Biologists

19-1029.04

Median wage $98,920/yr55,850 employed (US)Rank #358 of 923 scored · top 39% by substitution

Research or study basic principles of plant and animal life, such as origin, relationship, development, anatomy, and functions.

Sub-scores

0–100 · band = confidence interval from rater disagreement

Substitution31

Exposure28

Augmentation67

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

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

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

Tasks on the substitution scale

22 rated tasks, binned by substitution score.

Position among all scored occupations

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

Tasks with substitution ≥ 70

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%29

panel mean rating 2.2/5 → substitution pressure 29/100

Technical feasibility todayw 20%26

panel mean rating 2.1/5 → substitution pressure 26/100

Cost vs. human wagew 15%30

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

Adoption barriersw 20%inverted — strong barriers lower the score41

panel mean rating 3.4/5 (barrier strength) → substitution pressure 41/100

Sector adoption velocityw 10%29

panel mean rating 2.2/5 → substitution pressure 29/100

Task breakdown (22 tasks)

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

Program and use computers to store, process, and analyze data.

CI 72–75 · exposure 75 · augmentation 100 · 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	4/5	Bioinformatics and biotech sectors have rapidly adopted AI-assisted coding, automated pipelines, and data analysis tools; GitHub Copilot and similar tools are widely deployed in research labs and biotech firms.
Sector adoption velocity	claude-sonnet-5	3/5	Academic and research biology has moderate digitization and growing use of AI coding tools, but adoption is uneven across labs and slower than in finance or software-native industries.
Augmentation potential	claude-haiku-4-5-20251001	5/5	AI substantially assists biologists in writing, debugging, and optimizing code, and in exploratory data analysis, dramatically reducing time spent on routine programming while keeping the scientist in control of experimental design and interpretation.
Augmentation potential	claude-sonnet-5	5/5	AI dramatically accelerates writing analysis code, debugging, and exploring datasets, letting biologists focus on experimental design and interpretation while staying in the loop for validation.
Task automatability	claude-haiku-4-5-20251001	4/5	Modern AI systems (LLMs with code generation, automated statistical tools) can write, execute, and refine data processing and analysis pipelines with minimal human oversight, often achieving substantial time savings. However, domain-specific validation and interpretation typically still require human biologist expertise, preventing a full 5 rating.
Task automatability	claude-sonnet-5	4/5	Coding assistants and AI-based data analysis tools can generate scripts, perform statistical analysis, and produce visualizations for typical biological datasets with significant time savings, though complex custom pipelines still need human validation.
Adoption barriers	claude-haiku-4-5-20251001	2/5	Minimal legal or licensing barriers exist to automating the programming and data storage itself; however, accountability for analytical accuracy and interpretation means organizations typically require human oversight and sign-off on results.
Adoption barriers	claude-sonnet-5	2/5	No licensing requirement blocks using AI for data programming/analysis, though scientific rigor and reproducibility norms in publications create some institutional friction and require human verification of results.
Cost vs. human wage	claude-haiku-4-5-20251001	4/5	AI inference and cloud compute for data processing are substantially cheaper than the loaded cost of a biologist's labor per task-equivalent, often by an order of magnitude when accounting only for the mechanical programming and analysis steps.
Cost vs. human wage	claude-sonnet-5	4/5	AI coding/analysis assistance costs a small subscription fee versus substantial biologist or data analyst hourly wages, though oversight and validation of scientific analyses add some cost overhead.
Technical feasibility today	claude-haiku-4-5-20251001	4/5	Deployed products (GitHub Copilot, ChatGPT for coding, specialized bioinformatics platforms with AI assistants) reliably perform code generation and data processing tasks in production. Error rates on routine analysis are low, though interpretation of novel biological findings still requires human review.
Technical feasibility today	claude-sonnet-5	4/5	Products like GitHub Copilot, ChatGPT code interpreter, and specialized bioinformatics AI tools are widely deployed and reliably assist with data processing scripts and statistical analysis in production research settings today.

Identify, classify, and study structure, behavior, ecology, physiology, nutrition, culture, and distribution of plant and animal species.

CI 30–84 · exposure 58 · augmentation 88 · importance 4.0/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	4/5	Academic biology, conservation organizations, and pharmaceutical/agricultural sectors are actively deploying AI for genomic analysis, species identification, and ecological modeling in production workflows. Adoption is measurable and accelerating, especially in genomics and computer vision applications.
Sector adoption velocity	claude-sonnet-5	2/5	Academic and field biology remains a slower-adopting sector compared to information/finance industries, with AI used mainly in narrow analytic pilots rather than broad production deployment.
Augmentation potential	claude-haiku-4-5-20251001	5/5	AI dramatically amplifies biologist productivity: automated specimen sorting, instant species ID from images, rapid literature synthesis, and large-scale genomic analysis allow researchers to focus on hypothesis design, field strategy, and interpretation—classic augmentation of expert judgment.
Augmentation potential	claude-sonnet-5	4/5	AI substantially aids literature review, image-based species identification, data analysis, and pattern detection in ecological datasets, meaningfully boosting researcher productivity while humans remain central to fieldwork and interpretation.
Task automatability	claude-haiku-4-5-20251001	5/5	Modern AI systems can perform large portions of species identification, classification, and distribution mapping using computer vision (image recognition) and ecological modeling. Genomic sequencing paired with AI analysis can substantially automate physiology and nutrition studies, meeting the ≥50% time-saving threshold when integrated with databases and literature mining tools.
Task automatability	claude-sonnet-5	2/5	This task spans field observation, experimental design, and specialized taxonomic judgment that current AI cannot perform end-to-end; AI can assist with data analysis and literature review but not the full study workflow.itezed.
Adoption barriers	claude-haiku-4-5-20251001	2/5	No legal licensing prevents AI use in species study; biologists retain agency over research design and interpretation. Organizational barriers are minimal—academic and industry labs already integrate automated tools—and peer review remains the primary human checkpoint.
Adoption barriers	claude-sonnet-5	3/5	No strict licensing requirement for most biological research, but publication standards, peer review, and institutional science norms create real friction against pure AI-driven output.
Cost vs. human wage	claude-haiku-4-5-20251001	4/5	Once trained and integrated, AI inference for species identification, genomic analysis, and distribution modeling costs pennies per specimen compared to highly trained biologists earning $70k+. The cost advantage approaches an order of magnitude for routine identification and data processing.
Cost vs. human wage	claude-sonnet-5	2/5	AI tools for identification or data processing are cheap, but the core scientific work still requires expensive human expertise, fieldwork, and equipment that AI cannot substitute for at scale.
Technical feasibility today	claude-haiku-4-5-20251001	4/5	Deployed products (iNaturalist AI, image-based species identification APIs, genomic analysis pipelines) reliably handle identification and classification at scale in production. However, nuanced behavioral ecology and culture studies remain partially manual, limiting full end-to-end deployment to approximately 70–80% of subtasks.
Technical feasibility today	claude-sonnet-5	2/5	Products exist for species identification (image classifiers, iNaturalist-style tools) and literature synthesis, but no deployed system reliably performs the full breadth of structural, behavioral, and ecological study described.

Prepare technical and research reports, such as environmental impact reports, and communicate the results to individuals in industry, government, or the general public.

CI 43–56 · exposure 58 · augmentation 75 · importance 4.1/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	While research and consulting organizations increasingly use AI for drafting support, the adoption of AI-generated reports without significant human revision remains limited in regulated or high-stakes contexts (environmental reporting, government compliance). The sector lags high-tech adoption patterns.
Sector adoption velocity	claude-sonnet-5	2/5	Biological/environmental science sectors have historically slower AI adoption compared to finance or information services, though report-drafting tools are gradually being piloted.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI can meaningfully accelerate report writing by drafting sections, organizing data, suggesting visualizations, and handling formatting, while the biologist retains control over interpretation and stakeholder communication. This assistive pattern is already observable in practice.
Augmentation potential	claude-sonnet-5	4/5	AI tools significantly help biologists draft, structure, and summarize technical content, speeding up report writing while the scientist retains responsibility for accuracy and interpretation.
Task automatability	claude-haiku-4-5-20251001	4/5	Large portions of report generation—data aggregation, visualization, initial draft synthesis, and formatting—can be automated with current AI systems at significant time savings. However, the interpretive framing, novel synthesis of complex findings, and stakeholder-specific messaging typically require biologist judgment, preventing full 5-rating end-to-end automation.
Task automatability	claude-sonnet-5	3/5	AI can draft substantial portions of technical reports (summarizing data, generating standard sections, formatting) but requires domain expertise, judgment on data interpretation, and validation that current systems cannot fully replace end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	3/5	Environmental impact and regulatory reports often require professional sign-off or compliance with specific government/institutional standards, and stakeholders (industry, government) frequently expect credible human authorship and accountability. These create friction but do not absolutely prohibit AI-assisted or templated generation.
Adoption barriers	claude-sonnet-5	4/5	Environmental impact reports often require credentialed professionals to certify accuracy and are subject to regulatory review (e.g., NEPA), creating strong professional and legal barriers to full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	3/5	AI writing, data processing, and formatting tools cost substantially less per task than senior biologist time, but oversight, fact-checking, and revision by a qualified biologist remain necessary, bringing the total cost closer to rough parity.
Cost vs. human wage	claude-sonnet-5	3/5	AI drafting reduces time on boilerplate and summarization, but human expert review, fieldwork data integration, and liability checks remain costly, keeping overall costs roughly comparable to a human-led process augmented by AI.
Technical feasibility today	claude-haiku-4-5-20251001	3/5	AI writing tools and report generators exist and are deployed in some research and consulting contexts, but they still produce material errors in technical accuracy, selective citation, and domain-specific nuance that require expert review. Production use remains inconsistent across organizations.
Technical feasibility today	claude-sonnet-5	3/5	AI writing assistants and specialized tools are used in production to draft scientific/technical documents, but scientists still heavily edit and verify for accuracy, especially for regulatory environmental impact reports.

Write grant proposals to obtain funding for biological research.

CI 29–59 · exposure 38 · augmentation 88 · importance 3.8/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Grant writing automation adoption remains in early-stage pilots within academic institutions; most biologists still write proposals directly or with human collaborators. Sector digitization and risk-aversion (funder skepticism of AI-authored proposals) slow deployment significantly.
Sector adoption velocity	claude-sonnet-5	3/5	Academic and research sectors are adopting AI writing tools moderately quickly for drafting support, but institutional caution, integrity concerns, and funder policies slow full-scale production use.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI excels at assisting grant writers by generating literature summaries, outlining sections, spotting formatting gaps, and offering alternative phrasings. These augmentations meaningfully accelerate proposal development when a human investigator remains in control of scientific content and strategy.
Augmentation potential	claude-sonnet-5	5/5	AI substantially speeds up drafting, editing, formatting, and literature summarization for grant proposals, letting biologists focus more time on experimental design and novel ideas while staying fully in control of scientific content.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can draft sections (literature review, methods summary), grant proposals require substantive scientific judgment, novel hypothesis framing, and persuasive argumentation tailored to specific funding agencies—elements that demand human expertise and creativity. Current AI cannot reliably produce competitive proposals end-to-end without extensive human revision.
Task automatability	claude-sonnet-5	3/5	LLMs can draft substantial portions of grant proposal text (background, significance, methods framing) given input from the researcher, but crafting a compelling, novel, fundable proposal still requires deep domain expertise, original ideas, and iterative refinement that AI cannot fully replace.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Funding agencies typically require that grant proposals reflect the investigator's own scientific vision and institutional accountability; delegating proposal writing to automated systems raises ethical and funder-credibility concerns. Additionally, many agencies implicitly or explicitly expect human authorship and originality.
Adoption barriers	claude-sonnet-5	2/5	No licensing requirement, but funding agencies expect the PI's genuine intellectual contribution and there are norms/policies at some agencies restricting undisclosed AI-generated content, creating moderate friction.
Cost vs. human wage	claude-haiku-4-5-20251001	3/5	Using AI for grant drafting assistance (e.g., outline generation, literature synthesis) is cheaper than hiring a dedicated grant writer, but the loaded cost of a researcher's time to oversee and refine AI output approaches the cost of writing proposals directly, especially given revision overhead.
Cost vs. human wage	claude-sonnet-5	4/5	AI drafting assistance costs a small fraction of the many hours a biologist would spend drafting sections from scratch, though human oversight and revision costs remain necessary, keeping it below the top tier.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	No mature product reliably generates full grant proposals that reviewers would fund at production scale. AI writing assistants exist for brainstorming and drafting components, but deployed systems lack the domain-specific credibility and strategic fit required for grant success.
Technical feasibility today	claude-sonnet-5	3/5	Deployed AI writing tools (ChatGPT, Claude, specialized grant-writing assistants) are used in production by researchers today for drafting and editing, but reliability varies and human review/rewriting is essential for scientific accuracy and persuasiveness.

Prepare requests for proposals or statements of work.

CI 25–56 · exposure 38 · augmentation 63 · importance 3.6/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Academic and research institutions adopt AI tools slowly for contract and proposal work due to risk aversion, regulatory oversight, and the requirement for human expert review before submission.
Sector adoption velocity	claude-sonnet-5	2/5	Academic and research institutions adopt general AI writing tools unevenly and often with policy restrictions, so uptake for formal procurement documents remains slow and inconsistent.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can meaningfully assist by generating drafts, organizing sections, and checking formatting, allowing a biologist to iterate and refine faster; however, final judgment and domain-specific content remain human responsibilities.
Augmentation potential	claude-sonnet-5	4/5	AI is well-suited to drafting boilerplate sections, formatting, and generating first drafts, meaningfully speeding up the biologist's work of preparing these documents while they retain final control.
Task automatability	claude-haiku-4-5-20251001	2/5	AI can draft sections of RFPs and SOWs (e.g., boilerplate text, scope templates), but the task requires domain expertise, negotiation context, and compliance considerations specific to each biological research project that current systems handle unreliably without substantial human revision.
Task automatability	claude-sonnet-5	3/5	AI can draft substantial portions of an RFP or statement of work from templates and background material, but it requires human input on scientific scope, budget, and specific deliverables, limiting full end-to-end automation.
Adoption barriers	claude-haiku-4-5-20251001	4/5	RFPs and SOWs carry legal and contractual weight; institutional review, compliance with funding agency rules, and PI sign-off are typically mandatory, creating strong friction against full automation.
Adoption barriers	claude-sonnet-5	2/5	There's no licensing requirement to write an RFP, but institutional procurement policies and grant compliance rules mean a qualified person typically must review and finalize it.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI assistance with drafting may reduce time on routine boilerplate, but the cost of oversight, correction, and legal review often approaches or exceeds the cost of a junior scientist drafting from scratch, especially given liability concerns.
Cost vs. human wage	claude-sonnet-5	4/5	Drafting text with an LLM is far cheaper per page than scientist or administrative staff time, though the human review and domain-specific customization still add cost.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	While LLMs can generate RFP/SOW text, no mature production system reliably produces legally sound, scientifically precise, and project-specific proposals without significant human oversight and rework by domain experts.
Technical feasibility today	claude-sonnet-5	3/5	General-purpose LLM products (e.g., ChatGPT, Copilot) are routinely used to draft procurement and grant-related documents, but no specialized biology-domain product reliably automates this task in production without heavy editing.

Review reports and proposals, such as those relating to land use classifications and recreational development, for accuracy, adequacy, or adherence to policies, regulations, or scientific standards.

CI 25–43 · exposure 33 · augmentation 63 · 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	2/5	Adoption in government and conservation agencies remains slow; these sectors are not at the forefront of AI deployment. Most reviews continue to rely on human experts, with limited pilot adoption of AI-assisted tools even in more digital organizations.
Sector adoption velocity	claude-sonnet-5	2/5	Environmental and government/regulatory sectors tend to adopt AI more slowly due to compliance concerns, procurement cycles, and lower digitization compared to finance or software industries.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can usefully assist by summarizing documents, flagging inconsistencies with published standards, and organizing information for the reviewer, meaningfully raising the productivity of the human expert doing the review, even though the expert retains final authority.
Augmentation potential	claude-sonnet-5	4/5	AI can efficiently summarize lengthy reports, cross-check against regulations, and highlight potential inconsistencies, meaningfully speeding up the reviewer's workflow while the biologist retains final judgment.
Task automatability	claude-haiku-4-5-20251001	2/5	AI can assist with document parsing and flagging potential deviations from known policies or standards, but reviewing for scientific accuracy, adequacy of evidence, and complex adherence judgments requires domain expertise and contextual judgment that current systems cannot reliably perform end-to-end. Substantial human oversight would remain necessary.
Task automatability	claude-sonnet-5	3/5	AI can draft summaries and flag inconsistencies or missing elements in reports against known standards, but full evaluation of scientific adequacy and policy adherence requires domain judgment that current systems only partially replicate.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Strong barriers exist: regulatory frameworks often explicitly require a qualified biologist or scientist to sign off on compliance and accuracy; liability for missed errors in land-use or environmental policy is high; and institutional policy typically mandates expert human review before proposals are approved.
Adoption barriers	claude-sonnet-5	3/5	While no strict licensing mandates a biologist sign off on every report, agencies and regulatory bodies typically require professional review and accountability for land-use and environmental decisions, creating moderate institutional friction.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI-assisted review (document parsing + human verification) is unlikely to be significantly cheaper than direct human expert review, since the human still performs the critical judgment work. Integration and oversight costs add overhead without proportional savings.
Cost vs. human wage	claude-sonnet-5	3/5	AI can cheaply do first-pass screening for formatting, completeness, or regulatory keyword adherence, but human expert review is still needed for scientific validity, keeping overall cost savings moderate rather than dramatic.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	No deployed product reliably performs comprehensive scientific and regulatory review of complex proposals. LLMs can summarize and highlight keywords, but they lack the specialized biological knowledge, contextual understanding, and liability tolerance to serve as a primary review tool in production.
Technical feasibility today	claude-sonnet-5	2/5	Document review and compliance-checking tools exist but are not widely deployed specifically for biological/land-use scientific review with reliable accuracy; most current use is ad hoc via general LLMs rather than validated production systems.

Measure salinity, acidity, light, oxygen content, and other physical conditions of water to determine their relationship to aquatic life.

CI 25–42 · exposure 25 · 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	3/5	Environmental agencies and research institutions are piloting automated water quality sensors and data analytics platforms, but adoption remains mixed; many organizations still rely on manual sampling and expert interpretation. Production-scale AI-driven aquatic ecosystem monitoring is nascent rather than widespread.
Sector adoption velocity	claude-sonnet-5	2/5	Environmental and biological sciences are slower to adopt AI/automation for field data collection compared to information-heavy sectors, though sensor networks are growing.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI excels at processing large volumes of water quality sensor data, identifying trends, and flagging anomalies that biologists can then investigate. Machine learning can assist in pattern recognition across multiple physical parameters and historical records, meaningfully raising a biologist's ability to identify ecosystem changes and generate hypotheses about causation.
Augmentation potential	claude-sonnet-5	4/5	AI and automated sensors substantially aid data collection, real-time monitoring, and analysis of physical water conditions, improving efficiency while humans still perform or oversee sampling.
Task automatability	claude-haiku-4-5-20251001	2/5	Water quality measurement relies heavily on specialized instrumentation (probes, sensors) that aquatic biologists operate, but interpreting the relationship between physical conditions and aquatic life requires ecological expertise and field judgment. Current AI cannot autonomously design sampling protocols, select appropriate measurement sites, or synthesize physical data with biological observations to draw causal inferences about ecosystem health.
Task automatability	claude-sonnet-5	2/5	Physical measurement of water samples requires field/lab instrumentation and handling that AI cannot perform end-to-end; AI can assist with data logging and analysis but not the physical measurement itself.'
Adoption barriers	claude-haiku-4-5-20251001	3/5	Environmental monitoring and reporting often fall under regulatory frameworks (EPA, state water quality standards) that require human expert sign-off on data validity and interpretation. Liability and accuracy requirements for environmental compliance create organizational friction, though the instrumentation itself is not strictly licensed.
Adoption barriers	claude-sonnet-5	2/5	No strict licensing requirement, but fieldwork often requires physical presence, safety training, and site access that create practical barriers to full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Specialized water quality monitoring equipment and skilled biologists remain costly to deploy. While AI data analysis can reduce some processing costs, the instrumentation, field labor, and expert interpretation needed to measure conditions and establish biological relationships means total automation cost remains comparable to or slightly higher than current human-intensive workflows.
Cost vs. human wage	claude-sonnet-5	2/5	Sensors and automated probes exist but still require human deployment, calibration, and maintenance, so overall cost savings versus a technician are limited.
Technical feasibility today	claude-haiku-4-5-20251001	3/5	AI can assist with data processing and statistical analysis of water quality measurements (e.g., correlating sensor readings with species populations), and automated water quality sensors exist in production. However, no deployed AI system reliably performs the full task of determining relationships between physical conditions and aquatic life; this typically requires human expertise to interpret complex ecological patterns and field conditions.
Technical feasibility today	claude-sonnet-5	1/5	No deployed AI product autonomously conducts field water sampling and measurement; this remains a manual/instrument-driven task requiring human or robotic sensor deployment.

Collect and analyze biological data about relationships among and between organisms and their environment.

CI 30–35 · exposure 25 · 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	2/5	Life sciences sectors show slower AI adoption than tech/finance; adoption is mostly in pilot image-recognition and database tools, with limited production displacement of biologist roles in field and analytical work.
Sector adoption velocity	claude-sonnet-5	2/5	Academic and environmental research sectors are relatively slow adopters of AI compared to finance or tech, with pilots for data analysis emerging but fieldwork largely unchanged.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI demonstrates strong augmentation potential: image-based species ID, automated data logging, statistical analysis assistance, and literature mining meaningfully accelerate biologists' workflows while they retain design, interpretation, and judgment roles.
Augmentation potential	claude-sonnet-5	4/5	AI substantially assists with statistical modeling, pattern detection in ecological datasets, literature synthesis, and image-based species classification, meaningfully boosting biologist productivity while they remain central to hypothesis generation and fieldwork.
Task automatability	claude-haiku-4-5-20251001	2/5	AI can assist with data collection (e.g., image recognition for species identification, automated sensor data processing) and parts of analysis (statistical modeling, literature review), but current systems lack the field judgment, hypothesis generation, and adaptive sampling needed for end-to-end task performance with 50% time savings at equal quality.
Task automatability	claude-sonnet-5	2/5	Data collection is often physical fieldwork requiring human presence and judgment, while analysis can be partially automated, but the full end-to-end task including sampling, experimental design, and interpretation of ecological relationships still requires substantial human expertise.
Adoption barriers	claude-haiku-4-5-20251001	3/5	Regulatory and professional barriers are moderate: field collection often requires permits and institutional oversight, and many organizations value human expertise for rigor and accountability, but no hard legal requirement mandates human performance of the entire task.
Adoption barriers	claude-sonnet-5	2/5	No licensing requirement for the task itself, though scientific publication and grant-funded research often require named human researchers for accountability and peer review, creating mild organizational friction.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI tools for image analysis and data processing are relatively cheap, but integration, field deployment, and necessary human validation add overhead; moreover, biologists' labor for hypothesis-driven field work and interpretation remains hard to undercut on pure cost.
Cost vs. human wage	claude-sonnet-5	2/5	Field data collection still requires human labor and equipment costs comparable to or exceeding AI-assisted analysis costs, so overall task cost savings are modest rather than order-of-magnitude.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Deployed products exist for narrow components (species identification via image, basic data logging), but reliable production systems for the full task of designing and executing ecological surveys, interpreting organism-environment relationships, and validating findings remain limited and typically require significant human oversight.
Technical feasibility today	claude-sonnet-5	2/5	AI tools exist for statistical analysis, species identification from images, and data processing, but no deployed product autonomously performs field data collection and holistic biological interpretation reliably in production.

Develop and maintain liaisons and effective working relations with groups and individuals, agencies, and the public to encourage cooperative management strategies or to develop information and interpret findings.

CI 7–57 · exposure 28 · augmentation 63 · importance 4.0/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	3/5	Life sciences organizations are moderately digitizing their communication and administrative workflows (email automation, report generation), but stakeholder engagement remains a high-touch function where adoption of AI agents is still in pilot phases rather than mature production deployment across the sector.
Sector adoption velocity	claude-sonnet-5	2/5	Biology and environmental management sectors have shown slow, uneven AI adoption, particularly for public-facing relational and policy liaison work.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI systems can powerfully assist biologists in preparing communications, compiling and visualizing research findings, and managing liaison schedules, freeing them to focus on substantive negotiation and relationship-building. This augmentation materially boosts productivity while preserving human judgment in the critical interpersonal dimension.
Augmentation potential	claude-sonnet-5	3/5	AI can help draft communications, summarize findings for stakeholders, and prepare briefing materials, meaningfully supporting but not replacing the human relationship-management core of this task.
Task automatability	claude-haiku-4-5-20251001	4/5	A significant portion of this task—drafting communications, synthesizing findings into accessible formats, and scheduling liaison meetings—can be automated or substantially accelerated by AI systems. However, the core requirement of developing 'effective working relations' through nuanced interpersonal negotiation and building trust across diverse stakeholders remains difficult for AI to fully replicate end-to-end, limiting this to high rather than maximal automatability.
Task automatability	claude-sonnet-5	1/5	This task centers on building trust, relationships, and negotiated cooperation with stakeholders over time, which requires genuine human presence, credibility, and social judgment that current AI cannot replicate end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	3/5	Public agencies and conservation groups often expect human judgment and accountability in stakeholder engagement, creating some organizational and reputational friction against full automation. However, no strict legal or licensing requirement mandates human performance, and AI-assisted liaison work is increasingly acceptable if outcomes remain credible.
Adoption barriers	claude-sonnet-5	4/5	Effective stakeholder liaison often requires accountable, authorized human representatives, especially when interpreting scientific findings for regulatory or public trust purposes, creating strong organizational and credibility barriers to automation.
Cost vs. human wage	claude-haiku-4-5-20251001	4/5	AI-driven tools for drafting communications, managing information workflows, and tracking stakeholder engagement are substantially cheaper than full human salary equivalents, though oversight and human relationship-building remain necessary. The informational and administrative components represent significant cost savings once overhead is amortized.
Cost vs. human wage	claude-sonnet-5	1/5	Since AI cannot substitute for the relationship-building itself, any AI cost would be additive to, not replacing, the human labor required, making it more expensive relative to output.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	While AI can assist with information synthesis and initial communication drafting, no deployed product reliably handles the full task of building and maintaining effective stakeholder relationships at production scale. Current AI systems lack the contextual judgment, relationship memory, and adaptive interpersonal skills required for sustained liaison work; deployment remains experimental and narrow.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product manages ongoing interpersonal liaison relationships with agencies and the public; this remains firmly outside current AI product capabilities.

Study basic principles of plant and animal life, such as origin, relationship, development, anatomy, and function.

CI 30–32 · exposure 25 · augmentation 75 · importance 4.0/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	While research institutions are adopting AI for data analysis and literature synthesis, the core investigative work of studying foundational biological principles remains human-centric. Adoption is concentrated in data-heavy subfields (bioinformatics) rather than broad organism-level or developmental biology research.
Sector adoption velocity	claude-sonnet-5	3/5	Academic and biotech research sectors are adopting AI tools (literature search, protein structure prediction, data analysis) at a moderate pace, with pilots and augmentation common but full automation of foundational research rare.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI significantly assists biologists in literature review, data analysis, hypothesis generation, and identifying patterns in large datasets. Machine learning for microscopy image analysis, sequence databases, and literature mining measurably raise researcher productivity while humans remain essential for experimental design and interpretation.
Augmentation potential	claude-sonnet-5	4/5	AI substantially aids literature review, data analysis, hypothesis generation, and summarization, meaningfully boosting biologist productivity while humans retain scientific judgment and experimental control.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can assist in analyzing existing biological data, literature review, and hypothesis generation, the core work of studying foundational biological principles requires empirical observation, experimental design, and theoretical synthesis that humans must oversee. Automated systems cannot yet replace the full investigative loop of forming questions, designing experiments, and interpreting novel biological phenomena at equal quality.
Task automatability	claude-sonnet-5	2/5	AI can retrieve, summarize, and synthesize biological literature but the underlying task involves original research, hypothesis generation, and experimental investigation that current AI cannot fully execute end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	3/5	Regulatory requirements (IRB/IACUC approval, biosafety oversight) and institutional norms require human researchers to take responsibility for experimental design and ethical conduct. However, these are governance barriers rather than hard legal prohibitions on tool use, allowing partial automation and augmentation.
Adoption barriers	claude-sonnet-5	3/5	No licensing barrier prevents AI use, but scientific credibility, peer review, funding accountability, and institutional norms create friction against fully automating research conclusions.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Current AI infrastructure (compute, software, human oversight) for biological research remains expensive relative to direct human researcher labor, especially when factoring in validation, troubleshooting failed automation, and domain expertise required to steer the research.
Cost vs. human wage	claude-sonnet-5	2/5	AI can cheaply assist with literature synthesis and data analysis, but the core research process still requires expensive human expertise, lab work, and validation, keeping overall costs comparable to human-led research.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	AI systems can perform narrow subtasks like literature mining and sequence analysis in production, but no deployed product reliably conducts independent biological research or replaces a biologist's understanding of fundamental principles. Research-stage tools exist for specific assays or data interpretation, but end-to-end principle discovery remains human-driven.
Technical feasibility today	claude-sonnet-5	2/5	Products like literature-review assistants and AI research tools exist and are used by scientists, but no deployed system autonomously conducts basic biological research at reliable quality without heavy human direction.

Research environmental effects of present and potential uses of land and water areas, determining methods of improving environmental conditions or such outputs as crop yields.

CI 25–30 · exposure 25 · augmentation 75 · importance 4.2/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Environmental and agricultural research sectors adopt AI tools slowly for analytics, but primary research design and hypothesis testing remain human-driven. Even digitally native sectors move cautiously on this task given regulatory and credibility requirements; production automation is rare.
Sector adoption velocity	claude-sonnet-5	2/5	Environmental and biological sciences sectors show slower AI adoption compared to finance or tech, with AI mainly used for auxiliary data analysis rather than core field research.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI substantially assists biologists by processing satellite and sensor data, running environmental models, synthesizing literature, and generating preliminary analyses. These capabilities markedly improve researcher productivity in data exploration and hypothesis testing while the biologist retains central control over scientific method and conclusions.
Augmentation potential	claude-sonnet-5	4/5	AI substantially aids literature reviews, statistical modeling, remote sensing data interpretation, and report drafting, meaningfully boosting researcher productivity while humans retain core judgment and fieldwork roles.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can analyze existing environmental data and model land/water impacts, the core task requires field research design, hypothesis formulation, and iterative problem-solving that demand human expertise and judgment. Data processing could be partially automated, but determining novel 'methods of improving' conditions requires scientific creativity and contextual understanding beyond current AI capabilities.
Task automatability	claude-sonnet-5	2/5	This task combines field data collection, experimental design, ecological judgment, and site-specific research that AI cannot perform end-to-end; AI can assist with literature review and data analysis but not the full research cycle.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Research credibility, peer review requirements, and institutional accountability create strong friction against full automation. Grant funding, scientific ethics boards, and regulatory compliance tied to human researcher responsibility mean that a qualified biologist must design and validate environmental research methodologically.
Adoption barriers	claude-sonnet-5	3/5	Environmental research often feeds into regulatory decisions (e.g., permitting, EPA compliance) requiring credentialed scientists and accountable sign-off, though not always strictly licensed like some professions.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Environmental research requires specialized infrastructure (field equipment, lab work, expert time) and AI tools still need significant human supervision and validation. The loaded cost of a research biologist remains substantially lower than the combined cost of AI systems, integration, and required human oversight for meaningful research output.
Cost vs. human wage	claude-sonnet-5	2/5	Fieldwork, sampling, and expert interpretation still require human labor and equipment costs that AI cannot substantially reduce, though literature review and data processing portions can be cheaper with AI.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	AI tools exist for environmental data analysis and satellite imagery processing, but no deployed product independently performs the full research task of determining methodological improvements to environmental conditions. Such work remains primarily researcher-driven with AI as a narrow analytical tool, not a end-to-end solution.
Technical feasibility today	claude-sonnet-5	2/5	AI tools exist for data analysis, remote sensing interpretation, and literature synthesis, but no deployed product autonomously conducts environmental impact research and determines improvement methods reliably.

Study aquatic plants and animals and environmental conditions affecting them, such as radioactivity or pollution.

CI 25–30 · exposure 25 · augmentation 63 · importance 3.8/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Adoption is slow; regulatory and compliance-heavy sectors (government agencies, environmental consulting firms) remain cautious, and field biology work is geographically dispersed and relationship-dependent, limiting rapid AI integration.
Sector adoption velocity	claude-sonnet-5	2/5	Environmental and biological field sciences are slower to adopt AI at scale compared to information-heavy sectors, though AI-assisted data analysis and imaging classification are gradually spreading.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI provides useful assistance on data analysis, species identification from images, and literature synthesis, but the human biologist remains essential for field design, sampling decisions, and environmental interpretation.
Augmentation potential	claude-sonnet-5	4/5	AI substantially aids literature review, data analysis, species identification from images, and modeling of pollutant impact, meaningfully boosting researcher productivity while humans still perform field and experimental work.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can assist with data analysis, image classification, and literature review on aquatic organisms and pollution metrics, the core task requires direct fieldwork (sampling, specimen collection, in-situ observation) and complex environmental interpretation that current systems cannot perform end-to-end with 50% time savings at equal quality.
Task automatability	claude-sonnet-5	2/5	Fieldwork such as sample collection, specimen observation, and in-situ measurement of aquatic organisms and conditions requires physical presence and manual skill that current AI cannot perform; only data analysis and literature portions are automatable.'
Adoption barriers	claude-haiku-4-5-20251001	4/5	Strong barriers exist: regulatory requirements for environmental monitoring (Clean Water Act, EPA compliance) often mandate certified human professionals; liability for pollution assessment and aquatic species assessments falls on the credentialed biologist; and direct environmental sampling requires in-person expert judgment.
Adoption barriers	claude-sonnet-5	3/5	No formal licensing requires a human specifically for this research, but safety protocols, ethical field research standards, and physical access to sites impose real logistical barriers to automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI tools for data processing are relatively cheap, but the essential fieldwork, specimen handling, and expert interpretation still require skilled human biologists whose loaded wage far exceeds current AI inference costs for the narrow automation possible.
Cost vs. human wage	claude-sonnet-5	2/5	AI can cheaply process data and literature, but the dominant cost of fieldwork, sample handling, and specialized equipment remains human-driven, keeping overall cost savings modest.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Deployed products exist for specific subtasks (water quality sensor data logging, automated species identification from images) but lack reliable end-to-end capability for the full scope of environmental study, field diagnosis, and causal analysis required in practice.
Technical feasibility today	claude-sonnet-5	2/5	Deployed products can analyze water quality datasets or classify species from images, but no integrated system autonomously conducts aquatic field studies or interprets radioactivity/pollution effects reliably in production.

Prepare plans for management of renewable resources.

CI 25–30 · exposure 25 · augmentation 63 · importance 3.8/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Renewable resource management remains concentrated in government agencies, nonprofits, and specialized firms with slow digital transformation. Pilot adoption of AI tools exists, but production deployment of autonomous planning is rare; sectors are risk-averse and process-heavy.
Sector adoption velocity	claude-sonnet-5	2/5	Environmental and natural resource management is a sector with relatively slow AI adoption; agencies and consultancies use AI mainly for auxiliary data tasks rather than plan authorship.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can meaningfully assist by processing ecological datasets, generating literature summaries, scenario modeling, and drafting plan sections. A biologist's productivity on analysis and documentation can improve, though final judgment and synthesis remain human-driven.
Augmentation potential	claude-sonnet-5	4/5	AI can meaningfully assist by processing large ecological datasets, drafting report sections, summarizing literature, and running predictive models, significantly speeding up plan preparation while the biologist retains final judgment.
Task automatability	claude-haiku-4-5-20251001	2/5	This task requires synthesizing complex ecological data, stakeholder input, and regulatory constraints into a coherent management strategy. While AI can help analyze data and generate draft components, the integration of expert judgment, site-specific knowledge, and long-term strategic reasoning means end-to-end automation with 50% time savings is not yet demonstrated by current systems.
Task automatability	claude-sonnet-5	2/5	Requires synthesis of ecological data, field knowledge, stakeholder considerations, and regulatory context that current AI cannot reliably integrate end-to-end; AI can assist with drafting and data analysis but not autonomously produce a defensible management plan.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Resource management plans often require sign-off by licensed biologists or environmental professionals, regulatory approval, and liability accountability. Many jurisdictions mandate human professional judgment and legal responsibility, creating hard barriers to full substitution.
Adoption barriers	claude-sonnet-5	3/5	While not always requiring a specific license, resource management plans often need agency sign-off, scientific credibility, and accountability under environmental regulations, creating moderate institutional and liability barriers.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI-assisted data analysis and drafting tools cost considerably less per unit output than hiring a biologist, but the task still requires senior professional oversight, validation, and revision. The loaded cost of a domain expert remains the dominant expense in practice.
Cost vs. human wage	claude-sonnet-5	2/5	AI can cheaply handle subtasks like literature synthesis or data visualization, but the overall plan still requires expert human oversight, fieldwork integration, and liability-bearing judgment, keeping all-in costs closer to human-comparable.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	No deployed products reliably perform autonomous renewable resource management planning at production scale. AI tools exist for data analysis and report generation, but the core task—crafting defensible, legally sound, and contextually appropriate management plans—remains largely manual and dependent on human expertise.
Technical feasibility today	claude-sonnet-5	2/5	No deployed product autonomously generates renewable resource management plans; existing tools support data analysis, GIS mapping, or literature review but human biologists still author and validate the plans.

Teach or supervise students and perform research at universities and colleges.

CI 25–25 · exposure 25 · augmentation 63 · importance 3.6/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Universities remain conservative adopters of AI for core teaching and supervision roles; while AI tools are being piloted for grading and content support, actual displacement of teaching and mentoring roles in production is minimal, and sector-wide adoption is slow.
Sector adoption velocity	claude-sonnet-5	2/5	Higher education adopts AI unevenly and cautiously, mostly for administrative or supplementary tasks, with slow institutional change processes limiting deep adoption of AI in teaching/research roles.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI offers useful assistance on parts of this task—generating lecture outlines, summarizing student work, or automating grading—but does not transform overall productivity because the supervisory and interpersonal core remains human-driven and cannot be easily augmented.
Augmentation potential	claude-sonnet-5	4/5	AI substantially aids literature reviews, data analysis, drafting course materials, grading, and generating research ideas, meaningfully boosting productivity while the biologist retains primary responsibility.
Task automatability	claude-haiku-4-5-20251001	2/5	Teaching and supervising students involves complex human interaction, judgment, and mentoring that AI cannot replace end-to-end. While AI can assist with lecture content generation or grading, the core mentoring, real-time classroom interaction, and individualized supervision of students remain heavily human-dependent and far short of 50% time savings at equal quality.
Task automatability	claude-sonnet-5	2/5	Teaching and supervising students plus conducting original research involves mentorship, live instruction, and creative experimental design that current AI cannot autonomously replace end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Strong regulatory and institutional barriers exist: universities legally employ tenured faculty to teach, research, and mentor; accreditation and student outcomes rely on credentialed human instruction; and institutional norms heavily favor human faculty as primary educators and supervisors.
Adoption barriers	claude-sonnet-5	4/5	University teaching and research supervision require accredited, credentialed faculty, tenure processes, and institutional accountability, creating strong regulatory and organizational barriers to full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	The integrated cost of AI-assisted teaching tools (including oversight, customization, and human review) remains comparable to or higher than employing faculty, especially when accounting for the irreducibility of human mentorship and real-time student interaction.
Cost vs. human wage	claude-sonnet-5	2/5	Human faculty combine irreplaceable credentialing, mentorship, and grant-winning research capacity; AI can cut some prep costs but cannot substitute for the whole role, so cost comparison favors humans overall.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	No deployed product reliably performs teaching or student supervision at scale. AI tools exist for content assistance and automated grading, but they are supplementary; no system handles the full pedagogical and supervisory task in production university settings.
Technical feasibility today	claude-sonnet-5	2/5	AI tools exist for lecture drafting, grading assistance, and literature review but no deployed product independently teaches courses or supervises student research reliably in production.

Develop pest management and control measures, and conduct risk assessments related to pest exclusion, using scientific methods.

CI 25–25 · exposure 25 · augmentation 75 · importance 3.3/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Agriculture and pest control remain moderately digitized sectors with slower AI adoption; while some companies pilot AI pest identification, production deployment of autonomous risk assessment and control strategy development remains limited and adoption has lagged information/finance sectors.
Sector adoption velocity	claude-sonnet-5	2/5	Biology and agricultural/environmental sectors are historically slower to adopt AI compared to information or finance sectors, with pilots more common than production deployment for this specific task.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI systems demonstrate strong augmentation potential: image-based pest identification, rapid literature and toxicology database queries, statistical risk modeling, and scenario simulation all meaningfully assist biologists in faster, more comprehensive assessment and strategy design while keeping human expertise central.
Augmentation potential	claude-sonnet-5	4/5	AI can meaningfully augment biologists by rapidly analyzing pest occurrence data, modeling spread scenarios, summarizing research, and drafting portions of assessments, improving efficiency while the scientist retains decision authority.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can assist with literature synthesis, data analysis, and risk modeling for pest management, the core task requires designing novel control strategies, conducting field-specific assessments, and integrating complex ecological knowledge that demands expert judgment—current systems cannot reliably handle the full end-to-end task at 50% time savings with equal quality.
Task automatability	claude-sonnet-5	2/5	AI can support literature review, data analysis, and drafting risk assessment reports, but designing pest control measures requires field expertise, site-specific judgment, and physical verification that AI cannot perform end-to-end.a
Adoption barriers	claude-haiku-4-5-20251001	4/5	Regulatory frameworks often require licensed scientists to conduct pest risk assessments and sign off on management protocols; liability for control measure failures and crop/environmental damage creates strong legal and organizational barriers to autonomous AI decision-making.
Adoption barriers	claude-sonnet-5	4/5	Regulatory frameworks (e.g., biosecurity, agricultural policy) often require certified specialists to sign off on pest risk assessments and control measures, creating strong professional and legal barriers to full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI inference for pest identification and data processing is cheap, but the expert biologist still handles strategy design, field validation, and regulatory sign-off; the all-in cost of AI integration plus required expert oversight remains comparable to direct human labor.
Cost vs. human wage	claude-sonnet-5	2/5	AI can cheaply handle data synthesis and literature scanning, but the core scientific judgment, fieldwork, and risk assessment still require expensive expert labor, keeping overall cost savings modest.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	AI tools exist for pest identification via image analysis and some risk assessment frameworks, but no deployed product reliably performs integrated pest management strategy development and field risk assessment autonomously; applications remain research-stage or narrowly scoped.
Technical feasibility today	claude-sonnet-5	2/5	Some AI tools assist with species identification, data analysis, and modeling pest spread, but no deployed product independently develops full pest management plans or conducts risk assessments reliably in production.

Communicate test results to state and federal representatives and general public.

CI 25–25 · exposure 25 · augmentation 63 · importance 3.3/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Scientific institutions and government agencies have been slow to adopt AI for official communications due to liability concerns, regulatory conservatism, and the need for human expert accountability in statements to policymakers.
Sector adoption velocity	claude-sonnet-5	2/5	Environmental and biological sciences sectors are slower to adopt AI-driven external communications due to regulatory sensitivity and accuracy requirements, with pilots more common than production use.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can assist by drafting initial summaries of test results and generating multiple messaging variants for different audiences, which a biologist can then refine and tailor, improving productivity on the communication component of the work.
Augmentation potential	claude-sonnet-5	4/5	AI can meaningfully help draft, summarize, and tailor communications of technical results for different audiences, significantly speeding up report preparation even though humans must verify and finalize content.
Task automatability	claude-haiku-4-5-20251001	2/5	AI can draft communications and summarize test results, but cannot independently decide what findings to emphasize, navigate political sensitivities, or determine appropriate messaging for different audiences—tasks requiring human judgment and accountability.
Task automatability	claude-sonnet-5	2/5	Drafting communications can be AI-assisted, but synthesizing test results accurately for diverse audiences (regulators, public) and ensuring scientific accuracy and accountability requires human judgment; full end-to-end automation with equal quality is not yet reliable.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Federal and state communication, especially to representatives, often involves legal review, compliance with agency protocols, and implicit or explicit requirements that a qualified scientist sign off on public statements about test results.
Adoption barriers	claude-sonnet-5	4/5	Communicating official test results to regulators and the public often requires credentialed scientific authority, agency approval, and accountability for accuracy, creating strong institutional and legal barriers to full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI-generated drafts still require substantial biologist time for review, editing, and approval; the cost savings from draft generation are modest compared to the loaded wage of a scientist doing this communication work.
Cost vs. human wage	claude-sonnet-5	2/5	While AI drafting reduces some writing time, the need for expert verification, compliance checks, and liability oversight keeps the effective cost comparable to human-led communication for this task.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	While AI can generate text summaries of scientific data, no deployed product reliably handles the full task of communicating complex results to both policymakers and the public while maintaining scientific accuracy and contextual appropriateness.
Technical feasibility today	claude-sonnet-5	2/5	AI writing tools can draft summaries or reports, but no deployed product autonomously communicates scientific test results to government bodies and the public without human review and validation.

Study reactions of plants, animals, and marine species to parasites.

CI 20–30 · exposure 20 · augmentation 63 · importance 2.7/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Academic and government biology labs adopt image analysis and data management tools, but core experimental study of parasite reactions remains labor-intensive and human-led; adoption of AI replacement is minimal and slow.
Sector adoption velocity	claude-sonnet-5	2/5	Academic and applied biology research adopts AI tools unevenly, mostly for literature review and data analysis, with slow uptake for core experimental science.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can meaningfully assist biologists through automated image classification of parasites, literature synthesis, and statistical analysis of reaction data, moderately raising productivity on observational and analytical components while the biologist retains experimental design and interpretation.
Augmentation potential	claude-sonnet-5	4/5	AI can strongly assist with literature synthesis, statistical analysis, image/data pattern recognition, and drafting reports, meaningfully boosting researcher productivity while the scientist remains in charge of experiments.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can analyze images of parasites and extract data from literature, the core task of observing, interpreting, and drawing causal inferences about organism reactions to parasites requires live experimental design, field observation, and nuanced biological judgment that current AI cannot perform end-to-end at the required quality.
Task automatability	claude-sonnet-5	2/5	This involves designing experiments, field/lab observation, and interpreting biological interactions, which requires physical manipulation, novel hypothesis generation, and judgment that current AI cannot perform end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Institutional review requirements (animal ethics boards, field permits), liability for experimental design integrity, and professional credentialing create legal and organizational friction that prevent unilateral AI automation of biological studies.
Adoption barriers	claude-sonnet-5	3/5	No licensing requirement blocks AI use, but institutional review, animal welfare protocols, and scientific credibility norms create moderate friction against full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI tools for microscopy and literature review add cost and overhead compared to their benefit; the irreducible need for field work, live organisms, and expert interpretation means total cost per output remains comparable to or higher than a trained biologist.
Cost vs. human wage	claude-sonnet-5	2/5	Human researchers still must conduct fieldwork, husbandry, and physical experimentation; AI only reduces some analysis/writing time, so overall cost savings are modest.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Deployed products can assist with image analysis and data mining, but no production system reliably performs the full task of studying organism-parasite reactions independently; live specimen observation, hypothesis formation, and interpretation remain human-dependent in practice.
Technical feasibility today	claude-sonnet-5	1/5	No deployed AI product independently conducts host-parasite reaction studies; AI use is confined to research-stage data analysis support, not the full task.

Develop methods and apparatus for securing representative plant, animal, aquatic, or soil samples.

CI 7–30 · exposure 13 · augmentation 50 · importance 3.3/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Life sciences and field ecology remain relatively low-digital sectors; adoption of sampling automation is limited to research labs with specialized equipment and has not penetrated standard practice broadly.
Sector adoption velocity	claude-sonnet-5	2/5	Environmental and field biology sectors are slower adopters of AI for physical apparatus design, with most AI use concentrated in data analysis rather than method/apparatus development.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can assist biologists by recommending optimal sampling strategies, automating data logging from sensors, and identifying contamination patterns, but the human expert must remain in the loop for field execution and quality assurance.
Augmentation potential	claude-sonnet-5	3/5	AI can assist with literature review, design ideation, simulation, and data analysis to support method development, though the physical design and testing remain human-led.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can assist with experimental design and sample protocol optimization, the physical act of securing representative samples in the field requires hands-on fieldwork, environmental judgment, and real-time adaptation to conditions—tasks that current embodied AI cannot reliably perform autonomously.
Task automatability	claude-sonnet-5	1/5	Designing physical sampling apparatus and field methods for representative plant, animal, aquatic, or soil samples requires hands-on engineering, field testing, and domain judgment that current AI cannot execute end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Field sampling often requires licensed scientists and regulatory permits (e.g., environmental permits, species protection laws); professional accountability for sample integrity and chain-of-custody creates legal and organizational barriers to full automation.
Adoption barriers	claude-sonnet-5	3/5	No licensing requirement specifically bars AI, but the task demands physical fabrication, field validation, and scientific judgment that create substantial practical barriers to automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI assistance in sampling design and documentation may reduce overhead, but the core labor cost remains human fieldwork and specimen handling; automation does not substantially undercut the loaded cost of trained biologists performing collection.
Cost vs. human wage	claude-sonnet-5	1/5	AI cannot substitute for the physical prototyping, testing, and fieldwork involved, so there is no meaningful cost offset versus human labor.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Some AI tools exist for planning sampling strategies and data analysis, but no deployed autonomous systems reliably execute representative sample collection across diverse environmental contexts without human supervision and manual intervention.
Technical feasibility today	claude-sonnet-5	1/5	No deployed AI product designs or fabricates physical sampling equipment or field protocols; this remains a research-stage, human-driven activity.

Plan and administer biological research programs for government, research firms, medical industries, or manufacturing firms.

CI 7–25 · exposure 13 · augmentation 50 · importance 3.0/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Academic and government research institutions are digitizing workflows slowly; adoption of AI for research administration remains in pilot phase at most universities and labs. Conservative institutional culture and regulatory caution limit rapid deployment of automation in this domain.
Sector adoption velocity	claude-sonnet-5	2/5	Research institutions and government labs adopt AI for specific analytical tasks but administrative/managerial functions remain largely untouched by automation.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can usefully assist with data synthesis, grant opportunity identification, compliance tracking, and scheduling, raising a program director's productivity. However, these aids remain partial; strategic vision, stakeholder negotiation, and ethical judgment still centrally depend on human expertise.
Augmentation potential	claude-sonnet-5	3/5	AI can assist with literature synthesis, grant writing drafts, data analysis, and scheduling, providing meaningful support to biologists managing research programs.
Task automatability	claude-haiku-4-5-20251001	2/5	Planning and administering biological research programs requires strategic judgment, stakeholder alignment, ethical oversight, and adaptive management of complex projects. While AI can assist with literature reviews, data analysis, and report drafting, the core governance, decision-making, and human oversight functions demand sustained human leadership that current systems cannot fully replace end-to-end.
Task automatability	claude-sonnet-5	1/5	Planning and administering research programs requires strategic judgment, resource allocation, stakeholder management, and adaptive decision-making that current AI cannot perform end-to-end.'
Adoption barriers	claude-haiku-4-5-20251001	4/5	Regulatory bodies (NIH, FDA, institutional review boards) require a named human research administrator accountable for compliance, ethical oversight, and institutional risk. Many jurisdictions legally mandate human sign-off on research protocols and budgets, creating strong legal and liability barriers to substitution.
Adoption barriers	claude-sonnet-5	4/5	Program administration often involves regulatory compliance, funding accountability, institutional signoffs, and legal responsibility that require credentialed human leadership.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI tools cost less than senior researcher time for individual tasks like data processing, but comprehensive program administration demands a senior scientist's judgment and accountability. The all-in cost of AI oversight and integration approaches the salary of the researcher being assisted, making cost savings marginal.
Cost vs. human wage	claude-sonnet-5	1/5	There is no viable AI substitute for program management, so cost comparison favors humans entirely; any attempt would require extensive human oversight negating savings.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	No deployed AI product reliably performs end-to-end research program administration today. Tools exist for components (project scheduling, literature mining, compliance checking) but integrating these into actual program governance and institutional decision-making remains an open problem requiring human oversight.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product manages entire biological research programs; AI tools assist with literature review or data analysis but not program administration.

Study and manage wild animal populations.

CI 5–21 · exposure 5 · augmentation 50 · importance 3.8/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Conservation organizations and wildlife agencies are slow adopters; most work remains field-based with limited digitization. Uptake of AI-assisted analysis tools is gradual, but replacement of on-the-ground biologists is not occurring; sectors are geographically dispersed and regulatory-constrained.
Sector adoption velocity	claude-sonnet-5	1/5	Wildlife biology and conservation are low-digitization, field-based sectors with minimal AI agent deployment in production compared to information or finance sectors.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI provides meaningful assistance through automated image classification from camera traps, population modeling, and spatial analysis of movement data, improving biologists' productivity and decision-making. However, augmentation is limited to data processing; field presence and expert judgment remain wholly human-driven.
Augmentation potential	claude-sonnet-5	3/5	AI assists with data analysis, population modeling, image/audio species identification from camera traps and acoustic sensors, and predictive analytics, meaningfully boosting productivity on the analytical components.
Task automatability	claude-haiku-4-5-20251001	1/5	This task requires direct field observation, live animal handling, population tracking, and adaptive decision-making in highly variable natural environments. Current AI cannot autonomously perform the core activities of field surveying, capture, health assessment, and in-situ management decisions that define the work.
Task automatability	claude-sonnet-5	1/5	This involves fieldwork, animal capture/tagging, habitat assessment, and adaptive management decisions in physical environments that current AI cannot perform end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Legal and regulatory barriers are substantial: protected species management typically requires licensed wildlife biologists, permits, and documented expert oversight. Liability for failed management of endangered populations or ecosystem harm creates strong legal requirements for credentialed human decision-makers.
Adoption barriers	claude-sonnet-5	3/5	While no strict licensing mandates a biologist perform every aspect, wildlife management often requires permits, government oversight, and species-specific expertise that create moderate institutional friction against automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI tools for remote monitoring and analysis are relatively cheap, but the core task—physical field presence, capture, veterinary assessment, and adaptive management—requires human expertise whose loaded cost far exceeds any AI system. Full substitution is not economically viable.
Cost vs. human wage	claude-sonnet-5	1/5	The physical fieldwork, equipment, travel, and hands-on monitoring required cannot be replaced by AI inference, so cost comparison heavily favors human labor for the core task.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	While AI-assisted tools (camera traps with automated species detection, population modeling software) exist in production, no system autonomously manages wild populations end-to-end. Deployed tools assist rather than execute; human biologists remain essential for all field operations and management decisions.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product manages wild animal populations autonomously; AI tools exist only for narrow subtasks like camera-trap image classification or population modeling assistance.

Supervise biological technicians and technologists and other scientists.

CI 5–7 · exposure 0 · augmentation 38 · 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	Supervision is fundamentally a human-relations function resistant to automation despite high digitization in life sciences. Organizations continue to rely on human managers even as other functions automate, indicating low velocity of substitution for supervisory roles.
Sector adoption velocity	claude-sonnet-5	2/5	Life sciences R&D settings show moderate AI adoption for technical/analytical tasks, but management and supervisory functions see little to no automation deployment.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI can provide data analytics on technician productivity, literature summaries, or scheduling assistance, but these are narrow support functions. The core supervisory relationship—setting priorities, mentoring, evaluating performance—remains fundamentally human-centered.
Augmentation potential	claude-sonnet-5	3/5	AI tools can help with scheduling, tracking technician performance data, drafting evaluations, or summarizing lab output, aiding but not transforming the supervisory role itself.
Task automatability	claude-haiku-4-5-20251001	1/5	Supervision requires real-time judgment about personnel performance, conflict resolution, career development, and dynamic team coordination. Current AI systems cannot reliably handle the interpersonal nuance and context-dependent decision-making required to oversee human teams at 50% time savings with equal quality.
Task automatability	claude-sonnet-5	1/5	Supervision involves interpersonal leadership, performance evaluation, mentoring, and real-time judgment calls that AI cannot execute end-to-end; no off-the-shelf system replaces a human supervisor of staff.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Legal and organizational liability for employment decisions creates significant barriers; supervisory authority typically requires human accountability. Employment law, union agreements where applicable, and institutional policy generally require human decision-makers in personnel matters.
Adoption barriers	claude-sonnet-5	4/5	Supervisory responsibility often carries organizational accountability, personnel management duties, and sometimes regulatory/safety sign-off requirements that necessitate a qualified human in the role.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	Effective human supervisors with domain expertise command substantial salaries, and any attempt to automate supervision would require expensive oversight infrastructure plus ongoing human review, making the total cost exceed that of direct human supervision.
Cost vs. human wage	claude-sonnet-5	1/5	There is no viable AI substitute performing this task, so cost comparison favors the human by default; any AI role is only a minor cost-saving add-on, not a replacement.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed AI system reliably performs supervisory functions including performance reviews, disciplinary decisions, and team motivation. While AI can assist with scheduling or documentation, autonomous supervision of biological teams is research-stage only.
Technical feasibility today	claude-sonnet-5	1/5	There are no deployed products that autonomously supervise laboratory personnel; this remains a fundamentally human management function.

Represent employer in a technical capacity at conferences.

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	There is no adoption of AI for autonomous conference representation in any sector. This task is inherently tied to human professional presence and relationship-building, so velocity remains near zero.
Sector adoption velocity	claude-sonnet-5	2/5	While biologists work in increasingly digitized research settings, the specific act of in-person representation at conferences shows minimal AI adoption or displacement to date.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI can modestly assist by drafting presentation slides, preparing talking points, or analyzing conference data, but these are peripheral to the core task of personal representation and networking, which requires the human to remain the active agent.
Augmentation potential	claude-sonnet-5	3/5	AI can help prepare presentation materials, summarize research, generate talking points, or draft follow-up communications, meaningfully aiding preparation even though it cannot perform the representational task itself.
Task automatability	claude-haiku-4-5-20251001	1/5	This task requires human presence, social interaction, and real-time networking at conferences—core elements that cannot be meaningfully automated. Current AI systems cannot independently represent an organization, build professional relationships, or navigate dynamic interpersonal situations.
Task automatability	claude-sonnet-5	1/5	Representing an employer at conferences requires physical presence, real-time networking, spontaneous technical discussion, and organizational representation that current AI cannot perform end-to-end.'
Adoption barriers	claude-haiku-4-5-20251001	5/5	Legal and organizational barriers are substantial: employers have fiduciary and reputational responsibility for who represents them; professional norms and client expectations strongly prefer direct human contact; liability and trust considerations make substitution infeasible regardless of technical capability.
Adoption barriers	claude-sonnet-5	4/5	Representing an employer carries reputational, relational, and sometimes contractual/legal weight (e.g., speaking authority, negotiation), and organizations require a trusted, credentialed human presence.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	An AI system attempting to replace human conference representation would require expensive custom development, continuous oversight, and likely still fail to deliver comparable value; the cost would far exceed hiring a biologist to attend.
Cost vs. human wage	claude-sonnet-5	1/5	There is no AI product performing this role, so cost comparison favors the human by default; AI cannot substitute at any price point today.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed AI system can autonomously attend and represent an employer at conferences. While AI can assist with presentation materials or drafting talking points, the actual representation task—which is inherently human-centric—remains outside the scope of production-ready automation.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product substitutes for a human representative attending and engaging at a conference on an employer's behalf; this remains squarely human territory.

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.