Soil and Plant Scientists

19-1013.00

Median wage $78,850/yr15,730 employed (US)Rank #490 of 923 scored · top 53% by substitution

Conduct research in breeding, physiology, production, yield, and management of crops and agricultural plants or trees, shrubs, and nursery stock, their growth in soils, and control of pests; or study the chemical, physical, biological, and mineralogical composition of soils as they relate to plant or crop growth. May classify and map soils and investigate effects of alternative practices on soil and crop productivity.

Sub-scores

0–100 · band = confidence interval from rater disagreement

Substitution28

Exposure24

Augmentation66

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

27 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%25

panel mean rating 2.0/5 → substitution pressure 25/100

Technical feasibility todayw 20%22

panel mean rating 1.9/5 → substitution pressure 22/100

Cost vs. human wagew 15%28

panel mean rating 2.1/5 → substitution pressure 28/100

Adoption barriersw 20%inverted — strong barriers lower the score39

panel mean rating 3.5/5 (barrier strength) → substitution pressure 39/100

Sector adoption velocityw 10%25

panel mean rating 2.0/5 → substitution pressure 25/100

Task breakdown (27 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.

Provide information or recommendations to farmers or other landowners regarding ways in which they can best use land, promote plant growth, or avoid or correct problems such as erosion.

CI 34–87 · 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	Precision agriculture and data-driven farm advisory are rapidly deployed in developed agricultural sectors (commodity crops, large operations), with companies like Bayer, John Deere, and regional advisors integrating AI into operations; adoption is fastest in digitized, capital-intensive farming regions.
Sector adoption velocity	claude-sonnet-5	2/5	Agriculture is a historically slow-adopting, physically-grounded sector with uneven digitization, so AI tool uptake among soil/plant scientists and farmers remains in early pilot stages.
Augmentation potential	claude-haiku-4-5-20251001	5/5	AI-powered soil and crop monitoring tools, satellite imagery analysis, and decision-support dashboards substantially augment agronomists' productivity by automating data collection and preliminary assessment, freeing experts to focus on complex site-specific decisions and farmer consultation.
Augmentation potential	claude-sonnet-5	4/5	AI can meaningfully assist by synthesizing research literature, drafting reports, analyzing soil data trends, and suggesting management options, boosting scientist productivity while they retain final judgment.
Task automatability	claude-haiku-4-5-20251001	5/5	AI systems can analyze soil data, weather patterns, crop health imagery, and historical yield records to generate personalized land-use recommendations and erosion prevention strategies at scale, meeting the ≥50% time-saving threshold for information synthesis and preliminary advisory generation.
Task automatability	claude-sonnet-5	2/5	AI can generate generic recommendations from text descriptions, but the task requires site-specific assessment, soil sampling interpretation, and field judgment that current systems cannot fully replicate end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	2/5	No licensing requirement mandates that a human scientist must deliver agricultural advice; liability concerns are moderate and manageable through disclaimers; adoption is largely voluntary and driven by economics and farmer preference rather than legal gatekeeping.
Adoption barriers	claude-sonnet-5	3/5	No formal licensing typically required, but liability for erosion/crop failure recommendations and farmer trust in expert judgment create meaningful friction against pure AI substitution.
Cost vs. human wage	claude-haiku-4-5-20251001	5/5	AI-driven advisory tools cost pennies to dollars per recommendation once deployed, whereas hiring a soil scientist or agronomist for farm visits and consultation costs hundreds to thousands per engagement, making AI solutions one to two orders of magnitude cheaper.
Cost vs. human wage	claude-sonnet-5	3/5	AI-generated general advice is cheap to produce, but combined with necessary site visits, soil testing, and expert validation, overall cost savings versus a human scientist are moderate, not dramatic.
Technical feasibility today	claude-haiku-4-5-20251001	4/5	Multiple deployed platforms (precision agriculture tools, satellite-based crop monitoring, soil analysis APIs) reliably perform aspects of this task in production, though end-to-end autonomous advisory still requires human agronomist review in many jurisdictions; systems exist and are used operationally but not universally autonomous.
Technical feasibility today	claude-sonnet-5	2/5	Some ag-tech advisory chatbots and decision-support tools exist but are narrow in scope and generally supplement rather than replace scientist-provided recommendations in production settings.

Identify or classify species of insects or allied forms, such as mites or spiders.

CI 30–56 · exposure 42 · augmentation 75 · 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	Adoption remains slow in traditional soil and plant science organizations, which are often smaller, less digitized, and deeply dependent on taxonomic expertise; while research labs pilot AI tools, production-level displacement is minimal and sector-wide adoption lags information-intensive industries.
Sector adoption velocity	claude-sonnet-5	2/5	Agricultural and scientific research sectors adopt digital tools at a moderate pace, but formal taxonomic classification work remains largely manual and specialized, with slow institutional uptake of AI tools.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI image recognition significantly assists human taxonomists by rapidly narrowing candidate species, sorting specimens, and flagging unusual morphologies, enabling faster workflows and reducing eye strain; the human expert remains essential for confirmation but productivity gains are substantial.
Augmentation potential	claude-sonnet-5	4/5	AI-assisted identification tools significantly speed up preliminary species sorting and narrowing possibilities, letting scientists focus expert judgment on confirming difficult cases.
Task automatability	claude-haiku-4-5-20251001	2/5	AI vision systems can assist with insect identification via image analysis, but field conditions, specimen damage, and the need to correctly differentiate closely related species often require human expertise; current systems lack the contextual judgment for consistent 50% time-saving at equal quality across diverse insects and mites.
Task automatability	claude-sonnet-5	3/5	AI image classifiers can identify many common insect and arachnid species from photos with reasonable accuracy, but expert-level taxonomic classification of ambiguous or rare specimens still requires human verification and physical examination.tissue-level features.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Regulatory and quality-assurance contexts (pest monitoring, disease vector identification, ecological surveys) often require a licensed or certified taxonomist to sign off on species identifications; liability and accuracy requirements create strong organizational and legal friction against full automation.
Adoption barriers	claude-sonnet-5	2/5	No licensing requirement for species identification itself, though scientific publications or regulatory submissions may require credentialed verification, creating mild friction.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI identification tools are relatively cheap to deploy, but they typically require human validation, microscopy equipment, and domain expertise to confirm borderline cases, making the all-in cost comparable to or higher than a trained entomologist's labor for critical identifications.
Cost vs. human wage	claude-sonnet-5	4/5	Image-based identification apps are free or very cheap compared to an expert's time for routine identification tasks, though complex cases still need costlier expert review.
Technical feasibility today	claude-haiku-4-5-20251001	3/5	Computer vision and AI-powered identification tools (e.g., iNaturalist, specialized entomology platforms) exist and perform reasonably well on common species, but error rates remain material for rare or cryptic taxa, and integration with laboratory workflows is inconsistent.
Technical feasibility today	claude-sonnet-5	3/5	Deployed apps (iNaturalist, Google Lens, specialized entomology tools) perform species identification in production, but accuracy drops for closely related species, juveniles, or damaged specimens, requiring expert confirmation.

Communicate research or project results to other professionals or the public or teach related courses, seminars, or workshops.

CI 25–39 · exposure 30 · 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	Academic and research institutions have adopted AI writing tools incrementally for preliminary drafting and administrative communication, but core teaching and research dissemination remain primarily human-driven; adoption in this sector lags information-services industries.
Sector adoption velocity	claude-sonnet-5	2/5	Agricultural and soil science domains have historically slower AI adoption than digital-native, information-heavy sectors; most AI use here is in data analysis rather than public communication tasks.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI substantially aids research scientists by drafting manuscript sections, generating figures/summaries, organizing data narratives, and creating course syllabi, allowing scientists to focus on content accuracy, pedagogy, and live interaction with audiences.
Augmentation potential	claude-sonnet-5	4/5	AI tools substantially help scientists draft reports, prepare slides, generate visualizations, and organize course materials, meaningfully boosting productivity while the human still delivers and owns the communication.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can draft presentations, write research summaries, and generate explanatory content, delivering effective scientific communication—especially live teaching or seminars—requires contextual judgment, responsiveness to audience questions, and credibility signaling that current AI cannot fully replicate at the quality level expected of soil and plant scientists.
Task automatability	claude-sonnet-5	2/5	AI can help draft written reports or slides summarizing research, but live teaching, presenting to audiences, and answering nuanced questions requires human presence and expertise that current AI cannot fully replace end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Scientists retain gatekeeping authority over research dissemination and course instruction; institutional credentialing, peer review standards, and expectations for human expertise create strong organizational and professional norms against full AI substitution for communicating novel research findings.
Adoption barriers	claude-sonnet-5	3/5	No formal licensing requirement to communicate research, but credibility, subject-matter authority, and audience trust in a human expert create moderate friction against AI substitution.
Cost vs. human wage	claude-haiku-4-5-20251001	3/5	AI can reduce time spent on drafting communications and preparing materials, but a professional scientist's communication still requires significant human input for verification, customization, and delivery; overall cost savings are meaningful but not transformative.
Cost vs. human wage	claude-sonnet-5	2/5	Drafting support is cheap, but the actual delivery (teaching, presenting, engaging with audiences/media) still requires a human expert, so overall cost savings are limited.
Technical feasibility today	claude-haiku-4-5-20251001	3/5	Products exist for content generation (writing abstracts, creating slides, drafting course materials), but deployed systems rarely handle the full end-to-end teaching or live presentation task reliably; human oversight and substantive revision remain necessary for accuracy and pedagogical effectiveness.
Technical feasibility today	claude-sonnet-5	2/5	Products like AI writing assistants and presentation tools help draft content, but no deployed product independently delivers scientific communication, teaches courses, or represents an organization publicly at scale.

Investigate responses of soils to specific management practices to determine the use capabilities of soils and the effects of alternative practices on soil productivity.

CI 25–35 · exposure 25 · augmentation 75 · 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 of AI in soil science remains limited to research institutions and large agricultural enterprises; most soil management decisions are made by smaller farms and regional extension services with lower digital maturity and slower technology uptake.
Sector adoption velocity	claude-sonnet-5	2/5	Agricultural and soil science remain a low-to-moderate digitization sector with slow uptake of AI in field research relative to information-sector professions.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI can significantly assist soil scientists by automating data processing, generating predictive models of soil responses, synthesizing research findings, and flagging anomalies in field observations—augmenting the scientist's ability to interpret complex soil–management relationships while they retain investigative and decision authority.
Augmentation potential	claude-sonnet-5	4/5	AI can meaningfully assist with statistical modeling, literature synthesis, remote sensing data interpretation, and drafting reports, significantly boosting researcher productivity while humans retain fieldwork and judgment roles.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can assist in analyzing soil data, modeling soil responses to practices, and synthesizing literature, the core work requires field investigation, experimental design, and interpretive judgment that depends on tacit knowledge of local conditions. Current systems cannot reliably conduct end-to-end investigations at the scale and quality a soil scientist provides.
Task automatability	claude-sonnet-5	2/5	This requires designing field experiments, collecting physical soil samples, and interpreting results in context—AI can assist analysis but cannot perform the experimental fieldwork or judgment-based interpretation end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Regulatory frameworks in agriculture and environmental science often require licensed or credentialed professionals to design and certify soil studies, and liability for soil management recommendations falls on the expert. Customers and regulatory bodies typically require human expertise and sign-off, creating high adoption barriers.
Adoption barriers	claude-sonnet-5	2/5	No licensing requirement strictly mandates a human perform this, but scientific credibility, funding/grant requirements, and organizational trust in novel findings create moderate friction against pure AI substitution.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI-assisted analysis can reduce some analytical overhead, but the task's cost structure is dominated by field work, sample collection, and laboratory analysis, which remain labor-intensive and cannot be fully automated. All-in AI costs would not undercut the loaded wage of a skilled soil scientist.
Cost vs. human wage	claude-sonnet-5	2/5	Fieldwork, sampling, and lab analysis costs dominate and are not reduced by AI; only the data-analysis/reporting portion is cheaper via AI, so overall cost savings are limited.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	AI tools exist for soil data analysis, predictive modeling, and literature review, but no deployed product reliably performs the full investigative task—field sampling, experimental setup, field observation, and context-specific interpretation—without substantial human oversight and decision-making.
Technical feasibility today	claude-sonnet-5	2/5	No deployed product autonomously investigates soil-management responses; existing tools (data loggers, ag-analytics platforms) provide supporting data but the investigative task itself remains human-led with research-stage AI assistance.

Conduct research to determine best methods of planting, spraying, cultivating, harvesting, storing, processing, or transporting horticultural products.

CI 25–35 · exposure 25 · augmentation 63 · importance 3.5/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Research institutions and agricultural science remain relatively conservative; AI adoption in horticultural research is mostly pilots and supplementary tools (data analysis, literature mining) rather than autonomous research execution. Small-to-medium farms and traditional research centers adopt slowly.
Sector adoption velocity	claude-sonnet-5	2/5	Agricultural science sectors show moderate but slower AI adoption relative to information/finance industries, given the physical and field-based nature of the work and reliance on long trial cycles.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI tools demonstrably assist plant scientists in analyzing experimental data, processing imagery from field trials, predicting outcomes, and literature synthesis, but human judgment on experimental design, interpretation, and method validation remains central.
Augmentation potential	claude-sonnet-5	4/5	AI significantly aids research design, statistical analysis, predictive modeling, and literature synthesis, meaningfully speeding up parts of the research workflow while humans direct and validate the science.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can assist with literature review, data analysis, and modeling of horticultural variables, the task fundamentally requires field-based empirical research, experimentation, and judgment about complex biological systems. Current AI cannot independently design, execute, and validate agricultural experiments at scale or quality comparable to human plant scientists.
Task automatability	claude-sonnet-5	2/5	This requires designing experiments, field trials, and iterative empirical testing under variable environmental conditions, which AI cannot fully execute end-to-end today; AI can assist with data analysis and literature review but not the physical research process.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Research institutions and regulatory bodies (USDA, EPA) typically require human scientists to design, conduct, and certify findings on agricultural methods. Liability and validation standards for food production methods create strong institutional and legal requirements for human oversight and certification.
Adoption barriers	claude-sonnet-5	2/5	No licensing requirement blocks AI use for research design, but scientific credibility, peer validation, and practical field verification create organizational friction against pure AI-driven conclusions.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	The specialized equipment, field trials, biological materials, and human expertise required for rigorous horticultural research cannot be meaningfully replaced by AI inference costs. Integration and validation overhead would be substantial relative to a PhD-level plant scientist's salary amortized over research output.
Cost vs. human wage	claude-sonnet-5	2/5	AI can reduce costs for data analysis and literature review portions, but the field trials, sample collection, and physical experimentation still require human labor and equipment, keeping overall costs comparable to human-led research.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Deployed AI products exist for crop monitoring and yield prediction, but no mature systems independently conduct horticultural research or design optimal planting/harvesting protocols. Most practical deployment remains in narrow sub-tasks (image analysis) rather than the full research cycle.
Technical feasibility today	claude-sonnet-5	2/5	No deployed product autonomously conducts full agronomic research programs; existing tools assist with data analysis, modeling, or literature synthesis but the core experimental research remains human-led.

Study ways to improve agricultural sustainability, such as the use of new methods of composting.

CI 25–35 · exposure 25 · augmentation 63 · importance 3.4/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Adoption of AI in agricultural research is still primarily pilot-stage (data analytics, crop modeling); most soil and plant science remains empirically grounded in university and government labs with slow institutional technology adoption, especially for core research methodology.
Sector adoption velocity	claude-sonnet-5	2/5	Agricultural science and sustainability research sectors show slower, more uneven AI adoption compared to fast-moving digital/professional service sectors, with pilots more common than production deployment.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI meaningfully assists with literature synthesis, dataset analysis, and modeling scenarios that scientists then test and refine, improving research efficiency; however, it does not transform the core experimental design and hypothesis-testing work that remains central to the discipline.
Augmentation potential	claude-sonnet-5	4/5	AI can meaningfully assist scientists by summarizing literature, analyzing experimental data, suggesting research directions, and drafting reports, significantly boosting research productivity while humans retain scientific judgment.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can analyze composting literature and optimize parameters computationally, the task fundamentally requires empirical field testing, long-term observation of soil outcomes, and novel hypothesis generation grounded in physical experimentation—activities that current AI cannot perform end-to-end without substantial human direction and validation.
Task automatability	claude-sonnet-5	2/5	This is an open-ended research task requiring hypothesis generation, field experimentation, and novel synthesis, which AI can support but not perform end-to-end at the 50%-time-saving-at-equal-quality bar.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Agricultural science publication requires peer review and expertise certification, and novel sustainability claims require regulatory or institutional validation before adoption; these institutional and credibility barriers mean human scientists must remain centrally accountable for the research claims.
Adoption barriers	claude-sonnet-5	2/5	No licensing requirement blocks AI use in research support, though scientific credibility, peer review, and institutional trust in novel findings create moderate friction against full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI inference for literature mining and parameter optimization is cheap, but the bulk cost in this task is human expertise (field trials, experimental design, interpretation), equipment, and time—making the total human cost per validated innovation still substantially lower than outsourcing to AI plus human oversight.
Cost vs. human wage	claude-sonnet-5	2/5	AI can cheaply assist with literature synthesis and data crunching, but the core research (fieldwork, experimentation, expert judgment) still requires costly skilled human scientists, so overall cost savings are modest.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	AI tools can assist with literature review, data analysis, and computational modeling of composting scenarios, but no deployed product performs the full innovation cycle of studying and validating new agricultural sustainability methods independently; production systems remain advisory, not autonomous researchers.
Technical feasibility today	claude-sonnet-5	2/5	AI tools can assist with literature review, data analysis, and drafting hypotheses, but no deployed product independently conducts sustainability research or designs composting experiments reliably.

Conduct experiments to develop new or improved varieties of field crops, focusing on characteristics such as yield, quality, disease resistance, nutritional value, or adaptation to specific soils or climates.

CI 25–30 · 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	Adoption of AI in crop science is growing but remains primarily in the pilot and research phase. Most major seed companies and agricultural research institutions are experimenting with AI for breeding selection and modeling, but autonomous field experimentation is not yet a standard production practice, and adoption remains concentrated in large agribusiness rather than broad-based.
Sector adoption velocity	claude-sonnet-5	2/5	Agricultural R&D is a physically-oriented, moderately-digitized sector where AI adoption is growing (e.g., genomic selection tools) but production-scale autonomous experimentation is not widespread.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI substantially augments crop scientists by accelerating data analysis of field trials, optimizing experimental designs, predicting phenotypes from genotypes, and identifying disease or stress patterns in imagery. These tools meaningfully enhance human productivity while scientists remain responsible for hypothesis development, field decisions, and regulatory sign-off.
Augmentation potential	claude-sonnet-5	4/5	AI significantly aids scientists via predictive modeling, genomic data analysis, experimental design optimization, and literature synthesis, meaningfully boosting research productivity while humans still run experiments.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can assist in data analysis, experimental design, and literature review, conducting field experiments requires physical manipulation, biological observation, sample collection, and adaptive decision-making in real environmental conditions that current AI systems cannot perform end-to-end. The core experimental work remains fundamentally manual and requires on-site expertise.
Task automatability	claude-sonnet-5	2/5	This task involves physical experimentation, field trials, breeding cycles, and hands-on data collection that AI cannot perform end-to-end; AI can assist in design and analysis but not execute the core experimental work.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Regulatory barriers are substantial: crop variety development must meet USDA approval, EPA requirements, and food safety standards; liability for crop failure or unintended effects is significant; and authorization from land-grant institutions or government agencies often governs trial work. Human agronomists and plant scientists must legally conduct and sign off on variety testing.
Adoption barriers	claude-sonnet-5	3/5	No strict licensing requirement for the researcher role itself, but regulatory oversight of new crop varieties (biosafety, seed certification, GMO regulations) and institutional review add friction to full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI tools for data analysis and modeling are cheap, but the bottleneck cost is the field work, infrastructure, and labor of scientists conducting multi-season trials. AI cannot eliminate these costs, so the total cost of developing a crop variety remains dominated by human labor and field operations, making AI's marginal contribution modest relative to overall expenses.
Cost vs. human wage	claude-sonnet-5	2/5	AI tools reduce some analytical costs but the dominant costs (land, labor, multi-year field trials, greenhouse space) remain human/physical infrastructure costs that AI does not replace.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Current AI products can support crop analysis via image recognition and predictive modeling, but no deployed system reliably conducts the full experimental cycle of field crop variety development, which demands living biological work, environmental control, and iterative hypothesis testing over multiple growing seasons. Research prototypes exist but not production-ready autonomous systems for this work.
Technical feasibility today	claude-sonnet-5	2/5	Products exist for genomic analysis, predictive breeding models, and experimental design optimization, but no deployed system independently conducts multi-season crop breeding experiments.

Investigate soil problems or poor water quality to determine sources and effects.

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	Soil and plant science remains embedded in government agencies and smaller consulting firms with low digitization; pilot projects for AI-assisted data analysis exist, but production-scale replacement of investigation workflows is rare and slow across the sector.
Sector adoption velocity	claude-sonnet-5	2/5	Agricultural and environmental science sectors have historically slow AI adoption for field-based diagnostic work, with pilots more common than production deployment.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can assist by analyzing large datasets, generating preliminary hypotheses, organizing monitoring data, and drafting technical reports, but the scientist retains primary responsibility for field strategy, sample interpretation, and causal determination.
Augmentation potential	claude-sonnet-5	4/5	AI tools can strongly assist with data analysis, pattern recognition in soil/water chemistry data, and report drafting, meaningfully boosting scientist productivity even though the human must remain central to fieldwork and judgment.
Task automatability	claude-haiku-4-5-20251001	2/5	Some early-stage components like data analysis, historical pattern matching, and report generation could be partially automated, but the task fundamentally requires field investigation, hands-on sampling, expert interpretation of environmental context, and professional judgment about causation that current AI systems cannot reliably perform end-to-end at sufficient quality or speed.
Task automatability	claude-sonnet-5	2/5	This requires field sampling, on-site investigation, and physical diagnostics that AI cannot perform end-to-end; AI can assist with data analysis but not the full investigative task.:
Adoption barriers	claude-haiku-4-5-20251001	4/5	Environmental investigation results carry high liability and regulatory stakes (EPA, state agencies, remediation decisions); stakeholders require credentials and professional judgment from licensed scientists, and regulatory frameworks expect documented expert attribution rather than algorithmic determination.
Adoption barriers	claude-sonnet-5	3/5	While not universally licensed, environmental and agricultural investigations often require professional certification, regulatory reporting, and liability accountability that create moderate barriers to full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Current AI deployment (data analysis tools, image recognition for sample classification) requires significant human oversight, integration with lab and field workflows, and expert validation; the all-in cost remains comparable to or exceeds the loaded wage of a soil scientist for a typical investigation.
Cost vs. human wage	claude-sonnet-5	2/5	Human field investigation, sampling, and site-specific judgment remain necessary, so AI only reduces some analytical costs while overall cost is still dominated by field labor and expertise.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	While AI tools exist for data analysis and preliminary pattern detection in environmental datasets, no deployed product reliably performs the full investigation workflow—field diagnostics, sample collection decisions, contamination source determination, and multi-factor causal analysis—at production scale with acceptable error rates for environmental liability contexts.
Technical feasibility today	claude-sonnet-5	2/5	No deployed product autonomously investigates soil/water problems in the field; existing tools support lab analysis or data interpretation but not the full diagnostic workflow.

Study soil characteristics to classify soils on the basis of factors such as geographic location, landscape position, or soil properties.

CI 25–30 · exposure 25 · augmentation 63 · importance 3.5/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Soil science remains a relatively conservative, small sector with limited digitization and slow tech adoption outside large agricultural corporations or government agencies. Most soil work still requires field presence and relies on small teams of qualified practitioners rather than scalable digital workflows.
Sector adoption velocity	claude-sonnet-5	2/5	Agriculture and earth sciences are relatively slow adopters of AI compared to information/finance sectors, with AI mainly used in precision agriculture pilots rather than widespread soil classification workflows.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can meaningfully assist soil scientists by automating data integration from spectroscopy, satellite imagery, and lab results, and by flagging potential classification edges for review. These tools raise productivity in data synthesis and preliminary screening, though the final judgment and field-based reasoning remain fundamentally human.
Augmentation potential	claude-sonnet-5	4/5	AI and machine learning models (e.g., satellite imagery analysis, predictive soil mapping) significantly aid soil scientists in identifying patterns and classifying soils faster, though human validation remains essential.
Task automatability	claude-haiku-4-5-20251001	2/5	Soil classification involves synthesizing complex, multimodal field data (geographic, visual, textural, chemical) and applying domain expertise that requires trained judgment. While AI can assist in analyzing lab results or satellite imagery, end-to-end autonomous classification from raw fieldwork with 50% time savings remains limited without human verification of field conditions.
Task automatability	claude-sonnet-5	2/5	Soil classification requires field observation, sample collection, and expert interpretation of physical/chemical properties that current AI cannot independently perform end-to-end; AI can assist with data analysis but not full replacement of fieldwork and judgment.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Soil classification is regulated in many jurisdictions for agricultural, environmental, and engineering purposes, with formal taxonomies (USDA Soil Taxonomy) requiring certified soil scientists. Legal liability for incorrect classification and organizational reliance on credentialed experts create substantial barriers to autonomous AI deployment.
Adoption barriers	claude-sonnet-5	3/5	While not legally licensed in most jurisdictions, soil classification for regulatory, agricultural, or engineering purposes often requires certified professional judgment and physical site presence, creating moderate barriers to full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Current soil classification AI systems require significant setup, domain expertise to interpret outputs, and ongoing human oversight. The total cost (software, infrastructure, expert validation) remains comparable to or higher than hiring soil scientists for most operational scales, especially given liability for misclassification.
Cost vs. human wage	claude-sonnet-5	2/5	AI tools can reduce some data processing costs, but the need for physical sampling, lab testing, and expert validation keeps overall costs comparable to or only modestly cheaper than human-led work.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Some tools exist for soil image analysis and laboratory data interpretation, but deployed systems typically support rather than replace the full classification task. Field verification, spatial reasoning about landscape position, and integration of heterogeneous data sources still depend heavily on expert human judgment in production settings.
Technical feasibility today	claude-sonnet-5	2/5	Some GIS and remote-sensing tools exist to aid soil mapping and classification, but no deployed product reliably performs full soil classification without expert field verification and lab analysis.

Survey undisturbed or disturbed lands for classification, inventory, mapping, environmental impact assessments, environmental protection planning, conservation planning, or reclamation planning.

CI 25–30 · 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	Adoption of remote-sensing and GIS tools is steady but slow in many sectors, particularly in smaller environmental consulting firms and government agencies. Field-based soil science remains largely traditional, with AI integration limited to data processing rather than task displacement.
Sector adoption velocity	claude-sonnet-5	2/5	Environmental science and land management are moderate-to-low digitization sectors with slow AI adoption for fieldwork-heavy tasks, though remote sensing analytics are gaining some traction.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI tools meaningfully assist through automated classification of remote-sensing imagery, spatial mapping, and preliminary impact analysis, reducing time spent on data interpretation. However, augmentation is constrained to pre- and post-processing rather than the core fieldwork and expert judgment.
Augmentation potential	claude-sonnet-5	4/5	AI-powered remote sensing, satellite imagery classification, and GIS tools significantly boost productivity in mapping and inventory analysis, even though fieldwork remains human-led.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can assist with image analysis and some remote-sensing data interpretation, the task requires on-site fieldwork to assess soil disturbance, vegetation, and environmental conditions that current AI systems cannot conduct independently. Significant human judgment about land classification and planning decisions remains essential.
Task automatability	claude-sonnet-5	2/5	Physical land surveying requires on-site presence, sampling, and sensor data collection that current AI cannot perform end-to-end; AI can assist with analysis but not the fieldwork itself.'
Adoption barriers	claude-haiku-4-5-20251001	4/5	Environmental assessments and conservation planning often require licensed environmental scientists or certified soil scientists to sign off on findings for regulatory compliance, permitting, and liability. Professional licensing and regulatory requirements create hard barriers to full automation.
Adoption barriers	claude-sonnet-5	3/5	Environmental impact assessments and conservation planning often require professional certification and regulatory sign-off, creating moderate barriers to full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI remote-sensing and analysis tools can reduce some fieldwork preparation costs, but the core task—physically surveying land and making expert environmental assessments—requires trained scientists whose loaded wages remain substantially lower than integrating autonomous field systems and AI interpretation.
Cost vs. human wage	claude-sonnet-5	2/5	Fieldwork, sample collection, and site-specific judgment still require paid human labor and equipment, so AI only reduces costs for the data-processing portion, not the whole task.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Products exist for remote-sensing analysis and partial mapping support (satellite imagery, GIS tools), but reliable end-to-end surveying and environmental impact assessment still requires trained soil scientists in the field. No deployed system can fully replace the sensory and contextual judgment needed for accurate land classification.
Technical feasibility today	claude-sonnet-5	2/5	Products exist for remote sensing classification and GIS mapping analysis, but full survey workflows including ground-truthing and physical assessment are not automated in deployed systems.

Develop ways of altering soils to suit different types of plants.

CI 25–30 · exposure 20 · augmentation 75 · importance 3.1/5 · 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	Agricultural sectors show modest AI adoption, concentrated in large commercial operations; most soil science remains conducted by traditional research institutions and extension services where digital transformation lags information-sector adoption rates.
Sector adoption velocity	claude-sonnet-5	2/5	Agricultural and soil science R&D sectors show slower AI adoption compared to information-heavy industries, with AI mainly used in narrow data-analysis pilots.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI can significantly assist soil scientists by rapidly analyzing soil samples, cross-referencing plant requirements, predicting outcomes of amendments, and suggesting experimental designs, allowing scientists to focus on validation and field decision-making rather than routine data processing.
Augmentation potential	claude-sonnet-5	4/5	AI tools substantially help analyze soil chemistry data, model plant-soil interactions, and suggest amendment formulations, meaningfully speeding up scientists' research workflow.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can assist in analyzing soil composition data and predicting plant-soil interactions through models, developing effective soil alteration methods requires field testing, practical judgment on local conditions, and iterative experimentation that humans must supervise. Current AI systems cannot independently design and validate soil modification protocols.
Task automatability	claude-sonnet-5	2/5	This requires original experimental research, field testing, and iterative hypothesis-driven soil science that current AI cannot execute end-to-end; AI can support analysis but not perform the core scientific development work.atable.-, so most of the task remains human-driven.
Adoption barriers	claude-haiku-4-5-20251001	3/5	Agricultural and environmental regulations require licensed professionals to oversee soil management practices, especially for commercial or large-scale applications. Additionally, liability for crop failure or environmental damage creates organizational friction against full automation, though advisory AI is acceptable.
Adoption barriers	claude-sonnet-5	3/5	No licensing mandate strictly requires a human, but scientific credibility, publication norms, and practical agronomic accountability create meaningful friction against pure AI substitution.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI tools for soil analysis and plant-soil matching have moderate costs, but the full value chain—from experimental design through field validation—remains labor-intensive and dominated by human scientist time, making the cost ratio unfavorable for automation.
Cost vs. human wage	claude-sonnet-5	2/5	AI can cheaply generate literature summaries or data analysis, but the physical experimentation, soil sampling, and validation loop still require costly human labor and lab/field work.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Some agricultural AI tools can analyze soil samples and recommend amendments based on plant requirements, but no deployed products reliably perform end-to-end soil alteration strategy development at production quality. Existing systems operate as narrow decision-support tools rather than autonomous developers of new soil modification methods.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product autonomously develops novel soil amendment strategies for specific plant needs; this remains a research-stage capability at best.

Research technical requirements or environmental impacts of urban green spaces, such as green roof installations.

CI 25–30 · exposure 25 · augmentation 75 · importance 3.0/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Environmental science and urban planning sectors show slow to moderate AI adoption; while some organizations pilot modeling tools, production deployment of AI-driven environmental impact assessments remains limited, with most firms retaining human scientists in the critical path.
Sector adoption velocity	claude-sonnet-5	2/5	Environmental/soil science and urban planning sectors show slower AI adoption compared to information-heavy industries; AI tool use here remains mostly pilot-stage for literature synthesis rather than production research workflows.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI significantly assists by automating literature synthesis, running environmental simulations, processing satellite/sensor data, and generating baseline impact models that human scientists can refine and validate, substantially raising productivity while they remain the decision authority.
Augmentation potential	claude-sonnet-5	4/5	AI substantially aids in literature review, summarizing environmental regulations, drafting technical reports, and organizing research data, meaningfully boosting scientist productivity while judgment and fieldwork remain human-led.
Task automatability	claude-haiku-4-5-20251001	2/5	AI can assist with literature review, data analysis, and environmental impact modeling, but the task requires field observation, site-specific assessment, and judgment about complex environmental interactions that demand human expertise and on-site evaluation today.
Task automatability	claude-sonnet-5	2/5	AI can assist with literature review, data synthesis, and drafting reports, but original field research, site-specific environmental assessment, and technical judgment require human expertise and fieldwork that AI cannot perform end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Significant barriers exist: regulatory requirements often mandate that licensed professionals (soil scientists, environmental engineers) sign off on impact assessments; liability for incorrect environmental recommendations creates asymmetric error costs; and clients typically require human expert credibility and accountability.
Adoption barriers	claude-sonnet-5	3/5	No strict licensing requirement blocks AI use, but professional accountability, regulatory compliance documentation, and organizational trust in scientific judgment create moderate friction against full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI tools for environmental analysis and modeling are becoming cheaper, but the need for specialized human scientists to validate findings, conduct site visits, and make technical recommendations keeps total task cost comparable to or higher than human-only approaches.
Cost vs. human wage	claude-sonnet-5	2/5	AI can cheaply accelerate literature search and drafting, but the overall task still requires expert scientist time for site analysis, data interpretation, and validation, keeping all-in costs closer to human-comparable.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	While AI tools exist for environmental modeling and literature analysis, deployed products cannot reliably conduct the full end-to-end technical requirements assessment for specific urban green space projects without substantial human oversight and ground-truth validation.
Technical feasibility today	claude-sonnet-5	2/5	AI research assistants and literature-review tools exist and are used for background synthesis, but no deployed product independently conducts environmental impact research or technical requirement analysis for green infrastructure at professional reliability.

Study insect distribution or habitat and recommend methods to prevent importation or spread of injurious species.

CI 25–30 · exposure 25 · augmentation 63 · importance 2.6/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Agriculture and environmental sectors show moderate digitization; while species monitoring tools and distribution databases are being adopted, actual AI-driven decision-making on pest prevention remains in pilot phases rather than widespread production deployment.
Sector adoption velocity	claude-sonnet-5	2/5	Agricultural and environmental science sectors adopt AI more slowly than digital-native industries, with pilots for pest modeling but limited production-scale deployment of end-to-end recommendation systems.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can meaningfully assist by analyzing spatial distribution data, flagging anomalies in insect populations, and generating literature summaries on related species—supporting but not replacing the expert scientist's field judgment and regulatory decision-making.
Augmentation potential	claude-sonnet-5	4/5	AI can strongly assist by modeling species distribution, processing remote sensing/GIS data, and synthesizing research literature, greatly speeding up the scientist's analysis while they retain decision authority.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can assist with analyzing species distribution data and literature, the task requires field observation of insect behavior, expert judgment on ecological interactions, and contextual recommendations that current AI cannot reliably perform end-to-end. The recommendation phase demands nuanced understanding of complex ecological systems and regulatory contexts.
Task automatability	claude-sonnet-5	2/5	This requires fieldwork, sample collection, and expert judgment about ecological risk that AI cannot perform end-to-end; AI can assist with data analysis and literature review but not the full task.'
Adoption barriers	claude-haiku-4-5-20251001	4/5	Regulatory frameworks in agriculture and environmental protection typically require licensed or credentialed scientists to sign off on pest management recommendations due to liability and ecological impact concerns, creating hard barriers to full automation regardless of technical capability.
Adoption barriers	claude-sonnet-5	3/5	Regulatory and biosecurity decisions often require certified expert sign-off and are tied to legal frameworks (e.g., quarantine, invasive species regulations), creating moderate institutional and liability barriers.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	The specialized expertise required—field sampling, species identification, ecological modeling—commands significant human cost, while AI inference for this task remains limited in scope and accuracy, offering only partial support that still requires expert review and validation.
Cost vs. human wage	claude-sonnet-5	2/5	Fieldwork, sampling, and expert synthesis still require paid specialist labor and equipment, so AI only reduces a portion of costs (e.g., literature synthesis, data modeling) rather than replacing the overall cost structure.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	No deployed product reliably performs this full task in production. AI systems can help with data analysis and pattern recognition in insect distribution datasets, but the integrative ecological assessment and actionable prevention strategy formulation remain beyond current system capabilities at scale.
Technical feasibility today	claude-sonnet-5	2/5	No deployed product autonomously studies insect distribution/habitat and issues authoritative prevention recommendations; existing tools are decision-support aids used by human scientists.

Conduct research into the use of plant species as green fuels or in the production of green fuels.

CI 25–30 · exposure 25 · augmentation 75 · importance 2.3/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Agricultural and energy research sectors show slow AI adoption for novel experimental work; most applications remain in data analysis and modeling support rather than autonomous research direction. Full automation of green fuel research is in early pilot stages if at all.
Sector adoption velocity	claude-sonnet-5	2/5	Agricultural and plant science research is a slower-adopting sector for AI agents compared to finance or software, though AI-assisted literature review and bioinformatics tools are gaining some traction.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI significantly augments this task by accelerating literature synthesis, modeling plant biochemistry, optimizing experimental parameters, and analyzing large datasets, allowing scientists to focus on hypothesis testing and field validation. Human researchers remain in the loop but with substantially raised productivity.
Augmentation potential	claude-sonnet-5	4/5	AI significantly aids literature review, genomic/data analysis, and predictive modeling for candidate plant species, meaningfully boosting researcher productivity while humans still design and conduct experiments.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can assist with literature review, data analysis, and computational modeling of plant biochemistry, the core task requires hands-on experimental design, field research, lab work, and novel hypothesis generation that demand human judgment and tacit knowledge. AI cannot currently conduct end-to-end research with 50% time savings at equal quality.
Task automatability	claude-sonnet-5	2/5	Experimental research design, field trials, lab analysis of biomass yields, and novel scientific discovery cannot be end-to-end automated by current AI; AI can assist with literature review, data analysis, and hypothesis generation but the physical experimentation remains manual.'
Adoption barriers	claude-haiku-4-5-20251001	4/5	Research institutions and funding bodies typically require trained researchers to design and execute studies; liability, regulatory compliance, and peer-review requirements mean a credentialed scientist must lead and vouch for the work. Automation faces institutional and professional barriers.
Adoption barriers	claude-sonnet-5	3/5	No licensing requirement per se, but publication standards, peer review, funding body requirements, and scientific rigor norms create moderate friction against pure AI-driven research claims.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI tools (computational chemistry, data analysis software) cost significantly less than a soil/plant scientist's loaded salary, but they cannot replace the core research output; integration still requires expert human oversight and direction, making the all-in cost comparable or favoring the human.
Cost vs. human wage	claude-sonnet-5	2/5	AI can cheaply assist with data crunching and literature synthesis, but the bulk of cost is field trials, lab equipment, and specialized expertise that AI doesn't replace, keeping overall cost ratio close to human-comparable.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	No deployed product reliably performs independent research into green fuel production from plants. AI systems can support components (literature mining, data analysis) but not the integrated experimental research pipeline, which requires laboratory work, field validation, and novel scientific discovery.
Technical feasibility today	claude-sonnet-5	2/5	AI tools like literature-mining assistants and data analysis software support parts of this research, but no deployed product independently conducts biofuel crop research at scale.

Develop methods of conserving or managing soil that can be applied by farmers or forestry companies.

CI 20–34 · exposure 20 · 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	Adoption of AI-driven soil analysis is slow in the broader farming sector, especially among smallholder and traditional operations; most farmers still rely on extension agents and established practices rather than novel AI-generated methods.
Sector adoption velocity	claude-sonnet-5	2/5	Agricultural and forestry science sectors show slower AI adoption compared to information-intensive sectors, with pilots more common than production deployment.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI can significantly augment soil scientists by automating data processing, running simulations, and summarizing literature on conservation techniques, allowing experts to focus on field validation, stakeholder communication, and refining methods for local conditions.
Augmentation potential	claude-sonnet-5	4/5	AI can meaningfully assist by analyzing soil data, modeling scenarios, synthesizing research literature, and generating hypotheses, substantially speeding up the scientist's method-development process.
Task automatability	claude-haiku-4-5-20251001	2/5	AI can assist in analyzing soil data and modeling conservation strategies, but developing practical, context-specific methods for real farmers requires field validation, stakeholder engagement, and iterative refinement that current AI systems cannot do end-to-end autonomously.
Task automatability	claude-sonnet-5	2/5	Developing novel soil conservation methods requires field experimentation, site-specific judgment, and applied research synthesis that current AI cannot originate or validate end-to-end.5
Adoption barriers	claude-haiku-4-5-20251001	4/5	Regulatory and professional barriers are moderate-to-high: soil conservation practices often require certification or approval from agricultural extension services, environmental agencies, or forestry regulators, and stakeholder trust in methods typically depends on a recognized expert's validation and endorsement.
Adoption barriers	claude-sonnet-5	3/5	No licensing requirement to develop methods, but organizational trust, liability for agricultural outcomes, and need for empirical field validation create moderate friction.
Cost vs. human wage	claude-haiku-4-5-20251001	3/5	AI soil analysis and modeling tools can reduce some computational and data-processing costs compared to field surveys alone, but the overall cost of developing and validating practical methods is still primarily driven by expert scientist time and field trials.
Cost vs. human wage	claude-sonnet-5	2/5	AI can cheaply assist with literature review and data analysis, but the core work of designing, testing, and validating methods still requires costly human expertise and field trials.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	While AI tools exist for soil analysis and predictive modeling, no deployed product reliably develops complete, implementable conservation methods from scratch; most systems are lab-based or require significant expert interpretation to translate outputs into actionable farmer guidance.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product autonomously develops and validates soil management methods for real-world agricultural or forestry use; this remains a research and field-testing process.

Develop new or improved methods or products for controlling or eliminating weeds, crop diseases, or insect pests.

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	Agricultural science remains a relatively low-digitization sector outside large agrochemical firms; method development is still primarily conducted by universities and government agencies using traditional R&D models. AI adoption is slow, with tools used mainly as assistants rather than autonomous developers.
Sector adoption velocity	claude-sonnet-5	2/5	Agricultural science and agrochemical R&D are moderate-to-slow adopters of AI compared to digital-native sectors; AI is used in pilot/tool form (e.g., compound screening) but production-scale end-to-end automation is rare.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI meaningfully assists by accelerating literature review, identifying relevant research patterns, suggesting experimental conditions, and analyzing trial data. However, the assistance is bounded—creative hypothesis generation and experimental validation still center on the human scientist.
Augmentation potential	claude-sonnet-5	4/5	AI significantly aids literature review, data analysis, predictive modeling of pest resistance, and compound screening, meaningfully accelerating the scientist's research process while leaving experimental design and validation to humans.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can analyze data on pest/disease patterns and suggest experimental designs, developing genuinely new methods requires iterative laboratory and field testing, hypothesis formation grounded in domain expertise, and validation that current AI systems cannot conduct autonomously. AI assists in literature review and pattern recognition but cannot replace the hands-on experimental work and creative scientific reasoning needed.
Task automatability	claude-sonnet-5	2/5	This is an experimental, hypothesis-driven R&D task requiring field trials, lab work, and novel synthesis of biological insight; AI can assist ideation and data analysis but cannot independently develop and validate new methods or products end-to-end today.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Significant regulatory barriers exist: new pest-control products must pass EPA approval, phytotoxicity testing, and environmental safety reviews that require human expert sign-off. Liability for failed or harmful methods is high, and scientific credibility requires human authorship and responsibility. These create strong friction against full automation.
Adoption barriers	claude-sonnet-5	4/5	New pest/weed control products face heavy regulatory scrutiny (EPA registration, safety/efficacy testing) requiring licensed scientists and formal trials, creating strong institutional and legal barriers to full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI tools (computational biology platforms, literature analysis) reduce some overhead, but the core work—designing experiments, conducting trials, validating results—still requires highly trained scientists. The loaded cost of a soil scientist or agronomist far exceeds current AI assistance for this specialized task.
Cost vs. human wage	claude-sonnet-5	2/5	AI can reduce costs in screening and data analysis stages, but the overall R&D process still requires expensive wet-lab work, field trials, and regulatory testing that AI cannot substitute, keeping total cost comparable to human-led R&D.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	No deployed product reliably develops novel pest-control methods end-to-end. AI tools exist for data analysis and literature mining, but real-world method development remains researcher-driven, requiring wet-lab work, field trials, and regulatory validation that are not yet automated at scale.
Technical feasibility today	claude-sonnet-5	2/5	AI tools (e.g., molecule/compound screening, literature mining, predictive modeling) are used in agrochemical R&D pipelines, but no deployed product autonomously develops novel pest/weed control methods without extensive human experimentation and validation.

Identify degraded or contaminated soils and develop plans to improve their chemical, biological, or physical characteristics.

CI 20–30 · exposure 20 · 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	Adoption of AI in soil science remains limited largely to research and pilot applications in universities and larger consulting firms. The field is relatively small, process-bound by regulation, and reliant on field expertise; digitization and AI integration are slower than in information-intensive sectors.
Sector adoption velocity	claude-sonnet-5	2/5	Agricultural and environmental science sectors show slower AI adoption compared to information/finance industries, with pilots for soil analytics emerging but production-scale deployment still limited.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can usefully assist soil scientists by processing large analytical datasets, suggesting remediation options from literature, and modeling contamination fate—productivity gains in analysis and reporting phases. However, the core interpretive work of field assessment and plan customization remains heavily human-dependent despite these assists.
Augmentation potential	claude-sonnet-5	4/5	AI tools can meaningfully assist by analyzing soil test data, mapping contamination patterns via remote sensing/GIS, and drafting portions of remediation reports, improving scientist productivity while judgment and fieldwork remain human-led.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can assist in analyzing soil data and suggesting remediation strategies based on scientific literature, the task requires field assessment, contextual judgment, and integration of multiple environmental factors that currently demand human expertise. AI systems cannot reliably perform the full end-to-end task of identifying contamination and developing site-specific remediation plans without substantial human oversight.
Task automatability	claude-sonnet-5	2/5	AI can assist with data analysis and literature synthesis but cannot perform physical soil sampling, on-site diagnosis, or generate site-specific remediation plans without extensive expert validation and field verification.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Significant regulatory and liability barriers protect this task: environmental remediation plans typically require licensed soil scientists or hydrogeologists to sign off under state regulations, and contamination assessment carries legal liability if incorrect. Many jurisdictions mandate professional credentials for soil and groundwater work.
Adoption barriers	claude-sonnet-5	3/5	While not always legally mandated to be performed by a licensed professional, many contamination assessments feed into regulatory compliance and liability-sensitive decisions (e.g., environmental remediation), creating moderate oversight requirements.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Current AI applications in soil science are supplementary tools requiring expert human interpretation and validation. The all-in cost of AI infrastructure, expert oversight, and re-analysis when systems err likely approaches or exceeds the cost of direct human expertise for this complex, high-liability task.
Cost vs. human wage	claude-sonnet-5	2/5	Field sampling, lab analysis, and site-specific judgment still require human scientists and technicians, so AI only reduces some analytical/reporting costs rather than replacing the bulk of the cost structure.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Some AI tools exist for soil analysis from laboratory data and literature-based remediation recommendations, but no production systems demonstrably perform the full task of field identification and plan development reliably in deployed settings. Real-world soil assessment requires nuanced interpretation of spatial variability, regulatory context, and site history that current AI handles inconsistently.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product autonomously identifies soil degradation/contamination and develops remediation plans end-to-end; this remains a research-stage capability requiring expert integration with field data.

Develop improved measurement techniques, soil conservation methods, soil sampling devices, or related technology.

CI 20–30 · exposure 20 · augmentation 63 · importance 3.5/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Adoption of AI-assisted development tools in agricultural and soil science remains slow relative to IT or finance sectors. Field-intensive research, reliance on established methodologies, and risk-averse regulatory environments limit rapid deployment of AI agents in this domain.
Sector adoption velocity	claude-sonnet-5	2/5	Agricultural and soil science R&D sectors have historically slower digitization and AI adoption compared to information/finance sectors, with AI used more as a supplementary tool than a driver of core innovation.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can meaningfully assist soil scientists by automating literature search, running computational soil models, analyzing experimental data, and suggesting design variations. These tools raise researcher productivity in the exploratory and analytical phases, but human judgment and field validation remain central to the development process.
Augmentation potential	claude-sonnet-5	4/5	AI can meaningfully assist by analyzing large soil datasets, suggesting design parameters, modeling outcomes, and speeding up literature reviews, boosting scientist productivity in the ideation and analysis phases.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can assist in analyzing experimental data and simulating soil properties, the core task of developing novel measurement techniques and devices requires hands-on experimentation, prototype iteration, and tacit domain knowledge that current AI systems cannot execute end-to-end. AI may accelerate literature review or data analysis components, but cannot achieve the 50% time-saving threshold for the full development cycle.
Task automatability	claude-sonnet-5	2/5	This is a creative R&D task requiring physical prototyping, field testing, and novel engineering insight that current AI cannot execute end-to-end; AI can assist with literature review and design ideation only.7 characters aside, the bulk of hands-on device development and validation remains human-driven.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Soil and plant science is heavily regulated (EPA, USDA, environmental compliance), and new measurement techniques and soil conservation methods require peer review, field validation, and regulatory approval. Published standards and institutional credentialing create meaningful friction against fully automated development; human expert oversight is mandatory.
Adoption barriers	claude-sonnet-5	3/5	No strict licensing requirement mandates a human for invention, but scientific credibility, peer review, and publication norms create moderate friction against fully AI-driven innovation claims.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	The AI cost for supporting (not replacing) development—compute for simulation, data analysis, and literature synthesis—is modest, but the full human team cost (researchers, engineers, field testers) remains dominant. AI does not achieve cost parity, let alone order-of-magnitude savings, for the entire development pipeline.
Cost vs. human wage	claude-sonnet-5	2/5	AI tools can cheaply assist with data analysis or literature synthesis, but the core inventive and experimental work still requires expensive expert labor, fieldwork, and equipment, keeping overall costs comparable to or higher than pure AI use.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	No deployed products reliably perform end-to-end technology development for soil science applications. AI tools exist for data analysis and literature mining, but independent product or agent systems that can design, prototype, and validate soil conservation devices remain at the research or proof-of-concept stage.
Technical feasibility today	claude-sonnet-5	1/5	No deployed AI product autonomously invents or validates new soil sampling devices or conservation methods; this remains a research and engineering activity done by scientists in labs and fields.

Plan or supervise waste management programs for composting or farming.

CI 20–30 · exposure 20 · augmentation 50 · importance 3.1/5 · 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	Soil and plant science remains a moderately digitized sector with slower AI adoption rates; most farms and composting operations are small businesses or traditional agricultural entities that have not yet widely deployed AI-driven waste management systems in production.
Sector adoption velocity	claude-sonnet-5	2/5	Agriculture and environmental management are relatively slow adopters of AI compared to information/finance sectors, with pilots more common than production deployment.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can assist with data logging, compliance checking, scheduling recommendations, and predictive monitoring of composting conditions or soil quality, providing useful support to the planning process, though the human scientist must retain final supervisory authority over program design and regulatory approval.
Augmentation potential	claude-sonnet-5	3/5	AI tools can help analyze soil/compost data, forecast decomposition rates, or optimize logistics, providing moderate productivity gains while humans retain planning and supervisory roles.
Task automatability	claude-haiku-4-5-20251001	2/5	Composting and farming waste management involves site-specific decision-making, regulatory compliance, and field operations requiring human judgment and physical oversight. While AI could assist with monitoring data and scheduling, current systems cannot autonomously plan or supervise end-to-end waste management programs that meet the 50% time-savings threshold when accounting for necessary human validation and site-specific adjustments.
Task automatability	claude-sonnet-5	2/5	Planning and supervising waste management programs requires site assessment, regulatory knowledge, and adaptive decision-making that current AI cannot execute end-to-end; AI can assist with data analysis but not the physical planning/supervision.'
Adoption barriers	claude-haiku-4-5-20251001	4/5	Waste management programs are heavily regulated by federal and state environmental agencies, and a licensed scientist or qualified professional must typically sign off on program plans and compliance. Liability for environmental violations creates strong legal and organizational barriers to full automation.
Adoption barriers	claude-sonnet-5	3/5	While no license is strictly required, environmental regulations, safety compliance, and organizational accountability for waste handling create moderate friction against full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	The oversight, integration, and domain-specific customization required for waste management supervision are substantial compared to a scientist's salary, and the task involves ongoing supervision rather than discrete inference calls, making AI cost-competitive difficult to achieve at equal quality.
Cost vs. human wage	claude-sonnet-5	2/5	Since AI cannot substitute for the supervisory and on-site planning components, any AI use is supplementary, meaning the human cost remains dominant and AI adds cost rather than replacing labor.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	No mature production systems reliably autonomously plan or supervise waste management programs for composting or farming. AI tools can assist with data analysis and reporting, but deployed products do not demonstrably perform the full supervisory task at scale without substantial human oversight and decision-making.
Technical feasibility today	claude-sonnet-5	1/5	No deployed AI product autonomously plans or supervises composting/farm waste programs; this remains a human-led operational and field-based activity.

Conduct experiments regarding causes of bee diseases or factors affecting yields of nectar or pollen.

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	Agricultural and entomological research sectors adopt AI slowly; most bee disease and pollination studies remain in traditional university and research institute settings with limited digital-first workflows. Pilot projects exist, but production-level AI displacement in this domain is minimal.
Sector adoption velocity	claude-sonnet-5	2/5	Agricultural and entomological research sectors adopt AI more slowly than digital-native fields, mostly for data analysis rather than full experimental automation.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can meaningfully assist with literature mining, data visualization, statistical analysis of yield or disease correlations, and predictive modeling of nectar flow or pathogen spread. These tools raise researcher productivity but leave hypothesis testing, experimental design refinement, and field interpretation to human scientists.
Augmentation potential	claude-sonnet-5	4/5	AI can meaningfully assist with experimental design suggestions, statistical analysis, literature synthesis, and data pattern detection, improving researcher productivity substantially.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can assist with data analysis, literature review, and hypothesis generation, the core experimental work—field sampling, bee handling, disease diagnosis, and nectar/pollen measurement—requires hands-on laboratory and field work that current AI cannot perform. End-to-end automation would require robotics and wet-lab automation beyond standard deployments.
Task automatability	claude-sonnet-5	2/5	Designing and physically conducting bee disease experiments requires field/lab manipulation, sampling, and hands-on observation that AI cannot perform; AI can only assist with analysis and literature review portions.dayI
Adoption barriers	claude-haiku-4-5-20251001	4/5	Regulatory frameworks governing apiary research, animal welfare protocols, and biosafety requirements for disease work create substantial barriers. Institutional review, licensing, and legal liability for bee colony health outcomes protect human scientist involvement and oversight.
Adoption barriers	claude-sonnet-5	3/5	No licensing requirement specifically blocks AI, but scientific rigor, physical fieldwork, and institutional/peer review norms create practical barriers to full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	The specialized equipment, field access, and trained personnel required for bee disease and nectar/pollen experiments remain expensive. AI data-analysis tools reduce overhead, but the foundational experimental work and expertise cost dwarf current AI inference costs.
Cost vs. human wage	claude-sonnet-5	2/5	The core experimental labor (field trials, hive inspections, lab assays) still requires paid human researchers and technicians, so AI only marginally reduces overall cost via data analysis support.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	No deployed product reliably conducts the full experimental pipeline independently. AI tools exist for image analysis of bee samples and yield prediction from environmental data, but autonomous experimental design, specimen collection, and disease verification remain research-stage or require significant human oversight.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product autonomously conducts entomological or agronomic experiments end-to-end; this remains a human-driven research process with AI as a peripheral tool.

Perform chemical analyses of the microorganism content of soils to determine microbial reactions or chemical mineralogical relationships to plant growth.

CI 16–30 · exposure 20 · 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	Adoption of AI in soil microbiology is slow; most labs still rely on traditional culturing, PLFA, and sequencing workflows with human interpretation. While some research groups use AI for genomic data analysis, production-scale displacement in agricultural and environmental laboratories remains minimal and primarily in larger research institutions.
Sector adoption velocity	claude-sonnet-5	2/5	Agricultural and environmental science sectors show slow, uneven AI adoption, especially for wet-lab and fieldwork-dependent tasks compared to office-based analysis.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can meaningfully assist by analyzing large microbial datasets, pattern-matching against known microbial-plant relationships, and flagging anomalies in chemical profiles, freeing scientists to focus on hypothesis generation and field validation. However, the core interpretive and experimental work still centers on human expertise and judgment.
Augmentation potential	claude-sonnet-5	3/5	AI can help analyze resulting datasets, model microbial-plant relationships, and draft reports, meaningfully aiding interpretation even though it cannot perform the physical analysis.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can assist with data analysis of microbial composition and pattern recognition from lab outputs, the task fundamentally requires wet-lab chemical analysis, microscopy interpretation, and field sampling judgment that remain dependent on human scientists. AI cannot currently perform the experimental work itself or reliably interpret complex microorganism-mineral-plant interactions without significant human oversight.
Task automatability	claude-sonnet-5	2/5	The physical laboratory work of extracting, culturing, and chemically analyzing soil microorganisms cannot be done by AI; only data interpretation and reporting portions can be assisted, far short of 50% end-to-end time savings.'
Adoption barriers	claude-haiku-4-5-20251001	4/5	Significant barriers exist: regulatory frameworks govern soil and environmental analysis; professional credentials and licensing often required; liability for incorrect microbial or chemical assessments affects land management and agriculture decisions; and organizational expertise in soil science remains concentrated among credentialed scientists.
Adoption barriers	claude-sonnet-5	3/5	No formal licensure typically required, but scientific rigor, equipment access, and quality control create substantial organizational and methodological barriers to automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Current laboratory equipment, reagents, and trained soil scientists remain expensive; AI tools for data post-processing add modest cost savings on analysis time but do not offset the dominant costs of wet-lab work, sampling, and expert interpretation, keeping all-in costs comparable to or higher than human labor.
Cost vs. human wage	claude-sonnet-5	1/5	AI cannot replace the physical lab equipment, reagents, and technician labor needed for microbial and chemical soil analysis, so no meaningful cost substitution exists yet.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Some AI tools exist for analyzing sequencing data and chemical datasets post-collection, but no deployed systems reliably perform end-to-end microbial analysis and the interpretation of complex microorganism-plant-chemistry relationships in production settings. Lab automation handles routine chemistry but not the specialized microbial content analysis described here.
Technical feasibility today	claude-sonnet-5	1/5	No deployed AI product performs wet-lab chemical/microbial soil analysis; this remains a hands-on laboratory and fieldwork task requiring physical instrumentation and sample handling.

Develop environmentally safe methods or products for controlling or eliminating weeds, crop diseases, or pests.

CI 20–25 · exposure 20 · augmentation 63 · importance 3.5/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Agricultural R&D organizations are slow to adopt AI-driven product development; while computational screening is emerging in some biotech settings, the sector remains dominated by traditional field trial workflows, and most soil/plant science teams lack integrated AI pipelines in production.
Sector adoption velocity	claude-sonnet-5	2/5	Agricultural science and agronomy R&D sectors show slow, uneven AI adoption compared to digitized industries, with AI mainly used in ancillary data analysis rather than core method development.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can meaningfully assist with literature synthesis, data visualization of field results, and predictive modeling of pest/disease interactions, helping researchers accelerate analysis and hypothesis generation, though human expertise remains central to experimental design and validation.
Augmentation potential	claude-sonnet-5	4/5	AI substantially aids literature review, molecular screening, predictive modeling of pest behavior, and data analysis, meaningfully speeding up parts of the research process while scientists retain overall control.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can assist in literature review, data analysis of field trials, and computational screening of compounds, developing novel environmentally safe methods requires iterative wet-lab experimentation, ecological testing, and validation in complex real-world conditions that remain largely manual and require human scientific judgment.
Task automatability	claude-sonnet-5	2/5	This is a scientific R&D task requiring hypothesis generation, experimental design, field testing, and iterative validation that current AI cannot execute end-to-end; AI can assist literature review and data analysis but not the physical experimentation core.'
Adoption barriers	claude-haiku-4-5-20251001	4/5	Regulatory agencies (EPA, EFSA) require validated field data and human expert sign-off on safety claims for pest control and crop treatment products; liability and efficacy verification create hard barriers to full automation, and human scientists must legally certify environmental safety.
Adoption barriers	claude-sonnet-5	4/5	Regulatory approval (e.g., EPA pesticide registration) requires human-certified scientific validation and liability accountability, creating strong barriers to full automation of product development claims.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI computational tools are cheap, but the bulk of cost in this task lies in physical field trials, regulatory testing, and expert researcher time—domains where AI provides limited leverage. The loaded cost of a soil scientist conducting experiments typically exceeds the savings from AI-assisted analysis.
Cost vs. human wage	claude-sonnet-5	2/5	AI tools can cheaply assist with literature synthesis or compound screening, but the overall R&D cycle still requires expensive field trials, labs, and expert oversight, keeping costs comparable to human-led research.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	AI tools exist for chemical structure prediction and literature mining, but no deployed product end-to-end performs the full development pipeline (ideation through field validation) of new pest control methods reliably. Most systems remain research-stage or provide only narrow analytical support.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product autonomously develops novel environmentally safe pest/weed/disease control methods; this remains research-stage with human scientists driving discovery.

Plan or supervise land conservation or reclamation programs for industrial development projects.

CI 20–25 · exposure 20 · augmentation 50 · importance 2.5/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Adoption of AI in land conservation is slow; environmental and resource management sectors remain digitization laggards, and conservative regulatory environments discourage algorithmic decision-making on critical environmental projects.
Sector adoption velocity	claude-sonnet-5	2/5	Environmental science and industrial land management are relatively slow-adopting sectors with limited digitization of fieldwork and regulatory processes, though some GIS/remote-sensing tools are gradually being integrated.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can usefully assist with environmental data analysis, modeling, mapping, and documentation preparation, supporting scientists in their planning work; however, the augmentation remains partial because core strategic and supervisory decisions require human expertise and accountability.
Augmentation potential	claude-sonnet-5	3/5	AI can assist with data analysis, environmental modeling, and report generation, improving efficiency in planning stages, but does not transform the supervisory and on-site judgment components of the task.
Task automatability	claude-haiku-4-5-20251001	2/5	Planning land conservation requires domain expertise, stakeholder engagement, regulatory knowledge, and site-specific judgment that AI cannot fully replicate end-to-end. AI can assist with data analysis and documentation, but supervision of on-the-ground programs demands human decision-making and accountability.
Task automatability	claude-sonnet-5	2/5	This task combines site-specific field assessment, regulatory navigation, stakeholder coordination, and hands-on supervision of physical remediation work, which current AI cannot execute end-to-end.We are far from a 50% time-saving threshold on the full task.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Significant barriers exist: environmental regulations often require licensed professionals to sign off on conservation plans, liability for project outcomes rests with human experts, and many jurisdictions legally mandate qualified oversight of land reclamation work.
Adoption barriers	claude-sonnet-5	4/5	Reclamation and conservation programs are typically subject to environmental regulations requiring licensed professionals (e.g., certified soil scientists or engineers) to sign off on plans, creating strong legal and liability barriers to full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI tools for environmental analysis are available but represent a supplement to, not replacement for, specialized soil and plant scientist labor; the loaded cost of a qualified scientist still dominates the economics of these programs.
Cost vs. human wage	claude-sonnet-5	2/5	While some data analysis or report drafting could be cheaply automated, the core supervisory and planning work still requires expensive expert labor and site visits, keeping overall AI substitution costs high relative to savings.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	No deployed product reliably performs the full planning and supervisory role. While AI tools exist for environmental data analysis and reporting, the complex integration of regulatory compliance, stakeholder coordination, and adaptive management on actual projects remains a human responsibility.
Technical feasibility today	claude-sonnet-5	1/5	No deployed AI product plans or supervises land conservation/reclamation programs; at best GIS and modeling tools assist analysis, but the planning and supervisory role remains fully human-driven in practice.

Consult with engineers or other technical personnel working on construction projects about the effects of soil problems and possible solutions to these problems.

CI 11–30 · exposure 13 · 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	Construction and soil science are sectors with slower AI adoption; consulting work remains highly relational and site-specific, with engineers and contractors preferring direct interaction with credentialed professionals rather than AI intermediaries.
Sector adoption velocity	claude-sonnet-5	2/5	Construction and geotechnical engineering sectors have historically slow AI adoption due to physical, site-specific, and liability-heavy work.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can assist by rapidly synthesizing soil data, generating initial problem summaries, or flagging known risk patterns from databases, allowing the soil scientist to focus consultation time on site-specific judgment and solution innovation, though the human expert remains essential.
Augmentation potential	claude-sonnet-5	4/5	AI can help scientists synthesize soil data, generate reports, and draft recommendations, meaningfully speeding up parts of the consultation process while the human remains central to judgment and communication.
Task automatability	claude-haiku-4-5-20251001	1/5	This task requires direct consultation and real-time problem-solving dialogue with engineers on specific construction projects, necessitating domain expertise, site-specific judgment, and the ability to synthesize multiple factors in an interactive setting—capabilities that current AI systems cannot replicate end-to-end with reliable, production-ready quality.
Task automatability	claude-sonnet-5	2/5	This requires site-specific judgment, integration of field data, and real-time collaborative problem-solving with other professionals, which current AI cannot autonomously replace end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Soil and plant science consultation on construction projects typically requires a licensed professional (PE or relevant credentials) to sign off on engineering recommendations, and liability for incorrect soil assessments rests with the credentialed consultant, creating strong legal and regulatory barriers to full automation.
Adoption barriers	claude-sonnet-5	3/5	While not always legally mandated, engineering and geotechnical consulting often requires professional certification and sign-off, and liability for construction failures creates real friction against full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Even though inference cost is low, the task requires specialized domain expertise, site context, and integration with ongoing construction oversight—functions that would still require significant human supervision and domain-expert review, making end-to-end AI cost comparable to or higher than a direct soil scientist consultation.
Cost vs. human wage	claude-sonnet-5	2/5	AI could cheaply generate background reports, but the actual consultative judgment and liability-bearing recommendations still require a paid expert, keeping overall cost comparable to human labor.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	While AI can draft generic soil reports or summarize known soil problems, no deployed product reliably performs real-time consultation with construction engineers on site-specific soil issues requiring professional accountability and dynamic back-and-forth dialogue.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product exists that reliably consults on construction soil problems in place of a scientist; this remains a research-stage capability at best.

Conduct experiments to investigate the underlying mechanisms of plant growth and response to the environment.

CI 7–30 · exposure 13 · augmentation 63 · importance 3.7/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Academic and research institutions adopting AI for plant science remain slow; most experiments are still conducted manually by trained scientists. While computational tools for analysis are common, autonomous experimental platforms have seen limited institutional adoption outside a few research centers.
Sector adoption velocity	claude-sonnet-5	2/5	Agricultural and plant science research is a moderately slow-adopting sector for AI agents, though computational tools (modeling, genomics analysis) are gaining traction.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI meaningfully assists through experimental design recommendation, image-based plant phenotyping, statistical analysis of growth data, and literature synthesis, raising scientist productivity in planning and interpretation. However, the core experimental labor—growing plants under controlled conditions and manual measurement—remains largely human-dependent.
Augmentation potential	claude-sonnet-5	4/5	AI substantially aids experimental design, statistical analysis, literature review, and data interpretation, meaningfully boosting researcher productivity even though hands-on experimentation remains human-led.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can assist with experimental design and data analysis, conducting plant growth experiments requires hands-on manipulation of organisms and environmental conditions in real-world settings—seed sowing, transplanting, environmental control monitoring, and physical observation. Current AI cannot reliably perform these embodied experimental tasks end-to-end without substantial human intervention.
Task automatability	claude-sonnet-5	1/5	Designing and physically conducting biological experiments (growing plants, applying treatments, measuring physiological responses) requires hands-on lab/field work that current AI cannot perform end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Significant barriers exist: scientific peer review and publication require documented experimental integrity and human expertise signing off on results; liability for novel claims rests on the responsible scientist; institutional review boards and funding agencies expect direct human oversight of experimental methods and conclusions.
Adoption barriers	claude-sonnet-5	3/5	No licensing requirement, but scientific rigor, peer review, reproducibility standards, and physical lab access create real organizational friction against full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Current AI tools for experimental support (design software, image analysis) are relatively inexpensive, but the specialized hardware and ongoing human oversight required for autonomous plant experiments would be expensive compared to a trained plant scientist's wage, especially for complex multi-variable studies.
Cost vs. human wage	claude-sonnet-5	1/5	Physical experimentation still requires human labor, equipment, and specialized facilities, so AI cannot substitute cheaply for the core task.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Deployed products exist for experimental design optimization and data analysis, but no production systems autonomously conduct multi-week plant growth experiments with equivalent quality to human scientists. Robotic greenhouse systems exist but remain narrow, require customization, and are not widely deployed as general solutions for this task.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product autonomously runs plant physiology experiments; AI use is confined to research-stage data analysis and hypothesis generation support.

Conduct experiments investigating how soil forms, changes, or interacts with land-based ecosystems or living organisms.

CI 5–30 · exposure 13 · augmentation 63 · importance 3.1/5 · 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	Soil and plant science research remains largely in academic and government institutions that adopt new tools slowly. Field-based experimental work is inherently difficult to digitize and automate, and funding bodies and institutional review boards prioritize human expert oversight, resulting in low adoption of AI-driven automation in this domain.
Sector adoption velocity	claude-sonnet-5	2/5	Agricultural and environmental science sectors show moderate AI adoption for data analysis and modeling, but experimental fieldwork remains largely unautomated with slow uptake.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can meaningfully assist with data processing, statistical analysis, literature synthesis, and hypothesis suggestion from experimental outputs. However, these are ancillary to the core experimental design and execution; AI augmentation improves scientist productivity in analysis and ideation but does not transform their core investigative work.
Augmentation potential	claude-sonnet-5	4/5	AI can meaningfully assist with experimental design, statistical analysis, literature synthesis, and data interpretation, improving productivity around the core hands-on task.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can assist with data analysis and hypothesis generation from experimental results, soil and plant science experiments require hands-on field work, sample collection, and direct observation of complex ecological interactions that cannot be fully automated with current systems. Setup, execution, and adaptation of protocols based on real-world conditions remain dependent on human experimenters.
Task automatability	claude-sonnet-5	1/5	This is hands-on experimental science involving fieldwork, sample collection, and physical lab manipulation of soil and organisms that AI cannot perform end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Soil and plant science experiments often require professional licensing (in some jurisdictions), peer review and publication standards, liability for field safety, and regulatory compliance for certain work. Additionally, ecosystem experiments demand domain expertise and field experience that are difficult to delegate; organizations and funders expect credentialed scientists to design and oversee investigations.
Adoption barriers	claude-sonnet-5	3/5	No licensing requirement blocks AI use in analysis, but the physical nature of experimentation and need for scientific judgment in interpreting novel results creates practical friction.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI tools for analysis and modeling are relatively inexpensive, but they supplement rather than replace the core experimental work. The loaded cost of human soil and plant scientists conducting field experiments remains far lower than attempting to substitute with AI-driven robotic systems or specialized analysis platforms.
Cost vs. human wage	claude-sonnet-5	1/5	AI cannot substitute for the physical experimental work, equipment operation, and field sampling required, so no meaningful cost displacement occurs.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	No deployed AI systems reliably conduct full experimental investigations in soil and plant science independently. AI tools exist for data analysis and literature review, but the physical execution of experiments—field sampling, culturing organisms, measuring soil properties, monitoring ecosystem changes—requires human researchers in the field and laboratory.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product conducts physical soil-ecosystem experiments; AI at most assists with data analysis or literature review, not execution.

Provide advice regarding the development of regulatory standards for land reclamation or soil conservation.

CI 4–25 · exposure 13 · augmentation 63 · importance 3.5/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	1/5	Regulatory development is slow-moving, involves entrenched institutional processes, and remains dominated by human domain experts; adoption of AI for autonomous standard-writing is negligible because regulatory bodies require human expertise and accountability.
Sector adoption velocity	claude-sonnet-5	2/5	Environmental science and regulatory advisory sectors show slow, cautious AI adoption compared to fields like finance or software, largely limited to research support tools.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can assist by synthesizing scientific literature, organizing stakeholder input, generating draft language, and flagging inconsistencies, but the core advisory and standard-setting work remains human-driven and benefits from improved information synthesis rather than full automation.
Augmentation potential	claude-sonnet-5	4/5	AI can meaningfully assist by synthesizing literature, summarizing case studies, drafting policy language, and modeling scenarios, significantly speeding up the scientist's preparatory work.
Task automatability	claude-haiku-4-5-20251001	1/5	Developing regulatory standards requires synthesis of complex scientific evidence, stakeholder input, policy tradeoffs, and legal considerations that demand human judgment and accountability; no current AI system can autonomously produce defensible regulatory advice at equal quality.
Task automatability	claude-sonnet-5	2/5	This requires synthesizing scientific expertise, stakeholder input, and policy judgment into authoritative advice; AI can draft supporting material but cannot autonomously perform the expert advisory role at equal quality.
Adoption barriers	claude-haiku-4-5-20251001	5/5	Regulatory standards must be developed and approved by authorized government bodies or credentialed experts who bear legal responsibility; no jurisdiction allows AI systems to unilaterally author or sign off on formal regulatory standards without human professional oversight.
Adoption barriers	claude-sonnet-5	4/5	Regulatory advisory roles typically require credentialed expertise, accountability, and institutional trust, and agencies generally require named, qualified professionals to provide such advice, creating strong barriers to full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	While AI can reduce research and drafting time, the high-stakes nature of regulatory work demands expert human review and refinement, so total cost savings remain modest relative to the loaded wage of a soil scientist or policy specialist.
Cost vs. human wage	claude-sonnet-5	2/5	While AI can cheaply produce background research and drafts, the credentialed expert review, stakeholder engagement, and liability-bearing advice still require expensive human specialist time, keeping overall cost comparable to human-led work.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed product autonomously generates regulatory standards or advice. AI can assist with research synthesis and drafting, but actual standard-development work remains a human-led process requiring expert authority and accountability.
Technical feasibility today	claude-sonnet-5	2/5	No deployed product independently generates regulatory advice on soil conservation standards; AI is at best used as a research/drafting aid behind a human expert.

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.