual participation and professional standing.
Adoption barriers	claude-sonnet-5	4/5	Community and campus engagement inherently requires human presence, relationship-building, and representation, creating strong structural barriers to automation.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	AI cannot perform this task at all, so direct cost comparison is not applicable; the cost of attempting automation (if possible) would far exceed the value of human participation.
Cost vs. human wage	claude-sonnet-5	1/5	There is no AI substitute performing this task, so cost comparison favors the human by default since no viable AI alternative exists.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed AI product can autonomously participate in campus or community events as a substitute for human presence. This task fundamentally requires embodied human participation and cannot be performed by current AI systems.
Technical feasibility today	claude-sonnet-5	1/5	No deployed AI product attends or participates in campus/community events on behalf of a person; this is fundamentally a human presence task.

Act as advisers to student organizations.

CI 0–5 · exposure 0 · augmentation 25 · 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	1/5	Academic institutions have shown minimal adoption of AI for student advising roles; the relationship-dependent and professionally accountable nature of the role resists displacement even in digitization-forward sectors.
Sector adoption velocity	claude-sonnet-5	1/5	Higher education advising roles show minimal AI displacement; this is a low-digitization, relationship-driven duty with no adoption momentum.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI could marginally assist with scheduling, document tracking, or information retrieval for student organizations, but it offers limited augmentation for the core advisory function of providing mentorship and guidance.
Augmentation potential	claude-sonnet-5	2/5	AI can help with scheduling, drafting communications, or organizing event logistics for the advisor, but offers limited assistance to the core mentoring and oversight function.
Task automatability	claude-haiku-4-5-20251001	1/5	Advising student organizations requires nuanced understanding of student needs, interpersonal relationships, mentorship, and institutional context—elements that current AI cannot reliably handle end-to-end. The task involves ongoing human judgment about individual student development and organizational decision-making.
Task automatability	claude-sonnet-5	1/5	Advising student organizations requires ongoing relationship-building, mentorship, judgment calls, and institutional representation that current AI cannot perform end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	5/5	Institutional, legal, and duty-of-care barriers are substantial: students expect human advisers, faculty have contractual advisory responsibilities, and liability concerns around delegating mentorship to machines create hard adoption barriers.
Adoption barriers	claude-sonnet-5	4/5	Institutions typically require a designated faculty member for liability, signature authority, and official sponsorship of student organizations, creating strong organizational and administrative barriers.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	The cost of deploying AI systems to simulate human mentorship and organizational advisory, plus required oversight and correction, would exceed the instructional labor cost of a faculty adviser.
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 AI product reliably performs the advisory role for student organizations; this requires sustained relationships, emotional intelligence, and accountability that current systems cannot provide in practice.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product acts as a student organization advisor; this remains a human relational and institutional role with no commercial analog.

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

CI 0–3 · exposure 0 · augmentation 38 · 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	1/5	Academic institutions move slowly on automation of core supervisory duties, and this task sits at the heart of faculty responsibilities; sectors relying on this task (higher education) show minimal AI displacement of supervision functions to date.
Sector adoption velocity	claude-sonnet-5	2/5	Higher education adopts AI tools for grading or content support, but actual supervisory oversight roles remain untouched and slow to change due to accreditation and mentorship norms.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI tools could assist with routine administrative logging of student progress or summarizing research outcomes, but the core supervisory work—real-time guidance, feedback, mentoring, and judgment—remains fundamentally human-centric with limited scope for meaningful AI assistance.
Augmentation potential	claude-sonnet-5	3/5	AI can help supervisors track student progress, provide feedback drafts, or analyze research data, offering moderate assistance while the human remains the actual supervisor.
Task automatability	claude-haiku-4-5-20251001	1/5	Supervision of research and teaching work requires real-time observation, mentoring, interpersonal judgment, and adaptive feedback based on individual student progress—tasks that demand human presence, contextual understanding, and relationship-building that current AI systems cannot perform end-to-end.
Task automatability	claude-sonnet-5	1/5	Supervising students' teaching, internships, and research requires mentorship, judgment about individual progress, and real-time interpersonal guidance that current AI cannot perform end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	5/5	Strong legal and regulatory barriers protect this task: faculty must hold credentials, institutional accreditation requires qualified human supervision of research and teaching, and academic integrity standards mandate human judgment in evaluating student work and providing mentorship.
Adoption barriers	claude-sonnet-5	5/5	Institutional accreditation, degree-granting requirements, and faculty-of-record rules mandate that a qualified human supervise student research and teaching.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	Automation would require autonomous supervision AI with full liability; the cost of developing, integrating, and legally guaranteeing such a system would far exceed the loaded wage of a faculty member providing this supervision.
Cost vs. human wage	claude-sonnet-5	1/5	There is no AI system performing this supervisory role, so no meaningful cost comparison to a human supervisor exists.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed product reliably supervises academic research or teaching work in production at scale; AI systems lack the authority, contextual judgment, and accountability required to sign off on student progress or manage the social-mentoring dimensions of academic supervision.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product substitutes for faculty supervision of student research or teaching; AI tools exist only as ancillary aids, not as supervisors of record.

Perform administrative duties, such as serving as a department head.

CI 0–3 · 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	Academic institutions have shown virtually no movement toward AI-driven department leadership; governance, accreditation, and faculty culture strongly reinforce human leadership in administrative roles.
Sector adoption velocity	claude-sonnet-5	2/5	Higher education administration is a sector with slow AI adoption for leadership functions, though administrative support tools (scheduling, document drafting) are increasingly used.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI can assist narrowly with scheduling, document drafting, or data aggregation, but the core leadership and decision-making functions of a department head offer limited augmentation value beyond basic office tools.
Augmentation potential	claude-sonnet-5	3/5	AI can assist with drafting reports, scheduling, budget spreadsheets, and correspondence, providing moderate productivity gains for the administrative subtasks within the role.
Task automatability	claude-haiku-4-5-20251001	1/5	Administrative duties as a department head require complex human judgment, stakeholder management, personnel decisions, and institutional knowledge that current AI cannot handle end-to-end. AI can assist with scheduling or documentation but cannot autonomously lead a department.
Task automatability	claude-sonnet-5	1/5	Serving as a department head involves leadership, personnel decisions, budget oversight, conflict resolution, and institutional politics that require human judgment, authority, and accountability that current AI cannot perform end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	5/5	Department head roles are legally and contractually bound to specific individuals with institutional authority and liability; governance structures explicitly require a human administrator to hold this position and sign off on decisions.
Adoption barriers	claude-sonnet-5	5/5	Department head roles require institutional authorization, formal appointment, accountability structures, and often tenure/faculty governance rules that legally and organizationally restrict this to a qualified human.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	The cost of any AI system plus required human oversight and final decision-making far exceeds the marginal cost of the department head performing these duties themselves.
Cost vs. human wage	claude-sonnet-5	1/5	There is no AI substitute capable of performing this leadership role, so cost comparison favors the human entirely; any AI use is merely supportive, not substitutive.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No AI product reliably performs department head duties in production; this requires legal authority, fiduciary responsibility, and interpersonal leadership that remains entirely human-dependent in all deployed systems.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product performs the role of a department head; AI tools may assist with scheduling or drafting reports but do not execute the administrative leadership function itself.

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.