Epidemiologists
19-1041.00Median wage $87,220/yr12,090 employed (US)Rank #445 of 923 scored · top 48% by substitution
Investigate and describe the determinants and distribution of disease, disability, or health outcomes. May develop the means for prevention and control.
Sub-scores
0–100 · band = confidence interval from rater disagreement
Substitution29
Exposure27
Augmentation71
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
16 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
0%
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%27
panel mean rating 2.1/5 → substitution pressure 27/100
Technical feasibility todayw 20%27
panel mean rating 2.1/5 → substitution pressure 27/100
Cost vs. human wagew 15%32
panel mean rating 2.3/5 → substitution pressure 32/100
Adoption barriersw 20%inverted — strong barriers lower the score30
panel mean rating 3.8/5 (barrier strength) → substitution pressure 30/100
Sector adoption velocityw 10%30
panel mean rating 2.2/5 → substitution pressure 30/100
Task breakdown (16 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.
Identify and analyze public health issues related to foodborne parasitic diseases and their impact on public policies, scientific studies, or surveys.
51CI 25–76 · exposure 58 · augmentation 88 · importance 3.9/5 · click for rater detail
Identify and analyze public health issues related to foodborne parasitic diseases and their impact on public policies, scientific studies, or surveys.
51| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Public health agencies and research institutions show middling adoption: pilots with AI-assisted literature mining and surveillance dashboards are common, but full automation of epidemiological analysis remains uncommon; digitization is moderate and institutional inertia slows deployment relative to commercial sectors. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Public health agencies are historically slow adopters of AI tools, often constrained by data privacy, legacy systems, and government procurement processes, though pilots in surveillance analytics are emerging. |
| Augmentation potential | claude-haiku-4-5-20251001 | 5/5 | AI substantially augments epidemiologist productivity by automating literature screening, data aggregation, trend detection, and draft report generation, allowing epidemiologists to focus on causal interpretation, policy implications, and validation—transforming efficiency while keeping expert judgment central to the task. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can meaningfully assist by rapidly synthesizing scientific literature, flagging patterns in surveillance data, and drafting reports, significantly boosting productivity while the epidemiologist retains interpretive and decision-making control. |
| Task automatability | claude-haiku-4-5-20251001 | 5/5 | Current AI systems can systematically identify foodborne parasitic disease patterns from public health databases, scientific literature, and survey data using NLP and data mining; analyze epidemiological trends, extract statistical relationships, and synthesize findings into policy-relevant reports—potentially reducing manual literature review and data synthesis time by 50–70% while maintaining quality. |
| Task automatability | claude-sonnet-5 | 2/5 | AI can assist with literature review, data analysis, and drafting reports, but identifying novel public health issues and interpreting policy implications requires domain judgment and integration of field data that current AI cannot fully replicate end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Some regulatory and organizational friction exists: epidemiologists must validate AI findings before policy recommendations, institutional review may slow adoption, and public health agencies often require human accountability for disease surveillance outputs; however, no legal licensing requirement forbids AI analysis of existing data. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Public health surveillance and policy recommendations typically require credentialed epidemiologists and institutional review, with significant liability and regulatory oversight tied to public health decisions. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | AI-driven literature analysis, database querying, and report drafting cost a fraction of epidemiologist labor for equivalent output; the loaded cost of a public health epidemiologist ($120k–150k annually) far exceeds cloud-based AI services ($1–10k per analysis after setup), making AI 10–50× cheaper per task-equivalent. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI tools can reduce time on literature synthesis and data processing, but the need for expert epidemiological validation and field-specific knowledge keeps human oversight costs high relative to savings. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Deployed AI tools (literature mining, data analytics platforms, LLMs for report generation) perform components of this task reliably in public health settings, though full end-to-end automation requires human epidemiologist validation of disease causality and policy implications; production use exists but typically within human-guided workflows rather than fully autonomous operation. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Deployed products (statistical software, LLM-based literature summarizers) support pieces of this work, but no production system autonomously identifies and analyzes emerging foodborne parasitic disease issues with policy relevance. |
Educate healthcare workers, patients, and the public about infectious and communicable diseases, including disease transmission and prevention.
47CI 41–54 · exposure 42 · augmentation 75 · importance 4.2/5 · click for rater detail
Educate healthcare workers, patients, and the public about infectious and communicable diseases, including disease transmission and prevention.
47| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Public health agencies and healthcare systems are experimenting with AI for patient education and awareness campaigns, but production adoption remains patchy. Most organizations still rely on human-written or heavily human-reviewed materials, and adoption lags compared to faster-moving sectors like finance or software. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Public health and healthcare sectors are adopting AI content tools at a moderate pace, with pilots for chatbots and educational content common but full-scale deployment for authoritative communication still limited. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI excels at drafting, summarizing scientific literature, generating multilingual versions, and creating visual aids—all of which can dramatically accelerate an epidemiologist's ability to develop tailored educational programs. The human epidemiologist remains essential for accuracy review and contextual judgment, but AI substantially multiplies their productivity. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI substantially helps epidemiologists draft educational materials, translate content into plain language, and produce multilingual public health messaging, greatly increasing efficiency while the professional retains oversight. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI can draft educational materials, generate disease transmission diagrams, and create prevention infographics with significant time savings. However, the task requires tailoring content to diverse audiences (healthcare workers, patients, general public) with varying literacy levels and cultural contexts, which still demands substantial human judgment and refinement to ensure accuracy and appropriateness. |
| Task automatability | claude-sonnet-5 | 2/5 | AI can generate educational content and answer questions, but delivering tailored education involves audience assessment, credibility, and interactive dialogue that current systems only partially replace. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Health communication is not strictly licensed to epidemiologists, and many organizations (nonprofits, public health agencies) can delegate education to AI-assisted teams. However, liability concerns around health misinformation, regulatory expectations for accuracy (FDA/CDC guidance), and institutional preference for human expert sign-off on disease information create meaningful friction against full automation. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing requirement to distribute health information itself, but liability concerns around misinformation and the need for scientific accuracy from a credentialed source create moderate institutional friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | AI inference for generating educational content is very cheap per unit, and integration costs for templating and distribution are modest. The loaded cost of a professional epidemiologist or health educator creating equivalent content from scratch is substantially higher, making AI cost per output at least 5–10× lower. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI can cheaply draft educational materials, but designing accurate, context-sensitive public health messaging with review still requires expert oversight, keeping costs moderate rather than dramatically lower. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | AI tools (LLMs, content generators) reliably produce educational text and visual aids, and some organizations use them for patient-facing health information. However, deployed systems rarely handle the full lifecycle of audience segmentation, fact-checking against current epidemiological data, and ensuring regulatory compliance with health communication standards, so material error rates and scope limitations remain. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Chatbots and AI-generated materials (e.g., CDC-style FAQs, patient education handouts) are deployed today, but public health communication still relies heavily on human experts for nuance and trust. |
Write grant applications to fund epidemiologic research.
46CI 34–59 · exposure 38 · augmentation 75 · importance 3.8/5 · click for rater detail
Write grant applications to fund epidemiologic research.
46| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Academic and research institutions have been slow to formally adopt AI for grant writing, with most use informal and ad-hoc; conservative cultures and concern over disclosure, originality, and reviewer perception limit rapid uptake despite awareness of efficiency gains. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Academic and research institutions are adopting AI writing tools at a moderate pace, with growing but still cautious use due to concerns about originality, accuracy, and funder policies on AI-generated content. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI excels at rapid draft generation, literature synthesis, structure suggestions, and iterative editing—productivity gains are substantial when epidemiologists use AI to outline, refine, and check clarity while retaining control over scientific strategy and claims. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI substantially speeds up drafting, editing, literature summarization, and formatting for grant applications, meaningfully boosting researcher productivity while the epidemiologist retains ownership of study design and scientific judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | AI can draft sections of grant applications (background, literature review, methods outline) but cannot reliably generate the novel research hypotheses, strategic positioning, and institution-specific alignment that reviewers demand. Human oversight and substantial revision remain essential for competitive proposals. |
| Task automatability | claude-sonnet-5 | 3/5 | AI can draft significant portions of grant text (background, literature synthesis, boilerplate methods) given researcher input, but the core specific aims, novel study design, and strategic framing still require substantial human expertise and iteration. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Grants must be signed by a human researcher and institution, and reviewers' trust in the proposing scientist's voice is material; however, AI can assist in preparation without legal barriers. Organizational norms around human authorship and reputational risk create friction but not hard legal substitution barriers. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No formal licensing requirement bars AI assistance in writing grants, though funders expect PI authorship, scientific accountability, and institutional review, creating some friction against full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | AI-assisted drafting and editing costs (GPT API, copyedit tools) are orders of magnitude cheaper than the loaded salary of an epidemiologist spending weeks on grant writing, even accounting for oversight and revision time. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI drafting tools are cheap per query, but the human time needed for domain expertise, data specifics, and review to reach fundable quality remains substantial, keeping overall cost savings moderate rather than order-of-magnitude. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed product reliably writes fundable grant applications end-to-end; AI tools exist for drafting support and editing, but grant success depends on scientific novelty and institutional credibility that require human judgment. Products show high error rates in technical specificity and strategic framing. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | LLM-based writing assistants are already used by researchers to draft and edit grant sections, but no deployed product independently produces fundable, scientifically sound epidemiologic grant applications without heavy human revision. |
Write articles for publication in professional journals.
37CI 25–50 · exposure 38 · augmentation 88 · importance 4.0/5 · click for rater detail
Write articles for publication in professional journals.
37| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Academic and public health sectors are cautious adopters of AI for research output. While some epidemiologists use AI for drafting support, reliance on human-authored and peer-reviewed publication remains the norm, with slow and limited displacement of writing labor. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Academic and public health research is a professional/information-sector field with moderate AI tool adoption for writing assistance, but formal workflows integrating AI into manuscript production remain in pilot/informal stages. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI writing assistants meaningfully augment epidemiologists by accelerating drafting, editing, and clarity improvements, allowing researchers to focus on analysis and interpretation. Many epidemiologists now use such tools to boost productivity in manuscript preparation while retaining full intellectual control. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI writing tools substantially speed up drafting, editing, literature summarization, and formatting for journal submission, making them highly valuable augmentative aids for epidemiologists while they retain authorship and analytical responsibility. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | AI can draft sections and improve clarity, but epidemiological articles require original data analysis, interpretation of statistical results, and novel scientific claims that demand human judgment. Current systems cannot autonomously conduct the research, validate findings, or assume responsibility for accuracy. |
| Task automatability | claude-sonnet-5 | 3/5 | AI can draft literature reviews, structure sections, and generate prose from findings, saving significant time, but synthesizing original epidemiological findings, ensuring statistical rigor, and framing novel contributions still require substantial human input.4o |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Publication requires human authorship and accountability; journals demand that named authors take responsibility for content integrity and novelty. Regulatory and professional norms, plus liability concerns around false claims in public health literature, create substantial legal and ethical barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing requirement for authorship, but journal ethics policies, authorship accountability rules, and disclosure requirements around AI-generated content create meaningful friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | While AI writing tools have low per-use costs, epidemiologists earn substantial salaries, and the actual labor savings are modest since human review, revision, and intellectual contribution remain necessary. The savings do not offset the loaded wage significantly. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI drafting tools are cheap per use, but the need for expert review, fact verification, and correction of statistical/methodological content adds substantial oversight cost, making the net savings moderate rather than order-of-magnitude. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | AI writing assistants exist and can help with grammar and structure, but no deployed system reliably produces peer-review-ready epidemiological articles end-to-end. Journals require human authorship accountability and originality verification that AI cannot satisfy. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Products like ChatGPT and specialized writing assistants are widely used by researchers for drafting and editing manuscripts, but no deployed system reliably produces publication-ready epidemiological papers without heavy human revision and fact-checking. |
Communicate research findings on various types of diseases to health practitioners, policy makers, and the public.
27CI 25–29 · exposure 25 · augmentation 75 · importance 4.5/5 · click for rater detail
Communicate research findings on various types of diseases to health practitioners, policy makers, and the public.
27| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Public health and epidemiology are highly regulated, conservative sectors with strong norms favoring human expertise and accountability in communication. AI adoption for research-to-public pipelines remains at pilot stage; most agencies and practitioners still do this manually. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Public health and government sectors are traditionally slow adopters of AI tools for external-facing communications, with pilots more common than production deployment. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can powerfully assist epidemiologists by drafting initial summaries, generating multi-audience versions, creating data visualizations, and flagging clarity issues—substantially raising productivity while the epidemiologist retains judgment on framing, emphasis, and accuracy. This is a genuine augmentation use case already in early adoption. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI writing assistants meaningfully help epidemiologists draft summaries, translate technical findings for lay audiences, and prepare slides or reports, while the epidemiologist retains responsibility for accuracy and framing. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can draft clear written summaries and generate visualizations of research findings, the task fundamentally requires human judgment about how to frame nuanced epidemiological evidence for diverse audiences with different technical backgrounds and stakes, and how to handle uncertainty responsibly. Current systems cannot reliably adapt messaging across practitioners, policymakers, and the public in ways that meet the fidelity and ethical standards required. |
| Task automatability | claude-sonnet-5 | 2/5 | AI can draft summaries and visualizations but crafting audience-appropriate, accurate scientific communication with nuanced judgment about implications still requires substantial human expertise and accountability.summary drafting only covers a portion of the task. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Epidemiological communication to policymakers and the public carries high reputational and liability risk; health practitioners and government agencies often have regulatory or institutional requirements that a named human epidemiologist sign off on or author public health guidance. This legal/organizational friction strongly protects the role from full automation. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Public health communications often require credentialed epidemiologists or agency sign-off due to liability, regulatory oversight (e.g., CDC/WHO protocols), and the high cost of misinformation errors. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 3/5 | AI text generation and visualization tools are cheap to run, but integrating them into epidemiological communication workflows requires human oversight, fact-checking, and revision—offsetting the raw inference cost and keeping total-cost-per-output roughly comparable to a skilled communicator's time. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI can cheaply generate draft text, but the need for expert fact-checking, contextualization, and liability review for public health communications keeps overall cost comparable to or only modestly below human-only costs. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | AI systems can generate text and figures explaining study results, but no deployed product reliably handles the full communication task—selecting which findings matter for each audience, judging messaging tone, navigating conflicting evidence, and taking responsibility for public health implications. Pilots exist, but production deployment at epidemiological organizations remains minimal. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | LLM-based writing assistants are used to help draft reports and summaries, but no deployed product independently and reliably communicates epidemiological findings to diverse audiences without expert review. |
Conduct research to develop methodologies, instrumentation, and procedures for medical application, analyzing data and presenting findings.
26CI 25–28 · exposure 25 · augmentation 75 · importance 3.7/5 · click for rater detail
Conduct research to develop methodologies, instrumentation, and procedures for medical application, analyzing data and presenting findings.
26| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Adoption of AI in epidemiological research is still largely pilot and research-stage, concentrated in data-heavy analysis tasks. The methodological innovation and instrumentation development that define this task remain driven by human researchers; production deployment of AI-automated methodology development is not widespread in the sector. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Public health and academic research sectors are adopting AI tools for data analysis and literature synthesis at a moderate pace, with pilots common but full methodological automation rare. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI substantially augments epidemiologists by automating data processing, identifying patterns in large datasets, literature synthesis, and statistical modeling, allowing researchers to focus on hypothesis generation and experimental design. These tools notably raise productivity in the data-intensive parts of research while the epidemiologist retains central creative and validation roles. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI significantly aids epidemiologists in data analysis, statistical modeling, drafting reports, and literature review, substantially boosting productivity while humans retain control over methodology design and interpretation. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can assist with data analysis and literature review, the core work of developing novel methodologies and instrumentation requires domain expertise, creative problem-solving, and iterative experimental design that AI cannot perform end-to-end today. The task involves scientific innovation and validation that falls well short of the 50% time-saving threshold for full automation. |
| Task automatability | claude-sonnet-5 | 2/5 | This task combines novel methodology design, instrument development, and scientific judgment that current AI cannot execute end-to-end reliably; AI can assist with data analysis and literature synthesis but not the full research design and validation cycle.atable share is modest. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Regulatory barriers are significant: methodologies and procedures for medical application must typically be validated and approved by regulatory bodies (e.g., FDA, ethics boards, peer review), and institutional liability rests with credentialed human researchers. Publication and professional credentialing requirements create legal and organizational friction against full automation. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Research validity, peer review, IRB/ethical oversight, and publication standards require human epidemiologists to design and vouch for methodologies, creating strong professional and regulatory barriers to automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI-assisted analysis tools are inexpensive, but they cannot replace the loaded cost of a research epidemiologist (salary, benefits, institutional overhead) because human expertise is essential to the task. The cost of human scientists remains substantially lower than the combined cost of humans plus AI systems for this specialized research work. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Substantial human oversight, domain expertise, and validation are still required, so AI mainly supplements rather than replaces the costly expert labor, keeping cost savings modest relative to a fully human-driven process. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed product performs this task reliably end-to-end; research and data analysis tools exist, but developing new methodologies and instrumentation requires human epidemiologists' judgment and empirical work. AI systems lack the capability to independently design and validate novel scientific procedures at production scale. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | AI tools exist for statistical analysis and literature review support, but no deployed product autonomously designs epidemiological methodologies or validates new instrumentation in production settings. |
Consult with and advise physicians, educators, researchers, government health officials and others regarding medical applications of sciences, such as physics, biology, and chemistry.
26CI 25–28 · exposure 25 · augmentation 75 · importance 3.5/5 · click for rater detail
Consult with and advise physicians, educators, researchers, government health officials and others regarding medical applications of sciences, such as physics, biology, and chemistry.
26| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Epidemiology and public health organizations are adopting AI tools for data analysis and literature synthesis, but deep adoption of AI-autonomous consultation remains nascent; most deployment remains in supportive roles rather than autonomous decision-making. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Public health and research sectors have moderate AI adoption for literature review and data analysis, but advisory consultation roles show slower, more cautious integration. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can significantly assist epidemiologists by rapidly synthesizing relevant scientific literature, identifying patterns in epidemiological data, and generating candidate explanations or policy options that the epidemiologist then refines and takes responsibility for, substantially raising expert productivity in advisory work. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI significantly aids epidemiologists by rapidly synthesizing scientific literature, generating summaries, and supporting evidence-based recommendations, enhancing the quality and speed of advice given. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can retrieve and synthesize scientific literature and generate draft advice, the task requires nuanced expert judgment, contextual understanding of specific stakeholder needs, and the ability to weigh complex tradeoffs that current systems struggle with at production scale. The human-facing consultation element and need for accountability further limit full end-to-end automation. |
| Task automatability | claude-sonnet-5 | 2/5 | This task requires synthesizing expertise, contextual judgment, and interpersonal trust-building in real-time consultation, which current AI can support but not fully replace end-to-end., |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Professional licensing, liability exposure, government and institutional requirements for credentialed epidemiologists to sign off on health policy advice, and the necessity of human expert judgment create substantial adoption barriers. Regulatory frameworks and professional standards strongly protect this advisory role. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Advising on medical and public health matters carries significant liability and often requires credentialed expertise and professional accountability, creating strong barriers to full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | The cost of AI inference plus required human expert oversight to ensure accurate, context-appropriate advice likely approaches or exceeds the cost of direct expert consultation, especially when accounting for liability and quality assurance. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | While AI query costs are low, the oversight, verification, and liability management needed to trust AI-generated advice in this high-stakes advisory role raises effective cost closer to human parity. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed products reliably perform expert epidemiological consultation as a standalone service; AI tools exist for literature review and data analysis but require substantial human oversight to translate into actionable advice. Current systems lack the judgment and accountability framework needed for this high-stakes advisory role. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | AI tools can provide research summaries and information retrieval but no deployed product reliably serves as an authoritative consultant to physicians and officials on applied science questions in production settings. |
Investigate diseases or parasites to determine cause and risk factors, progress, life cycle, or mode of transmission.
25CI 25–25 · exposure 25 · augmentation 75 · importance 4.3/5 · click for rater detail
Investigate diseases or parasites to determine cause and risk factors, progress, life cycle, or mode of transmission.
25| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Public health agencies and research institutions are adopting AI surveillance and analytics tools slowly and cautiously; most deployment remains pilot-stage or limited to administrative support rather than displacement of investigation work itself. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Public health and government epidemiology sectors are slower adopters of AI agents compared to finance or tech, with pilots more common than production deployment. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI substantially augments epidemiologists through rapid literature synthesis, statistical modeling, data visualization, outbreak detection algorithms, and hypothesis prioritization—enabling faster and broader investigation while the epidemiologist retains analytical and decision-making authority. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI significantly aids literature synthesis, statistical modeling, outbreak pattern detection, and drafting reports, meaningfully boosting epidemiologist productivity while they retain decision-making control. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | AI can assist with literature analysis, data aggregation, and hypothesis generation, but epidemiological investigation requires integrating complex observational data, clinical judgment, field work, and causal inference that current systems cannot reliably execute end-to-end with substantial time savings at equal quality. |
| Task automatability | claude-sonnet-5 | 2/5 | Core epidemiological investigation requires field data collection, hypothesis-driven study design, lab collaboration, and judgment about causality that current AI cannot execute end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Professional licensure, regulatory requirements (public health authority approval of disease investigation protocols), liability for missed outbreak detection, and mandatory human sign-off on epidemiological determinations create substantial legal and organizational barriers to autonomous automation. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Public health investigations often require credentialed epidemiologists, institutional authority (e.g., CDC/health department oversight), regulatory reporting, and legal responsibility for conclusions affecting public health policy. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI tools (surveillance software, data analysis platforms) reduce costs on specific subtasks, but the total cost of AI-assisted investigation remains comparable to or higher than direct human epidemiologist effort when considering data collection, validation, and human oversight of outputs. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI can cheaply handle data analysis subtasks, but the full investigation still requires expensive expert labor, fieldwork, and lab coordination, keeping overall costs comparable to human-led work. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | AI products exist for narrow components (disease surveillance dashboards, literature mining, statistical analysis), but no deployed system reliably performs the full investigative task—determining cause, risk factors, transmission modes—which requires field investigation, specimen analysis, and integration of heterogeneous evidence sources. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | AI tools assist with literature review, data analysis, and modeling but no deployed product autonomously conducts disease investigations in production. |
Monitor and report incidents of infectious diseases to local and state health agencies.
25CI 25–25 · exposure 25 · augmentation 63 · importance 4.2/5 · click for rater detail
Monitor and report incidents of infectious diseases to local and state health agencies.
25| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Adoption of AI in public health surveillance remains limited; most agencies still rely on manual or semi-automated legacy e-surveillance systems. Pilot programs exist but production deployment of AI-assisted case classification is not yet mainstream in the field. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Public health agencies are historically slow adopters of AI due to legacy IT systems, funding constraints, and regulatory caution, despite pilot programs in disease surveillance. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist by auto-populating case report forms from clinical notes, flagging potential reportable conditions, and streamlining data entry, meaningfully reducing epidemiologist time on clerical tasks. However, the epidemiologist must always validate findings and authorize reporting, limiting the autonomy of AI assistance. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools significantly aid pattern detection, anomaly flagging, and data aggregation across large datasets, meaningfully speeding up an epidemiologist's monitoring workflow even though final reporting remains human-driven. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Disease surveillance reporting involves extracting structured data from clinical reports and submitting standardized forms, which LLMs could partially automate. However, the task requires judgment about what constitutes a reportable incident, verification against case definitions, and handling of incomplete/ambiguous clinical information—functions current AI cannot reliably perform end-to-end without substantial human oversight. |
| Task automatability | claude-sonnet-5 | 2/5 | Reporting infectious disease incidents involves case verification, judgment about classification/severity, and coordination with agencies that AI can support but not fully replace end-to-end today. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Disease reporting is legally mandated and typically requires a licensed epidemiologist or physician to verify, classify, and authorize submission to health authorities. Regulatory frameworks (CDC, state health law) embed accountability in the human reporter, creating a hard barrier to full automation. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Public health reporting is legally mandated and tied to credentialed epidemiologist/public health authority responsibility, with regulatory frameworks requiring accountable human reporting to agencies. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | The task is primarily clerical but requires expert judgment that justifies human labor cost. AI systems for data extraction and form drafting may reduce some administrative overhead, but integration, validation, and oversight costs mean AI is not substantially cheaper than deploying epidemiologic staff for this function. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI can cheaply process surveillance data streams, but the overall reporting task still requires expert verification and liability-bearing sign-off, keeping human cost dominant in the loaded total. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While draft-generation and data-extraction tools exist, no deployed AI system reliably completes disease surveillance reporting autonomously in production. Existing e-surveillance systems require manual human triage, case classification, and submission; AI assists but does not replace the epidemiologist's role. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some surveillance systems use automated data aggregation and flagging (e.g., syndromic surveillance tools), but actual case confirmation and formal reporting to health agencies remain human-led processes with narrow AI tool support. |
Plan and direct studies to investigate human or animal disease, preventive methods, and treatments for disease.
25CI 25–25 · exposure 25 · augmentation 75 · importance 4.2/5 · click for rater detail
Plan and direct studies to investigate human or animal disease, preventive methods, and treatments for disease.
25| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Public health and academic epidemiology show modest AI adoption in data analysis and reporting; study planning remains human-led with AI as a supporting tool, and organizational practices have not shifted toward algorithmic direction of studies. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Public health and academic research sectors are relatively slow adopters of AI for high-level strategic research direction, with pilots more common in narrower analytic tasks than in overall study leadership. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI substantially assists epidemiologists by automating literature synthesis, suggesting statistical methods, flagging data quality issues, and generating preliminary analyses, allowing the epidemiologist to focus on study design strategy and interpretation. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI substantially aids epidemiologists in literature review, study design brainstorming, statistical analysis, and drafting protocols, meaningfully increasing productivity while the human retains ultimate direction and judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can assist with literature review, data analysis, and hypothesis generation, the core work of planning and directing epidemiological studies requires human judgment on study design, ethical oversight, stakeholder coordination, and adaptive decision-making that current systems cannot perform end-to-end with 50% time savings at equal quality. |
| Task automatability | claude-sonnet-5 | 2/5 | Planning and directing research studies requires scientific judgment, hypothesis generation, stakeholder coordination, and strategic decision-making that current AI cannot perform end-to-end, though it can assist with literature review, protocol drafting, and data analysis subcomponents. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | IRB approval, regulatory compliance (FDA, CDC guidelines), liability for study design errors, and funder requirements typically mandate human epidemiologists sign off on study protocols; legal and institutional barriers prevent full substitution. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Institutional review boards, funding bodies, and public health authorities require credentialed epidemiologists to design and be accountable for studies, especially those involving human subjects, creating strong regulatory and liability barriers. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI inference for analytical components is cheap, but integration into study planning, expert review cycles, and oversight costs remain substantial; the all-in cost of AI-assisted study direction remains comparable to or higher than the human epidemiologist's expert labor. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI can cheaply support literature review and data crunching, but the overall task requires expert oversight, domain judgment, and accountability that still demand substantial (expensive) human epidemiologist time, keeping costs comparable rather than dramatically lower. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed product performs study planning and direction reliably; AI systems exist for specific subtasks (data analysis, literature mining) but production epidemiological systems do not independently plan, design, or direct full studies in real organizations. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | No deployed product autonomously plans and directs epidemiological studies; AI tools exist for literature synthesis, statistical modeling, and drafting but the directive/leadership function remains firmly human-led in production settings. |
Provide expertise in the design, management and evaluation of study protocols and health status questionnaires, sample selection, and analysis.
25CI 25–25 · exposure 25 · augmentation 63 · importance 4.1/5 · click for rater detail
Provide expertise in the design, management and evaluation of study protocols and health status questionnaires, sample selection, and analysis.
25| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Academic and public health institutions have adopted AI for data analysis pipelines, but protocol design and questionnaire development remain tightly held by human epidemiologists; adoption of AI for these high-stakes, regulated tasks is slow and cautious. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Public health and academic research sectors are relatively slow adopters of AI for core scientific design work, with pilots emerging but production-level autonomous protocol design not yet common. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can usefully assist epidemiologists by suggesting statistical approaches, generating questionnaire drafts, and reviewing literature for design precedents, but the epidemiologist retains responsibility for final decisions on study validity and ethics. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools substantially help epidemiologists draft questionnaires, suggest sampling frameworks, run statistical analyses, and summarize literature, meaningfully boosting productivity while the expert retains full control over design decisions. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can assist with data analysis and statistical methods for existing protocols, designing study protocols and health questionnaires fundamentally requires human domain expertise, regulatory knowledge, and judgment about research ethics that current systems cannot perform end-to-end at the quality bar required in peer-reviewed epidemiology. |
| Task automatability | claude-sonnet-5 | 2/5 | This task requires deep domain expertise, judgment on study design tradeoffs, and contextual scientific reasoning that current AI cannot fully replicate end-to-end, though AI can assist with drafting and statistical components. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Institutional Review Boards (IRBs) legally require human epidemiologist sign-off on study protocols; regulatory and liability frameworks mandate that a qualified human expert certify research design, ethics, and methodological soundness. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Study protocols in health research typically require institutional review board approval, credentialed principal investigators, and regulatory compliance (e.g., IRB, funding agency standards), creating strong barriers to full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI analysis tools are cheap for computation, but epidemiologists' specialized labor (doctorate-level expertise, regulatory compliance knowledge, liability) commands high wages; the value AI adds to protocol design specifically does not yet offset human engagement costs. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | While AI can cheaply assist with drafting questionnaires or running analyses, the overall protocol design and validation still requires expensive expert oversight, keeping the all-in cost comparable to or only modestly cheaper than human-led work. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | AI tools exist for statistical analysis and questionnaire generation assistance, but no deployed product reliably performs the full suite of protocol design, ethics review integration, and sample selection strategy that epidemiologists execute in production research environments. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | There are no deployed products that autonomously design and manage full epidemiological study protocols; existing tools (statistical software, LLM drafting assistants) only handle narrow sub-components under expert supervision. |
Prepare and analyze samples to study effects of drugs, gases, pesticides, or microorganisms on cell structure and tissue.
25CI 25–25 · exposure 25 · augmentation 63 · importance 2.7/5 · click for rater detail
Prepare and analyze samples to study effects of drugs, gases, pesticides, or microorganisms on cell structure and tissue.
25| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Adoption of computational tools for image analysis is increasing in research and some clinical labs, but the sector remains moderately conservative; many labs still rely on traditional manual microscopy and human interpretation, with limited production-scale deployment of end-to-end AI automation. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Public health and biomedical research labs adopt AI unevenly, with automation concentrated in data analysis rather than physical sample handling, resulting in slow pilot-stage adoption. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | Image analysis AI and automated microscopy platforms assist epidemiologists and lab technicians by accelerating data collection and flagging anomalies, improving throughput and consistency; however, the assistance is primarily in the analytical phase rather than transforming the entire workflow from sample preparation through interpretation. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools substantially aid image analysis, pattern recognition in tissue samples, and statistical modeling of drug/toxin effects, improving epidemiologist productivity even though physical steps remain human-driven. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Sample preparation involves substantial manual lab work (culturing, staining, mounting) that requires dexterous physical manipulation; analysis of microscopy/imaging data can be partially automated with image analysis tools, but interpreting complex cellular changes and tissue responses requires domain expertise and contextual judgment that current AI cannot fully replace end-to-end. |
| Task automatability | claude-sonnet-5 | 2/5 | Physical sample preparation and much of the wet-lab analysis require manual dexterity and judgment that current AI cannot perform end-to-end; AI can assist with data analysis but not the full task.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Regulatory frameworks (FDA, EPA, ISO standards for toxicology testing) often mandate specific protocols, chain-of-custody procedures, and human expert validation; many assays require certified technicians or oversight by licensed professionals, creating substantial legal and compliance barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Biosafety regulations, lab certification requirements, and the need for qualified personnel to handle hazardous agents and interpret results create strong institutional and regulatory barriers. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | While image analysis software reduces labor on the analytical side, sample preparation still requires trained lab technicians, and oversight of AI results by epidemiologists remains necessary; overall labor savings do not yet offset the cost of systems and integration. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Physical lab equipment, robotics, and validation costs remain high relative to skilled technician labor for this specialized task, so AI is not clearly cheaper overall. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Image analysis software and computational pathology tools exist for some analytical components, but no deployed system reliably handles the full pipeline of sample preparation, exposure control, and interpretive analysis of tissue responses without significant human oversight and validation. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | No deployed product autonomously prepares and analyzes biological samples for toxicological/microbial effects on tissue; lab automation exists but is narrow and requires human oversight. |
Plan, administer and evaluate health safety standards and programs to improve public health, conferring with health department, industry personnel, physicians, and others.
23CI 20–25 · exposure 20 · augmentation 75 · importance 3.8/5 · click for rater detail
Plan, administer and evaluate health safety standards and programs to improve public health, conferring with health department, industry personnel, physicians, and others.
23| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Epidemiology and public health agencies tend to be slower adopters of automation due to regulatory conservatism, data privacy constraints, and the requirement for human professional accountability. Adoption remains mostly in pilots and limited support tools rather than autonomous systems. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Public health agencies are historically slow adopters of AI due to bureaucratic structure, funding constraints, and data governance issues, with pilots more common than production deployment. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can substantially assist epidemiologists by automating data aggregation, statistical analysis, literature synthesis, and scenario modeling, allowing professionals to focus on stakeholder coordination, program design, and judgment-based decisions. These tools are increasingly in use to amplify productivity. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools can meaningfully assist by analyzing health data, drafting policy documents, and summarizing stakeholder input, enhancing efficiency while humans retain decision-making and relational responsibilities. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can assist with data analysis, literature review, and initial program design, the task critically requires human judgment in conferring with multiple stakeholders, interpreting complex regulatory landscapes, and making high-stakes decisions about public health standards. The coordination and negotiation components are not automatable today. |
| Task automatability | claude-sonnet-5 | 2/5 | This task involves strategic planning, cross-organizational negotiation, and judgment calls about public health standards that require contextual authority and stakeholder trust AI cannot replicate end-to-end today.4 |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Public health program planning and administration typically requires professional licensure (MPH, MD, or related credentials), legal accountability for safety standards, and often explicit regulatory requirements that a qualified human professional must sign off on recommendations. Liability asymmetry is high given public health consequences. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Public health program administration often requires credentialed epidemiologists, regulatory accountability, and legal responsibility for health outcomes, creating strong institutional and liability barriers to automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Even accounting for AI's efficiency in data processing, the task requires epidemiologists' specialized expertise, licensure, and responsibility. The all-in cost of AI infrastructure plus required human oversight and decision-making remains comparable to or higher than direct human performance. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI can cheaply draft reports or summarize data but the core conferring, negotiation, and program administration still requires paid expert labor, keeping costs comparable to human-driven work. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed products reliably perform the full scope of this task end-to-end. AI systems can support components (data analysis, report generation) but lack the capability to genuinely plan, administer, and evaluate programs while coordinating across diverse human stakeholders in production settings. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product plans and administers public health programs autonomously; this remains a human-led, multi-stakeholder governance function with no production-scale AI substitute. |
Teach principles of medicine and medical and laboratory procedures to physicians, residents, students, and technicians.
21CI 16–25 · exposure 17 · augmentation 63 · importance 2.9/5 · click for rater detail
Teach principles of medicine and medical and laboratory procedures to physicians, residents, students, and technicians.
21| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Medical education institutions are exploring AI-assisted learning tools (simulations, practice questions), but actual displacement of physician educators is minimal; adoption remains in pilot and supplementary phases. The sector's conservative posture toward credentialing and training responsibility slows adoption. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Academic medicine adopts AI slowly for core teaching functions, though e-learning platforms and case-based AI tools are gaining some traction as supplements. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist educators by generating example cases, providing instant access to literature, creating practice questions, or handling administrative grading, meaningfully improving instructor productivity. However, the core teaching relationship remains human-centered, limiting transformative impact. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can generate teaching materials, quizzes, case studies, and provide on-demand explanations, meaningfully supporting instructors preparing and delivering curricula. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Teaching complex medical principles and procedures requires real-time interaction, adaptive explanation to different learner levels, hands-on demonstration feedback, and nuanced judgment about comprehension—capabilities far beyond current AI systems operating autonomously. The task demands live Q&A, corrections, and mentoring that cannot be fully automated today. |
| Task automatability | claude-sonnet-5 | 2/5 | Teaching involves live demonstration, mentorship, hands-on lab supervision, and adaptive pedagogy that current AI cannot fully replicate end-to-end, though lecture content generation can be partially automated.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Teaching in medical education is bound by accreditation requirements, liability (students must learn correctly), institutional trust in faculty credentials, and regulatory expectations that a credentialed physician lead instruction. Organizational and regulatory friction against full automation is substantial. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Medical education often requires credentialed faculty, accreditation standards, and hands-on supervision for licensure-track training, creating substantial institutional and regulatory barriers. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI content generation and tutoring systems cost substantially less than a physician educator per hour, but the quality gap and need for human oversight and customization mean the effective cost per equivalent instructional outcome remains higher than deploying an AI-only solution would suggest. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Human instructors carry salary costs, but AI cannot fully substitute for supervised lab teaching, so cost comparison favors humans once oversight and liability are included. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While AI can generate educational content, create videos, or answer factual questions, no deployed system reliably performs the full teaching task (instruction, assessment, mentoring, real-time correction) in medical settings. Products exist for narrow components (video generation, Q&A) but fall far short of replacing an instructor. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | AI tutoring and content-generation tools exist for medical education support, but no deployed product independently teaches clinical/laboratory procedures to trainees in production settings. |
Oversee public health programs, including statistical analysis, health care planning, surveillance systems, and public health improvement.
18CI 11–25 · exposure 13 · augmentation 63 · importance 4.3/5 · click for rater detail
Oversee public health programs, including statistical analysis, health care planning, surveillance systems, and public health improvement.
18| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Public health agencies operate under strict regulatory and governance constraints with slow digital infrastructure modernization. While some agencies use AI for data analysis within programs, autonomous program oversight automation is nascent and adoption remains limited to analytical enhancements rather than operational replacement. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Public health agencies are generally slower adopters of AI compared to finance or tech, with pilots for surveillance analytics emerging but full programmatic AI oversight remaining rare. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can meaningfully assist epidemiologists by automating statistical computation, flagging surveillance anomalies, and generating preliminary health trend reports, thereby raising analytical productivity. However, augmentation is bounded to analytical components rather than program strategy and stakeholder management. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can meaningfully assist with statistical analysis, surveillance data processing, forecasting, and report drafting, significantly boosting productivity for epidemiologists who remain in charge of oversight. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can assist with statistical analysis and data surveillance components, overseeing public health programs requires integrated judgment across policy, resource allocation, stakeholder engagement, and real-time program adaptation that current systems cannot perform end-to-end at 50% time savings. The task fundamentally requires domain expertise and human accountability that AI cannot yet replace. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a high-level oversight and management task combining strategic planning, stakeholder coordination, and program leadership; AI cannot autonomously oversee programs or make accountable public health decisions today. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Public health program oversight involves regulatory authority, legal accountability for health decisions, mandatory reporting requirements, and stakeholder trust that effectively require a licensed epidemiologist or public health official to sign off. Organizations cannot substitute AI decision-making for human professional accountability. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Public health program leadership often requires credentialed epidemiologists, government accountability, and legal responsibility for public health decisions, creating strong institutional and regulatory barriers to full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI inference for analytical components is cheap, but full program oversight requires significant human epidemiologist oversight and integration costs that remain comparable to or exceed the cost of direct human epidemiologist labor given liability and decision quality requirements. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | While AI can cheaply support statistical analysis pieces, the overall oversight task requires sustained human judgment, coordination, and accountability, so AI cannot substitute for the human role at comparable output quality, keeping effective cost savings low. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Deployed AI products handle discrete sub-tasks (statistical analysis, data visualization, trend detection) but no production system reliably performs the integrated oversight function across planning, surveillance, and improvement initiatives. Existing tools are narrowly scoped components, not full program management systems. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs holistic public health program oversight; existing tools address narrow sub-components like data analysis, not the integrated management function described. |
Supervise professional, technical, and clerical personnel.
4CI 0–7 · exposure 0 · augmentation 38 · importance 3.4/5 · click for rater detail
Supervise professional, technical, and clerical personnel.
4| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Despite high digitization in epidemiology organizations, human supervision of personnel remains a core function with no meaningful production adoption of AI agents; organizational and legal barriers prevent displacement. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Public health and research organizations are slow to adopt AI for management functions, with AI use concentrated in analytical support rather than supervisory roles. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could assist with scheduling, performance metrics tracking, or documentation, but the core supervisory tasks of coaching, accountability, and decision-making remain human-centric with only marginal productivity gains possible. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI tools can help with scheduling, performance tracking, report generation, and communication drafting, aiding but not replacing supervisory judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Supervising personnel requires real-time judgment about performance, conflict resolution, motivation, and strategic direction—fundamentally human activities involving agency and accountability. Current AI cannot meaningfully manage or be held responsible for personnel decisions. |
| Task automatability | claude-sonnet-5 | 1/5 | Direct personnel supervision requires interpersonal leadership, performance evaluation, and contextual judgment that current AI cannot execute end-to-end.value |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Supervision carries legal and fiduciary responsibilities (employment law, discrimination protections, performance accountability) that require a licensed, legally responsible human to execute and sign off on personnel decisions. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Supervisory authority typically requires a designated human manager with accountability, HR/legal responsibility, and organizational trust, creating strong structural barriers to automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The full-cost AI solution (integration, oversight, legal liability) would exceed the loaded wage of a supervisory role, which already involves substantial human judgment and accountability that cannot be efficiently automated. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI substitute performing this managerial function, so no favorable cost comparison exists. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI product reliably performs personnel supervision; this task requires ongoing two-way relationships, legal employment authority, and professional judgment that AI systems cannot execute in production environments. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product autonomously supervises staff; management software only supports scheduling or tracking, not actual supervision. |
Related occupations — Life, Physical & Social Science
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.