Geoscientists, Except Hydrologists and Geographers

19-2042.00

Median wage $101,920/yr23,470 employed (US)Rank #491 of 923 scored · top 53% by substitution

Study the composition, structure, and other physical aspects of the Earth. May use geological, physics, and mathematics knowledge in exploration for oil, gas, minerals, or underground water; or in waste disposal, land reclamation, or other environmental problems. May study the Earth's internal composition, atmospheres, and oceans, and its magnetic, electrical, and gravitational forces. Includes mineralogists, paleontologists, stratigraphers, geodesists, and seismologists.

Sub-scores

0–100 · band = confidence interval from rater disagreement

Substitution27

Exposure25

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

32 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%26

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

Technical feasibility todayw 20%22

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

Cost vs. human wagew 15%26

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

Adoption barriersw 20%inverted — strong barriers lower the score36

panel mean rating 3.5/5 (barrier strength) → substitution pressure 36/100

Sector adoption velocityw 10%27

panel mean rating 2.1/5 → substitution pressure 27/100

Task breakdown (32 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.

Locate and review research articles or environmental, historical, or technical reports.

CI 67–79 · exposure 70 · augmentation 100 · 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	4/5	Academic and research institutions have rapidly adopted literature search and summarization tools; published adoption in geoscience labs and environmental consulting firms demonstrates real production use of AI-assisted research workflows.
Sector adoption velocity	claude-sonnet-5	3/5	Geoscience and earth sciences are moderately digitized with growing use of AI search tools in academic and industry research, but adoption lags behind finance or software sectors.
Augmentation potential	claude-haiku-4-5-20251001	5/5	AI dramatically augments researcher productivity by rapidly filtering, ranking, and summarizing thousands of articles—geoscientists review AI results to identify relevant sources orders of magnitude faster than manual browsing, keeping human expert judgment intact.
Augmentation potential	claude-sonnet-5	5/5	AI substantially accelerates locating, filtering, and summarizing relevant literature and reports, letting geoscientists focus on interpretation and judgment while staying in the loop.
Task automatability	claude-haiku-4-5-20251001	4/5	Current AI systems can efficiently search academic databases, retrieve relevant articles, and summarize reports with >50% time savings. Tools like semantic search, document classification, and summarization models handle the retrieval and initial review phases well, though expert validation remains necessary.
Task automatability	claude-sonnet-5	4/5	AI literature search and summarization tools (e.g., Elicit, AI-powered search engines, LLMs with retrieval) can locate and synthesize articles and reports quickly, saving substantial time though domain-specific technical reports may require verification.
Adoption barriers	claude-haiku-4-5-20251001	2/5	Minimal legal or regulatory barriers exist to automating article location and review; no licensing requirement mandates human involvement. Only modest organizational friction (preference for human curation, access controls on paywalled journals) slows adoption.
Adoption barriers	claude-sonnet-5	2/5	No licensing requirement mandates a human perform this task, but professional reliance on accurate technical interpretation creates some liability-driven caution about fully trusting AI summaries.
Cost vs. human wage	claude-haiku-4-5-20251001	5/5	Database search, document retrieval, and automated summarization cost pennies per query versus the loaded wage of a geoscientist spending hours on manual literature review. The cost ratio is orders of magnitude favorable to AI.
Cost vs. human wage	claude-sonnet-5	4/5	AI search/summarization tools cost a fraction of a geoscientist's hourly wage for equivalent literature scanning, though some oversight cost remains to verify accuracy.
Technical feasibility today	claude-haiku-4-5-20251001	4/5	Deployed products (Google Scholar, ResearchGate, specialized literature review platforms with ML ranking) reliably locate and filter research articles and technical reports at scale. Summarization features are increasingly mature in production, though human experts still verify relevance and quality.
Technical feasibility today	claude-sonnet-5	3/5	Deployed products like Elicit, Semantic Scholar, and LLM-based search assistants perform literature review reliably for general research but can miss or misinterpret specialized geoscience technical/historical reports, requiring human review.

Analyze and interpret geological data, using computer software.

CI 44–50 · exposure 50 · augmentation 75 · importance 4.4/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	3/5	Oil, gas, and mining companies have active AI pilots for data processing and anomaly detection, but production-scale autonomous interpretation is still emerging; adoption is middling, with human-in-the-loop workflows being the norm rather than full replacement.
Sector adoption velocity	claude-sonnet-5	2/5	Geosciences, tied to energy, mining, and environmental sectors, adopt AI tools at a slower pace than software-native industries, with pilots more common than widespread production deployment.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI excels at rapid data visualization, statistical analysis, and anomaly detection that augment a geoscientist's productivity; tools that surface patterns and alternative interpretations while leaving final judgment to the expert can significantly accelerate analysis and reduce manual computational work.
Augmentation potential	claude-sonnet-5	4/5	AI significantly aids data visualization, anomaly detection, and processing of large geological datasets, meaningfully speeding up analysis while geoscientists retain interpretive control.
Task automatability	claude-haiku-4-5-20251001	3/5	AI can automate significant portions of geological data processing, visualization, and pattern recognition using existing geospatial software and machine learning models, but interpretation often requires domain expertise, contextual judgment, and integration with field observations that humans currently retain.
Task automatability	claude-sonnet-5	3/5	AI can assist with pattern recognition, data processing, and interpretation suggestions in geological software, but final interpretation requires domain expertise and judgment on ambiguous, incomplete subsurface data that current systems cannot fully replicate.
Adoption barriers	claude-haiku-4-5-20251001	3/5	Regulatory requirements for resource estimation and subsurface characterization often require licensed professionals' sign-off, and organizational practices favor human expertise for high-stakes interpretations; these create friction but do not fully prevent AI tool adoption as assistants.
Adoption barriers	claude-sonnet-5	3/5	No licensing mandate requires a human for data interpretation, but liability for exploration decisions, safety-critical outcomes (e.g., drilling, hazard assessment), and organizational trust in expert judgment create moderate friction.
Cost vs. human wage	claude-haiku-4-5-20251001	3/5	AI software infrastructure and inference costs are becoming competitive with skilled geoscientist labor, but integration, validation, and the need for expert oversight to interpret results keep the all-in cost roughly comparable to human geoscientists' loaded wages.
Cost vs. human wage	claude-sonnet-5	2/5	Specialized geological AI software and data infrastructure remain costly to license, integrate, and maintain, and human expert oversight is still needed, keeping cost savings modest relative to a geoscientist's output.
Technical feasibility today	claude-haiku-4-5-20251001	3/5	Deployed products exist (e.g., automated rock classification, fault detection, mineral mapping via AI) but are narrow in scope and often require human validation; reliable end-to-end interpretation of complex geological datasets without expert oversight remains limited.
Technical feasibility today	claude-sonnet-5	3/5	ML-based tools exist in seismic interpretation, well-log analysis, and geospatial software (e.g., used in oil & gas and mining), but they still require significant geologist oversight and are narrow in scope rather than fully autonomous.

Review environmental, historical, or technical reports and publications for accuracy.

CI 39–48 · exposure 45 · augmentation 75 · importance 3.7/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Geoscience sectors (geology, mining, environmental consulting) have been slower to adopt automation compared to finance and tech. Document review is still predominantly manual, with AI tools in pilot phases rather than production at scale.
Sector adoption velocity	claude-sonnet-5	2/5	Geoscience and environmental consulting sectors are moderate-to-slow adopters of AI tools relative to finance or IT, with pilots emerging but production-scale review automation still limited.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI can meaningfully assist geoscientists by automatically flagging potential inconsistencies, cross-referencing citations, detecting formatting errors, and highlighting data anomalies, allowing human experts to focus review effort on complex interpretations and domain-specific accuracy.
Augmentation potential	claude-sonnet-5	4/5	AI tools can meaningfully speed up review by highlighting inconsistencies, summarizing lengthy reports, and checking references, substantially aiding the human reviewer even though final judgment remains human-led.
Task automatability	claude-haiku-4-5-20251001	3/5	AI can assist in flagging inconsistencies, data errors, and detecting some factual misstatements across technical reports, but reviewing for full accuracy requires domain expertise and contextual judgment that current systems only partially automate. Significant human oversight remains necessary, limiting time savings to roughly 30-50%.
Task automatability	claude-sonnet-5	3/5	AI can flag inconsistencies, cross-check figures, and summarize technical documents, but validating scientific accuracy against field data and domain expertise still requires human geoscience judgment for a large portion of the task.
Adoption barriers	claude-haiku-4-5-20251001	3/5	Regulatory and liability concerns apply: inaccurate geoscience reports can affect environmental decisions, permit approvals, and public safety, creating organizational hesitancy. Professional responsibility and sign-off by licensed geoscientists remain standard practice in many jurisdictions.
Adoption barriers	claude-sonnet-5	3/5	Many geoscience reports feed into regulatory, environmental compliance, or engineering decisions requiring professional geologist sign-off, creating moderate liability and credentialing barriers to full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	The inference cost of large-scale document review and the oversight burden of validating AI-flagged issues approach the cost of junior-level expert review, especially when false positives must be investigated by experienced geoscientists.
Cost vs. human wage	claude-sonnet-5	3/5	AI-assisted review can cut time on initial screening and summarization at low cost, but final accuracy verification still requires expert human review, keeping overall cost roughly comparable rather than dramatically cheaper.
Technical feasibility today	claude-haiku-4-5-20251001	3/5	Products exist (document analysis tools, plagiarism detection, fact-checking assistants) that can surface potential errors in technical documents, but they struggle with domain-specific accuracy verification and produce false positives requiring expert human validation. Deployment is limited and error rates are material for critical geoscience work.
Technical feasibility today	claude-sonnet-5	2/5	General-purpose LLM document review tools exist and are used for drafting/summarization, but no deployed product reliably verifies technical/geological accuracy of specialized reports in production workflows today.

Identify new sources of platinum group elements for industrial applications, such as automotive fuel cells or pollution abatement systems.

CI 13–62 · exposure 33 · augmentation 88 · importance 3.3/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	3/5	Mining and exploration firms have begun piloting AI-driven prospecting tools and remote sensing analysis, but production deployment remains uneven. Large multinational mining companies adopt faster; smaller explorers lag. Overall adoption is in the pilot-to-early-production phase rather than deep industry-wide integration.
Sector adoption velocity	claude-sonnet-5	2/5	Mining and geoscience exploration sectors are relatively slow AI adopters compared to information/finance, though machine learning is increasingly piloted for prospectivity mapping.
Augmentation potential	claude-haiku-4-5-20251001	5/5	AI significantly augments geoscientist productivity by rapidly filtering candidate deposits, generating exploration hypotheses from multi-source geological data, and accelerating literature synthesis. Geoscientists remain central to interpretation and decision-making, but AI transforms speed and scope of analysis they can manage.
Augmentation potential	claude-sonnet-5	4/5	AI/ML tools can meaningfully assist by analyzing geospatial, geochemical, and remote sensing data to flag promising target areas, significantly aiding geoscientists in narrowing exploration focus.
Task automatability	claude-haiku-4-5-20251001	4/5	Current AI can automate significant portions of this task: literature review, mineral deposit database analysis, geological property prediction from spectral/geochemical data, and candidate site ranking using machine learning models trained on geological data. However, final field validation and economic feasibility assessment still require human judgment, keeping it slightly below full end-to-end automation.
Task automatability	claude-sonnet-5	1/5	This is exploratory geoscience requiring fieldwork, sample collection, geochemical analysis, and expert interpretation of geological formations; no AI system can identify new mineral deposits end-to-end today.'
Adoption barriers	claude-haiku-4-5-20251001	3/5	While no legal license barrier prevents AI use for prospecting analysis, significant organizational friction exists: mining companies rely on senior geoscientist judgment for high-stakes capital decisions, environmental and regulatory approvals require credentialed professionals, and liability for exploration failures creates cautious adoption patterns.
Adoption barriers	claude-sonnet-5	3/5	While not licensed like medicine, mineral exploration claims and reporting often require professional geologist sign-off (e.g., NI 43-101 style disclosures), and high capital stakes create strong preference for expert human judgment.
Cost vs. human wage	claude-haiku-4-5-20251001	4/5	AI-driven analysis of geological data, satellite imagery, and databases costs orders of magnitude less than traditional field surveys and exploration campaigns. However, some integration costs and specialist review remain, preventing a full 5x advantage over the complete human exploration workflow.
Cost vs. human wage	claude-sonnet-5	1/5	Exploration requires physical sampling, lab assays, and expert geological judgment that AI cannot substitute for, so AI does not reduce the dominant cost drivers of this task.
Technical feasibility today	claude-haiku-4-5-20251001	3/5	Deployed AI tools exist for geological prospecting (remote sensing analysis, geochemical modeling, deposit prediction software), but they typically serve as decision-support rather than fully autonomous systems. Current products have material limitations in novel geological contexts and require specialist oversight, preventing reliable full automation without human expertise.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product autonomously discovers new PGE deposits; AI is used only as a minor aid within a research-stage pipeline that still requires extensive human fieldwork and validation.

Develop applied software for the analysis and interpretation of geological data.

CI 30–38 · exposure 25 · augmentation 75 · importance 2.8/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	3/5	Geoscience and energy sectors show moderate AI adoption in data analysis and interpretation, but software development for geological applications remains dominated by human teams with specialized domain knowledge; pilots are visible but production hand-off remains limited.
Sector adoption velocity	claude-sonnet-5	2/5	Geoscience and extractive industries have historically been slower to adopt AI-driven software development compared to pure information/tech sectors, though pilots are emerging.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI coding assistants meaningfully augment geoscientists by auto-completing code, suggesting algorithms, and generating test scaffolds, substantially raising productivity on routine software development tasks while geoscientists retain control over domain logic and validation.
Augmentation potential	claude-sonnet-5	4/5	AI coding assistants (e.g., Copilot-style tools) meaningfully speed up scripting, debugging, and prototyping of geological analysis tools, keeping the geoscientist in the loop for domain validation.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can assist with code generation and data processing pipelines, developing applied geological software requires domain expertise, algorithm selection, validation against real-world data, and iterative refinement that substantially exceeds 50% automation today. Current LLMs can draft components but cannot autonomously ensure the software correctly interprets complex geological phenomena.
Task automatability	claude-sonnet-5	2/5	Software development for niche geological data analysis requires deep domain knowledge, custom algorithms, and iterative validation against physical processes that current AI cannot fully automate end-to-end.iche
Adoption barriers	claude-haiku-4-5-20251001	3/5	Some regulatory and liability friction exists: software validating geological hazards or subsurface models may require human sign-off or certification, and organizations often prefer human accountability for domain-critical tools. However, these are not absolute legal barriers to automation.
Adoption barriers	claude-sonnet-5	2/5	No formal licensing requirement for writing internal analysis software, though correctness for resource/hazard decisions creates some liability and quality-control friction.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI code generation reduces junior developer hours but does not yet eliminate the need for experienced geoscientists to architect, test, and validate software against real geological samples and field data, keeping total cost roughly comparable to or exceeding human-only development.
Cost vs. human wage	claude-sonnet-5	2/5	While AI can accelerate coding, the specialized domain expertise and validation needed for geological software keeps human geoscientist involvement costly and necessary, limiting cost savings.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	No deployed product end-to-end develops domain-specific geological software autonomously. AI coding assistants (GitHub Copilot, Claude) help with boilerplate and syntax but require geoscientific expertise to validate, debug, and integrate outputs; they are research-adjacent rather than production-grade replacements for this task.
Technical feasibility today	claude-sonnet-5	2/5	AI coding assistants can help write portions of applied software, but no deployed product independently develops full geological analysis software reliably in production.

Investigate the composition, structure, or history of the Earth's crust through the collection, examination, measurement, or classification of soils, minerals, rocks, or fossil remains.

CI 30–30 · exposure 25 · augmentation 63 · importance 4.4/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Geoscience is traditionally conservative, embedded in universities and government agencies with slow digitization; adoption of AI classification tools is emerging but remains limited to research labs and larger firms, not mainstream production deployment.
Sector adoption velocity	claude-sonnet-5	2/5	Geoscience and extractive/exploration industries are physically grounded and generally slower AI adopters compared to information-sector benchmarks, though some AI-assisted geological analysis tools are emerging.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI-powered image recognition, spectroscopic analysis, and automated classification can meaningfully assist geoscientists in processing and interpreting samples faster; however, the human expert remains essential for field strategy, unusual specimens, and hypothesis formation.
Augmentation potential	claude-sonnet-5	4/5	AI substantially aids interpretation of geological data, image classification of samples, pattern recognition in stratigraphy, and literature synthesis, meaningfully boosting geoscientist productivity while humans still collect and verify samples.
Task automatability	claude-haiku-4-5-20251001	2/5	Field collection and physical examination of geological samples require hands-on manipulation and contextual judgment in situ; AI can assist with classification and analysis of images/data post-collection, but cannot autonomously perform the full investigative workflow end-to-end with 50% time savings.
Task automatability	claude-sonnet-5	2/5	Physical sample collection, field observation, and hands-on measurement/classification of specimens require in-person work AI cannot perform; only data analysis and literature portions are automatable today.
Adoption barriers	claude-haiku-4-5-20251001	3/5	No hard legal licensing barrier for data analysis, but peer review, publication standards, and organizational reliance on expert judgment create friction; field work in remote or regulated sites may require licensed personnel or permits that limit substitution.
Adoption barriers	claude-sonnet-5	3/5	No licensing requirement mandates a human geoscientist for classification itself, but field access, chain-of-custody, safety, and physical presence create substantial practical barriers to automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI image analysis and classification tools are inexpensive per sample, but field work, sample preparation, and specialized instrumentation (XRD, microscopy) remain human-dependent and costly; the marginal cost advantage of automation is modest relative to the loaded wage of trained geoscientists.
Cost vs. human wage	claude-sonnet-5	2/5	Physical fieldwork, equipment, and travel costs dominate and are not reduced by AI; only the analytical/interpretive sub-tasks see cost savings, keeping overall ratio close to human cost.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	AI tools exist for mineral identification from images and rock classification from spectroscopy data, but they operate on pre-collected samples with material error rates; no deployed system handles the full investigative chain (field collection, measurement, classification, interpretation) reliably in production at scale.
Technical feasibility today	claude-sonnet-5	2/5	AI tools exist for image-based rock/mineral classification and spectral data analysis, but no deployed product handles the full field collection-to-classification workflow reliably in production.

Locate and estimate probable natural gas, oil, or mineral ore deposits or underground water resources, using aerial photographs, charts, or research or survey results.

CI 25–32 · 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	3/5	Oil and gas and mining sectors have adopted AI-assisted interpretation (seismic, hyperspectral imaging) at the pilot and early-production stage, but human geoscientist sign-off remains standard; adoption is moderate, not yet sector-wide replacement.
Sector adoption velocity	claude-sonnet-5	2/5	Oil, gas, and mining sectors are historically slower adopters of AI compared to information/finance industries, though there is growing pilot use of machine learning in seismic and remote sensing interpretation.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI significantly enhances geoscientist productivity by rapidly processing aerial photographs, seismic surveys, and geochemical datasets to highlight anomalies and candidate areas, allowing the expert to focus interpretation on highest-value leads and reduce manual image scanning time.
Augmentation potential	claude-sonnet-5	4/5	AI substantially augments this task by rapidly processing aerial imagery, seismic data, and survey results to highlight probable deposit zones, letting geoscientists focus expert judgment on refined targets.
Task automatability	claude-haiku-4-5-20251001	2/5	AI can assist with analyzing aerial photographs and geophysical data to identify promising deposit locations, but final estimation of deposits and resource probability requires integration of diverse geological, geochemical, and contextual factors that current AI systems handle incompletely. The task demands expert judgment on uncertainty and risk that humans still perform better than 50%-time-saving automation.
Task automatability	claude-sonnet-5	2/5	AI can assist with data analysis and pattern recognition in geophysical datasets but the core task requires integrating multi-source geological judgment, field validation, and domain expertise that current systems cannot fully replicate end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Regulatory and liability frameworks often require a licensed geoscientist or petroleum engineer to sign off on resource estimates and drilling decisions; professional liability for missed or overestimated deposits creates strong legal and organizational friction against full automation.
Adoption barriers	claude-sonnet-5	3/5	No strict licensing mandates a human to make these estimates, but high capital stakes, liability for exploration decisions, and reliance on interpretive geological judgment create meaningful organizational and risk-based friction against pure automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI tools for image and data analysis are relatively inexpensive, but the task still requires expert geoscientists for interpretation, verification, and risk assessment; overall cost remains comparable to or higher than fully human-performed analysis when all integration and oversight are factored in.
Cost vs. human wage	claude-sonnet-5	2/5	Specialized geoscience AI tools require significant data infrastructure, domain-specific model training, and expert oversight, making all-in costs still substantial relative to human geoscientist output, though incrementally cheaper for repetitive screening tasks.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Products exist (seismic interpretation software, remote sensing analysis tools) that handle narrow components, but no deployed system reliably performs end-to-end deposit location and probability estimation at production scale without substantial human geological expertise and field validation.
Technical feasibility today	claude-sonnet-5	2/5	Some AI/ML tools are deployed in exploration geoscience for anomaly detection and seismic interpretation, but they function as decision-support aids within human-led workflows, not autonomous deposit identification systems.

Analyze and interpret geological, geochemical, or geophysical information from sources, such as survey data, well logs, bore holes, or aerial photos.

CI 25–30 · exposure 25 · augmentation 63 · importance 4.3/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Oil and gas, mining, and geotechnical engineering firms are experimenting with AI-assisted log analysis and imaging, but adoption remains pilot-stage; the conservative risk culture, regulatory requirements, and slow capital-project cycles limit rapid deployment compared to finance or software sectors.
Sector adoption velocity	claude-sonnet-5	2/5	Oil and gas, mining, and geoscience sectors have piloted AI/ML for subsurface interpretation but adoption remains uneven, dependent on legacy systems and slow-moving capital-intensive industry practices.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI image enhancement, automated anomaly detection in well logs, and rapid cross-survey data synthesis offer useful assistance in narrowing interpretive focus and accelerating data preparation, but the core interpretive judgment remains the geoscientist's responsibility and is only partially augmented today.
Augmentation potential	claude-sonnet-5	4/5	AI significantly enhances productivity by helping detect patterns in seismic and log data, flagging anomalies, and speeding up preliminary analysis, while final interpretation remains human-driven.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can extract data from images, logs, and surveys automatically, interpreting geological significance requires domain-specific judgment about regional context, risk tolerance, and integration with prior knowledge that current systems struggle to do reliably end-to-end. Partial automation of data extraction is achievable, but the synthesis and interpretation step that defines the task remains primarily human-dependent.
Task automatability	claude-sonnet-5	2/5	AI can assist with pattern recognition in well logs or seismic data, but full interpretation requires integrating multi-source domain expertise, geological judgment, and uncertain, noisy data that current systems cannot reliably synthesize end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Geological interpretations directly inform expensive capital decisions (mineral exploration, drilling, storage siting) where liability and regulatory requirements (environmental, mining, petroleum regulations) typically require a licensed professional geoscientist to validate and sign off on conclusions, creating a hard adoption barrier.
Adoption barriers	claude-sonnet-5	3/5	While no licensing mandate requires a human specifically for interpretation, high-stakes decisions (drilling, resource estimation, environmental risk) create liability and quality-assurance requirements that keep experienced geoscientists in the loop.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Current AI tools require significant setup, ground-truthing by geoscientists, and human oversight; the loaded cost of integration and validation approaches or exceeds the cost of a human geoscientist's time for the same analytical work, especially when rework is factored in.
Cost vs. human wage	claude-sonnet-5	2/5	Specialized geoscience AI tools require expensive domain-specific training data, software licensing, and integration with expert oversight, so total cost is not dramatically lower than employing a geoscientist for critical interpretation.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Image analysis and log-reading tools exist and are being piloted, but deployed systems are narrow in scope (e.g., specific rock type classification) and error rates remain material for high-stakes decisions like drilling location or mineral reserve estimation. No mature production system reliably performs full interpretive analysis across diverse geological data types.
Technical feasibility today	claude-sonnet-5	2/5	Some specialized ML tools exist for seismic interpretation and log analysis in production use at oil/gas companies, but they are narrow, require heavy human review, and are not general-purpose deployed solutions for the full interpretive task.

Prepare geological maps, cross-sectional diagrams, charts, or reports concerning mineral extraction, land use, or resource management, using results of fieldwork or laboratory research.

CI 25–30 · exposure 25 · augmentation 63 · 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	Geosciences operate in moderately digitized sectors (mining, oil & gas, environmental consulting) with adoption of AI-assisted tools in pilots and data processing, but production-scale displacement of map/diagram preparation remains limited due to the need for expert interpretation.
Sector adoption velocity	claude-sonnet-5	2/5	Geosciences and extractive industries are historically slower adopters of AI tools compared to information/finance sectors, though GIS and remote sensing AI adoption is growing in mining and land management.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can meaningfully assist by automating routine diagram rendering, cross-section generation from well logs, and data visualization, reducing drafting time; however, the human geoscientist must validate geological logic and interpretation, limiting the transformative effect.
Augmentation potential	claude-sonnet-5	4/5	AI tools significantly speed up drafting of reports, chart generation, and data visualization, letting geoscientists focus on interpretation and judgment, providing strong augmentation value.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can assist with data visualization and basic diagram generation from structured geological data, the task requires significant domain judgment about mineral extraction feasibility, stratigraphic interpretation, and resource assessment that depends on fieldwork context and professional discretion. End-to-end automation with 50% time savings and equal quality is not achievable today.
Task automatability	claude-sonnet-5	2/5	AI can assist with drafting reports, generating charts, and some GIS visualization, but interpreting field/lab data into accurate geological maps and cross-sections requires domain expertise and spatial reasoning that current AI cannot reliably perform end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Geological maps and resource assessments often inform regulatory compliance, permitting, and land-use decisions where professional geoscientist sign-off is required by law or industry standard. Liability and professional responsibility create strong barriers to full automation or delegating final authority to AI systems.
Adoption barriers	claude-sonnet-5	3/5	Many mineral/land-use reports require professional geologist certification or sign-off for regulatory, environmental, or resource-estimation purposes, creating moderate liability and licensing barriers.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Current AI mapping and diagramming tools require significant expertise to operate, review, and correct; integration and oversight costs are high relative to the time savings achieved, making the all-in cost comparable to or exceeding the cost of skilled geoscientist labor.
Cost vs. human wage	claude-sonnet-5	2/5	While AI can cheaply generate report text and basic visualizations, the core geological interpretation still requires expensive expert oversight, keeping overall costs comparable to human-only workflows.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Some specialized GIS and mapping software incorporate AI-assisted features, but no deployed product reliably generates novel geological maps and cross-sections from raw fieldwork data without substantial human expert review and correction. Most systems require manual input interpretation and professional validation.
Technical feasibility today	claude-sonnet-5	2/5	GIS software with AI-assisted features exists but production tools for autonomous geological map generation from raw fieldwork data are not deployed at scale; most mapping still requires expert geologist interpretation.

Communicate geological findings by writing research papers, participating in conferences, or teaching geological science at universities.

CI 25–30 · exposure 25 · 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	Geosciences remain relatively traditional and cautious sectors with slower digitization than finance or tech. While some researchers use AI writing tools experimentally, institutional adoption remains limited and pilot-stage; displacement is minimal in practice.
Sector adoption velocity	claude-sonnet-5	2/5	Academia and geoscience research are slower-adopting sectors relative to finance or software; AI writing aids are used but production-level automation of research communication and teaching is still nascent.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI can meaningfully augment geoscientists by drafting initial paper sections, generating literature summaries, refining prose, and assisting with figures and data presentation. When the human expert remains in control of scientific content and interpretation, AI substantially accelerates the writing and communication workflow.
Augmentation potential	claude-sonnet-5	4/5	AI tools meaningfully assist with drafting, literature summarization, slide preparation, and editing, significantly boosting productivity while the geoscientist retains authorship and delivery responsibility.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can draft segments of research papers and generate conference abstracts, the task requires original interpretation of geological data, novel synthesis of findings, and domain-specific judgment that current systems cannot reliably produce end-to-end. Teaching and conference participation demand real-time interaction and expert credibility that AI cannot fully replicate.
Task automatability	claude-sonnet-5	2/5	AI can draft text and summarize data but the underlying research, novel interpretation, and live teaching/conference engagement require human expertise and presence, limiting end-to-end automation well below the 50% threshold for the full task bundle.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Strong barriers exist: peer-review and publication gatekeepers require human expert authorship and accountability; academic institutions require credentialed humans to teach; conference organizers expect human presenters. Professional ethics and institutional norms create substantial friction against full automation of scientific communication.
Adoption barriers	claude-sonnet-5	3/5	No formal licensing barrier for writing, but academic norms, peer review, authorship accountability, and the expectation of human-delivered teaching create moderate institutional friction against full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI writing tools are inexpensive per use, but integrating them into peer-reviewed research workflows and replacing experienced geoscientists who command high salaries would require substantial setup and oversight. The cost advantage is modest given the high wage of domain experts and the need for human validation.
Cost vs. human wage	claude-sonnet-5	2/5	AI drafting tools are cheap for text generation, but the task also includes conference participation and university teaching, which still require costly human time and expertise, keeping blended cost comparable to human labor.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	AI tools can assist with writing and outline generation, but no deployed product reliably produces publication-quality geological research papers or handles the full scope of conference participation and university teaching independently. Products exist for writing assistance but fall short of the complete, credible scientific communication required.
Technical feasibility today	claude-sonnet-5	2/5	LLM writing assistants are used for drafting parts of papers, but no deployed product reliably conducts original geological research communication, teaching, or conference presentation at production scale.

Study historical climate change indicators found in locations, such as ice sheets or rock formations to develop climate change models.

CI 25–30 · exposure 25 · augmentation 75 · 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 for climate science is growing in research settings, but production deployment of automated end-to-end climate indicator analysis and model development remains limited; most adoption is in data preprocessing, not autonomous model generation.
Sector adoption velocity	claude-sonnet-5	2/5	Earth science and geoscience research adopt AI tools for data analysis at a moderate pace, but field-based paleoclimate research remains a niche, slower-adopting academic domain compared to fast-moving digital sectors.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI substantially assists geoscientists in analyzing large paleoclimate datasets, visualizing trends, and accelerating preliminary model construction; these tools measurably raise expert productivity while human geoscientists remain central to interpretation and validation.
Augmentation potential	claude-sonnet-5	4/5	AI significantly assists with statistical modeling, pattern recognition in ice-core/sediment data, simulation calibration, and literature synthesis, meaningfully boosting geoscientist productivity while humans handle field collection and interpretation.
Task automatability	claude-haiku-4-5-20251001	2/5	AI can assist in analyzing historical climate data from ice cores and rock formations (e.g., via spectral analysis, pattern recognition), but the full workflow—fieldwork, site selection, sample interpretation in context of geological complexity, and model development—requires expert human judgment and cannot achieve 50% time savings end-to-end with current systems.
Task automatability	claude-sonnet-5	2/5	This requires physical field sampling, specialized lab analysis, and interpretation of novel geological/ice-core data that AI cannot independently collect or fully interpret; AI can assist with data analysis and modeling components but not the full task end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Climate modeling and paleoclimate research are conducted within regulated scientific and policy contexts (e.g., IPCC processes, environmental agencies); model validation requires peer review and expert sign-off, and there is substantial institutional and regulatory friction against full automation of climate-critical research.
Adoption barriers	claude-sonnet-5	3/5	No strict licensing barrier, but rigorous scientific peer review, methodological rigor, and physical access to remote sites create substantial practical barriers to full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI inference for data analysis is cheap, but the full task requires fieldwork, domain expertise for interpretation, and model validation by trained geoscientists; the all-in cost of partial automation does not undercut the wage of specialized domain experts.
Cost vs. human wage	claude-sonnet-5	2/5	Field expeditions, sample collection, and specialized instrumentation remain costly and labor-intensive; AI reduces some data-processing costs but doesn't replace the expensive fieldwork component.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Tools exist for automated data analysis and climate modeling (e.g., ML for paleoclimate pattern recognition), but no deployed product reliably performs the complete task of site selection, field sampling, and integrating findings into validated climate models without substantial expert oversight.
Technical feasibility today	claude-sonnet-5	2/5	Some products exist for climate data analysis and statistical modeling, but no deployed system autonomously conducts field-based paleoclimate indicator study and model development at production reliability.

Identify deposits of construction materials suitable for use as concrete aggregates, road fill, or other applications.

CI 25–30 · exposure 25 · augmentation 50 · importance 3.4/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Geoscience remains a relatively traditional, field-based sector with slower digital transformation. While large mining and construction firms pilot AI-assisted mapping, most deposit identification still relies on conventional survey and sampling methods; production-scale autonomous deployment is uncommon.
Sector adoption velocity	claude-sonnet-5	2/5	Mining, construction materials, and geosciences sectors are relatively slow adopters of AI tools compared to information/finance industries, with most AI use limited to exploratory GIS analytics.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI assists geoscientists by accelerating analysis of satellite imagery, geological databases, and core sample data, speeding up the preliminary screening and mapping phases. However, the gains are moderate—final deposit qualification and suitability assessment still require human expertise and field verification.
Augmentation potential	claude-sonnet-5	3/5	AI-enhanced remote sensing, geospatial data integration, and predictive modeling can meaningfully speed up preliminary deposit identification, helping geoscientists prioritize field investigation sites.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can assist with analyzing geological data, satellite imagery, and core samples to identify aggregate deposits, the task requires field validation, geotechnical testing, and expert judgment about suitability for specific applications that cannot be fully automated end-to-end today. Current AI cannot reliably substitute for on-site assessment and regulatory sign-off without substantial human oversight.
Task automatability	claude-sonnet-5	2/5	This requires field surveys, geological sampling, lab testing of material properties, and site-specific judgment that current AI cannot perform end-to-end; AI can assist with data analysis but not the physical identification and vetting process.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Regulatory requirements and liability mean that aggregate suitability for construction must typically be certified by licensed geoscientists or engineers; permits and resource extraction decisions often require professional sign-off. Customer and regulatory demand for qualified human judgment creates strong adoption friction.
Adoption barriers	claude-sonnet-5	3/5	No strict licensing mandate for AI use, but engineering standards, permitting, and liability for material quality assurance typically require professional geoscientist sign-off, creating moderate friction.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI-assisted analysis of geological data and imagery reduces some analytical costs, but field sampling, lab testing, and expert review remain expensive and human-labor-intensive. The all-in cost of AI tools plus requisite human validation is still comparable to or higher than traditional geoscientist labor for this task.
Cost vs. human wage	claude-sonnet-5	2/5	AI can cheaply process satellite/geological data for initial screening, but the costly components—field verification, sampling, and material testing—still require human geoscientists, keeping overall cost comparable to human-led work.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Products exist for geological mapping and mineral identification from imagery and samples, but they are narrow in scope and typically used as analytical tools rather than standalone systems. Production deployment remains limited; geoscientists still perform most deposit characterization and qualification work manually.
Technical feasibility today	claude-sonnet-5	2/5	GIS and remote-sensing tools with ML components exist for preliminary resource mapping, but no deployed product reliably identifies and qualifies aggregate deposits without extensive human fieldwork and lab confirmation.

Locate potential sources of geothermal energy.

CI 25–30 · exposure 25 · augmentation 75 · importance 3.2/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	The geothermal sector is niche and capital-intensive with slow digitization compared to information or finance sectors; adoption of AI for site selection is emerging but remains limited to larger operators and research institutions rather than widespread industry practice.
Sector adoption velocity	claude-sonnet-5	2/5	Geoscience and energy exploration sectors are moderate-to-slow adopters of AI compared to software/finance, with digitization and adoption lagging due to physical fieldwork and specialized data needs.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI excels at processing large volumes of geological, thermal, and seismic data to highlight anomalies and rank candidate sites, significantly augmenting a geoscientist's ability to synthesize complex datasets and narrow exploration targets before field validation.
Augmentation potential	claude-sonnet-5	4/5	AI-driven geospatial analysis, satellite imagery processing, and pattern recognition in geological datasets substantially speed up identification of candidate sites, greatly aiding the geoscientist's workflow while they retain final judgment.
Task automatability	claude-haiku-4-5-20251001	2/5	AI can assist in analyzing geological data, thermal maps, and subsurface models to identify promising geothermal sites, but the task requires integration of multi-source field data, seismic interpretation, and domain expertise that current systems cannot reliably execute end-to-end without significant human judgment and field validation.
Task automatability	claude-sonnet-5	2/5	This requires integrating geological surveys, subsurface data interpretation, and field-verified geophysical signals; AI can assist with data analysis but cannot independently locate viable geothermal sources end-to-end today.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Geothermal projects require permits, environmental review, and stakeholder approval; exploration decisions carry financial and geological risk that typically demand a licensed geoscientist's professional judgment and liability sign-off before investment.
Adoption barriers	claude-sonnet-5	3/5	No strict licensing requirement mandates a human perform this specific task, but liability for costly drilling decisions and reliance on professional geological judgment create substantial organizational and risk-based friction.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Geothermal exploration requires expensive fieldwork, drilling, and specialized sensors regardless of AI involvement; current AI tools are relatively cheap but cannot eliminate the high capital and labor costs of actual exploration and site confirmation.
Cost vs. human wage	claude-sonnet-5	2/5	AI tools can cheaply process large geospatial datasets, but the overall task still requires expensive field surveys, expert interpretation, and validation, keeping all-in costs comparable to human-led work.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	While machine learning models exist for subsurface analysis and data interpretation tools are deployed, no mature production system reliably locates geothermal sources independently; most applications remain in research or pilot phases with human geoscientists validating all findings.
Technical feasibility today	claude-sonnet-5	2/5	Some geospatial AI tools and ML models exist for anomaly detection in geothermal prospecting, but these are research/pilot-stage aids rather than reliable standalone production systems replacing geoscientist judgment.

Identify possible sites for carbon sequestration projects.

CI 25–30 · exposure 25 · 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	Carbon sequestration is an emerging sector with limited project deployment and slow organizational adoption of digital tools. Pilot programs exist, but the regulatory immaturity and high stakes of site selection mean most organizations continue to rely on traditional human-led geological expertise rather than automated AI systems.
Sector adoption velocity	claude-sonnet-5	2/5	Oil & gas and geoscience sectors are adopting AI/ML for subsurface analysis but at a measured pace, with pilots and decision-support tools more common than fully automated production deployment.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI can meaningfully augment geoscientists by rapidly processing seismic data, screening vast geological databases, and modeling subsurface properties, significantly accelerating the preliminary assessment phase while the expert geoscientist retains final judgment on site viability and regulatory strategy.
Augmentation potential	claude-sonnet-5	4/5	AI-driven geospatial analysis, seismic interpretation, and predictive modeling substantially speed up the screening and narrowing of candidate sites, meaningfully boosting geoscientist productivity while final judgment remains human-led.
Task automatability	claude-haiku-4-5-20251001	2/5	AI can assist with site screening by analyzing geological data, but the task requires integrating complex domain knowledge, site-specific regulatory constraints, and economic feasibility that demands significant human interpretation. Current systems cannot reliably perform the full identification workflow end-to-end without substantial geoscientist oversight.
Task automatability	claude-sonnet-5	2/5	Site identification requires integrating geological survey data, seismic interpretation, reservoir modeling, and regulatory/geographic constraints—AI can assist with data analysis but cannot end-to-end replace the geological judgment and field validation required.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Carbon sequestration projects are heavily regulated by environmental and energy agencies, and site selection decisions carry significant liability for subsurface integrity and permanence. A licensed geoscientist's professional sign-off is typically required for permitting and regulatory compliance, creating a substantial legal barrier to full automation.
Adoption barriers	claude-sonnet-5	3/5	No explicit licensing mandate requires a human to identify sites, but liability, environmental regulation, and permitting processes create substantial oversight requirements and organizational caution around subsurface risk assessments.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI-assisted geological analysis is cost-effective for data processing, but the full site identification process still requires expert geoscientist time for validation, risk assessment, and stakeholder coordination, keeping total cost comparable to or exceeding traditional human-led assessment.
Cost vs. human wage	claude-sonnet-5	2/5	AI tools can cheaply process seismic and geological datasets, but the overall task still requires expensive expert analysis, field surveys, and regulatory compliance work, keeping all-in costs comparable to or only modestly below human-only workflows.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	While ML models can be trained on geological databases and analyze subsurface characterization data, no mature production system exists that reliably identifies carbon sequestration sites autonomously. Deployed tools are primarily data analysis aids rather than integrated site-identification systems meeting real project standards.
Technical feasibility today	claude-sonnet-5	2/5	Some geoscience software incorporates ML for subsurface characterization and basin screening, but no deployed product autonomously identifies and validates sequestration sites; human expert review is standard practice in production workflows.

Provide advice on the safe siting of new nuclear reactor projects or methods of nuclear waste management.

CI 0–55 · exposure 33 · augmentation 63 · importance 3.9/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Nuclear energy is a slow-moving, heavily regulated sector with long project timelines and organizational conservatism around AI adoption in safety-critical domains. Adoption of AI tools remains limited to augmentation rather than replacement, driven by sector inertia and compliance overhead.
Sector adoption velocity	claude-sonnet-5	1/5	Nuclear engineering and geoscience consulting for safety-critical infrastructure is a slow-moving, highly regulated, low-digitization-of-judgment sector with minimal AI agent deployment in production.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI significantly assists geoscientists by automating data integration, rapid scenario modeling, regulatory cross-checking, and visualization of geological hazards, substantially reducing the time for preliminary assessments and enabling exploration of more alternatives while the human expert retains judgment authority.
Augmentation potential	claude-sonnet-5	3/5	AI can assist with data analysis, seismic modeling, literature review, and report drafting, providing meaningful productivity gains while the geoscientist retains full judgment and liability.
Task automatability	claude-haiku-4-5-20251001	4/5	Modern AI systems can synthesize geological, hydrogeological, and regulatory data to generate comprehensive siting recommendations, hazard assessments, and waste management protocols with substantial time savings. However, the task requires expert judgment on edge cases and integration of multiple specialized scientific domains, preventing it from reaching full automatability without human validation.
Task automatability	claude-sonnet-5	1/5	This requires integrating site-specific geological survey data, seismic risk modeling, regulatory judgment, and high-stakes safety advice that current AI cannot autonomously perform end-to-end with equal quality.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Nuclear regulatory frameworks (NRC, IAEA) and licensing regimes explicitly require qualified hydrogeologists and geoscientists to conduct siting and waste-management assessments; liability and safety criticality create hard barriers to full automation. Regulatory approval demands licensed professional judgment and accountability.
Adoption barriers	claude-sonnet-5	5/5	Nuclear siting and waste management are subject to intense regulatory oversight (e.g., NRC) requiring licensed professional geoscientists and engineers to certify safety assessments, creating hard legal barriers to substitution.
Cost vs. human wage	claude-haiku-4-5-20251001	4/5	AI inference and integration costs for large-scale geological modeling and regulatory database synthesis are substantially lower than the loaded cost of senior geoscientists performing iterative site characterization and multi-month feasibility studies, though human experts remain needed for final validation.
Cost vs. human wage	claude-sonnet-5	1/5	Given the extreme liability and required expert sign-off, AI cannot substitute for the human cost structure; any AI use adds to rather than replaces the geoscientist's fee and oversight burden.
Technical feasibility today	claude-haiku-4-5-20251001	3/5	AI-assisted geospatial analysis, risk modeling, and document synthesis tools exist in production (used by energy companies and consultancies), but no deployed system independently advises on nuclear siting without expert geoscientist review and sign-off. The task involves high-consequence decisions where organizations maintain human gatekeeping.
Technical feasibility today	claude-sonnet-5	1/5	No deployed AI product independently provides authoritative siting or nuclear waste management advice; this remains firmly in expert consultant territory with only research-stage geospatial/ML tools as aids.

Research ways to reduce the ecological footprint of increasingly prevalent megacities.

CI 16–35 · exposure 13 · augmentation 75 · importance 3.2/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Environmental research and urban planning sectors are moderately digitized but adopt AI slowly for high-stakes ecological work; most geoscience organizations use AI for data preprocessing rather than autonomous research generation.
Sector adoption velocity	claude-sonnet-5	2/5	Academic and applied environmental science sectors are slower adopters of AI agents relative to finance or software, with AI use mostly limited to literature search and data analysis pilots.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI excels at augmenting geoscientists through rapid literature synthesis, pattern discovery in large environmental datasets, scenario modeling, and carbon footprint calculation—substantially raising productivity while the expert remains in the loop to validate and direct research.
Augmentation potential	claude-sonnet-5	4/5	AI can significantly assist geoscientists by synthesizing literature, modeling scenarios, analyzing large datasets, and drafting reports, meaningfully speeding up parts of the research process.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can analyze data on urban ecology, carbon emissions, and land use patterns, this task fundamentally requires creative problem-solving, stakeholder engagement, and synthesis of novel solutions tailored to specific megacity contexts. AI tools can assist with data processing but cannot independently conceive holistic reduction strategies that meet the ≥50% time-saving bar.
Task automatability	claude-sonnet-5	1/5	This is open-ended scientific research requiring novel synthesis, field data, fieldwork-informed judgment and original hypothesis generation that current AI cannot perform end-to-end at equal quality.
Adoption barriers	claude-haiku-4-5-20251001	3/5	This research often informs policy and environmental planning, creating some regulatory and institutional friction around who can author authoritative recommendations, though there is no hard legal requirement that only licensed humans perform the analytical work.
Adoption barriers	claude-sonnet-5	2/5	No licensing requirement mandates a human for this research, but publication credibility, peer review, and domain expertise create moderate organizational friction against pure AI substitution.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Geoscientist expertise commands high loaded wages; AI inference and data analysis tools are inexpensive, but integrating them into novel research workflows and validating outputs still requires substantial expert oversight, making the all-in cost closer to human equivalence.
Cost vs. human wage	claude-sonnet-5	2/5	AI can cheaply assist with literature synthesis and data processing, but the core research task still requires expensive expert scientists, so overall cost savings are limited.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	No deployed product reliably performs independent ecological footprint reduction research; existing AI systems can support literature review and data analysis but lack the domain reasoning and contextual judgment needed to generate vetted recommendations for complex urban systems.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product autonomously conducts original ecological-footprint research on megacities; existing tools are only research-stage aids for literature review or data analysis.

Test industrial diamonds or abrasives, soil, or rocks to determine their geological characteristics, using optical, x-ray, heat, acid, or precision instruments.

CI 21–30 · exposure 20 · augmentation 50 · 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	Geoscience and materials testing labs operate in traditionally slower-adopting sectors (academic, government, specialized industries). Digitization of these workflows is uneven, and most labs still rely on manual expert assessment rather than automated pipelines.
Sector adoption velocity	claude-sonnet-5	2/5	Geoscience and mining/exploration sectors are relatively slow adopters of AI for physical lab work, with automation concentrated in data analysis rather than hands-on testing.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI-powered image analysis and automated data logging can meaningfully assist geoscientists by speeding sample screening, flagging anomalies, and standardizing measurement recording, but the core interpretive judgment and instrument operation remain human-centered.
Augmentation potential	claude-sonnet-5	3/5	AI can assist by analyzing spectral data, X-ray diffraction patterns, or image data from optical instruments to help interpret results faster, though the physical testing itself remains manual.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can assist with image analysis of optical and x-ray data, the task requires hands-on operation of precision instruments, physical sample handling, and complex judgment calls about geological characteristics that demand human expertise and sensory feedback. End-to-end automation would require robotics integration that is not yet standard in geological labs.
Task automatability	claude-sonnet-5	2/5	This requires physical manipulation of samples with specialized lab instruments (XRD, optical microscopy, acid tests) and hands-on execution that current AI cannot perform end-to-end; AI can only assist with data interpretation afterward.'
Adoption barriers	claude-haiku-4-5-20251001	3/5	Professional geoscience roles may involve licensing or certification requirements in some jurisdictions, and liability concerns around material certification (especially for industrial diamonds and abrasives) create moderate friction, though no hard legal mandate prevents machine testing given human oversight.
Adoption barriers	claude-sonnet-5	3/5	While not licensed in the sense of medicine or law, specialized technical expertise, safety protocols (acid handling, X-ray safety), and instrument calibration create meaningful operational barriers to full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	The specialized equipment, integration costs, and ongoing maintenance of automated testing systems rival or exceed the cost of trained geoscientists performing manual testing, particularly given the low-volume, bespoke nature of many analyses.
Cost vs. human wage	claude-sonnet-5	1/5	AI cannot replace the physical testing equipment, sample handling, and instrument operation, so there is no cost substitution possible for the core task.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Image recognition systems exist for some geological sample analysis, but deployed products for comprehensive testing across multiple modalities (optical, x-ray, heat, acid, precision instruments) with reliable accuracy remain limited. Lab automation for these specific analyses is niche and requires significant customization.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product autonomously runs physical geological testing procedures; this remains a manual lab/field task performed by trained scientists or technicians.

Assess ground or surface water movement to provide advice on issues, such as waste management, route and site selection, or the restoration of contaminated sites.

CI 25–25 · exposure 25 · augmentation 63 · 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	Geoscience consulting and environmental remediation are relatively slow to digitize; adoption of AI tools is limited to academic modeling and some large firms' analytic pipelines; most advisory work remains human-expert-driven in consulting and government agencies.
Sector adoption velocity	claude-sonnet-5	2/5	Environmental consulting and geosciences are moderate-to-low digitization sectors with slow AI adoption relative to finance or software, though some firms pilot AI-assisted modeling tools.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can meaningfully augment geoscientists by automating data preprocessing, visualizing water movement models, and flagging anomalies in contaminant transport simulations; however, the human expert retains essential judgment on site risk and remediation strategy.
Augmentation potential	claude-sonnet-5	4/5	AI can meaningfully assist in data analysis, predictive modeling of contaminant transport, and report drafting, significantly boosting geoscientist productivity while human judgment and site work remain essential.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can assist with data analysis and modeling of water movement patterns, the task requires integrating hydrogeological expertise, site-specific physical assessment, contamination risk judgment, and complex advisory recommendations that cannot yet be fully automated end-to-end with equivalent quality at ≥50% time saving.
Task automatability	claude-sonnet-5	2/5	This requires field data collection, site-specific hydrogeological modeling, and professional judgment integrating multiple physical variables that AI cannot independently gather or validate; AI can assist analysis but not replace the end-to-end assessment and advisory role.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Regulatory and liability barriers are substantial: site assessment, remediation recommendations, and waste management advice often require licensed professional engineers or geoscientists to sign off; environmental compliance and risk accountability typically demand human professional responsibility.
Adoption barriers	claude-sonnet-5	4/5	Environmental consulting and contaminated site remediation typically require licensed professional geoscientist/engineer certification and regulatory compliance, creating strong legal barriers to full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI-based water flow modeling and analysis can reduce computational overhead, but the advisory component—integrating site conditions, regulatory requirements, and risk mitigation—still demands expert geoscientist review; full-task cost remains comparable to or higher than human specialist labor.
Cost vs. human wage	claude-sonnet-5	2/5	AI tools can reduce modeling and data-processing time, but the overall cost is still dominated by fieldwork, site-specific expertise, and liability-bearing sign-off, keeping the ratio closer to human-comparable.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Hydrogeological modeling software and AI-assisted tools exist, but they remain research or specialized-tool territory; no production system reliably performs the full advisory task (assessment + recommendations on waste management, routing, remediation) without substantial expert oversight and validation.
Technical feasibility today	claude-sonnet-5	2/5	Some modeling software incorporates AI/ML for groundwater simulation, but no deployed product performs full site assessment and regulatory-grade advisory work autonomously in production.

Identify risks for natural disasters, such as mudslides, earthquakes, or volcanic eruptions.

CI 25–25 · exposure 25 · augmentation 63 · 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	Adoption remains slow outside research institutions and large government agencies; most hazard assessment relies on traditional field surveys and expert judgment. Small- to mid-sized consultancies, which conduct much geohazard work, have limited automation incentives and high switching costs.
Sector adoption velocity	claude-sonnet-5	2/5	Geoscience and hazard assessment fields have moderate digitization but adopt AI more slowly than software-centric professional sectors, with pilots more common than production deployment.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can meaningfully assist by pre-processing satellite imagery, running automated hazard models, and flagging anomalies in seismic or geological data, allowing geoscientists to focus interpretation effort on high-risk areas and reducing analysis time for routine surveys.
Augmentation potential	claude-sonnet-5	4/5	AI substantially aids pattern recognition in seismic, satellite, and geospatial data, helping geoscientists identify risk indicators faster while they retain judgment and final assessment authority.
Task automatability	claude-haiku-4-5-20251001	2/5	AI can assist with pattern recognition in seismic data and geological surveys, but identifying disaster risks requires integration of complex spatial, temporal, and contextual factors that involve significant human geological judgment and field validation. Current systems cannot reliably end-to-end replace the expert interpretation needed to assess actual risk with sufficient confidence.
Task automatability	claude-sonnet-5	2/5	Risk identification requires integrating field geology, remote sensing, historical data, and site-specific judgment that current AI can partially support but not fully replace end-to-end at equal quality.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Disaster risk assessment directly informs public safety, emergency planning, and liability; errors carry high human and financial costs. Regulatory bodies and insurance/government agencies typically require certified geoscientist sign-off, creating both legal and reputational barriers to full automation.
Adoption barriers	claude-sonnet-5	4/5	Hazard assessments often feed into regulatory, insurance, and public safety decisions requiring licensed geoscientist sign-off, creating strong liability and authorization barriers.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	While inference costs are low, integrating AI systems into geohazard workflows requires substantial domain-expert oversight, validation, and field sampling—costs that often approach or exceed the salary of a geoscientist analyst for comparable output quality.
Cost vs. human wage	claude-sonnet-5	2/5	AI can cheaply process satellite and sensor data, but the overall task still requires expensive expert interpretation, fieldwork, and validation, keeping costs comparable to human-led analysis.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	ML models exist for earthquake prediction and hazard mapping, but they operate at research maturity with material error rates and cannot reliably forecast mudslides or volcanic activity. No deployed product consistently performs integrated multi-hazard risk identification at production scale with the accuracy required for decision-making.
Technical feasibility today	claude-sonnet-5	2/5	Some deployed tools exist for hazard mapping and seismic/volcanic monitoring analytics, but they assist rather than autonomously perform comprehensive risk identification in production workflows.

Conduct geological or geophysical studies to provide information for use in regional development, site selection, or development of public works projects.

CI 25–25 · exposure 25 · augmentation 75 · importance 3.7/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Adoption of AI in geoscience remains slow outside research and large oil/gas firms. Most regional planning and public works agencies rely on traditional consulting workflows; digitization and AI adoption lag relative to information-sector benchmarks.
Sector adoption velocity	claude-sonnet-5	2/5	Geoscience and civil engineering sectors are historically slow to adopt AI at scale, relying on established fieldwork and regulatory-driven workflows with limited digitization compared to other professional fields.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI significantly augments geoscientist productivity through automated interpretation of seismic and well-log data, 3D modeling, and spatial analysis. Geoscientists using these tools can process and synthesize data faster while maintaining final judgment and sign-off authority.
Augmentation potential	claude-sonnet-5	4/5	AI substantially aids data processing, geophysical modeling, GIS analysis, and report drafting, meaningfully speeding up analysis while a geoscientist remains responsible for field validation and interpretation.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can assist with data analysis and modeling of geological/geophysical datasets, the task requires site-specific judgment, integration of multiple data sources, stakeholder consultation, and regulatory navigation that current systems cannot do end-to-end at 50% time savings. AI tools are useful for processing surveys and simulations but not for the full advisory output.
Task automatability	claude-sonnet-5	2/5	This is a multi-stage field-to-analysis task involving site visits, sample collection, geophysical survey design, and integrative judgment about regional geology that current AI cannot execute end-to-end.'
Adoption barriers	claude-haiku-4-5-20251001	4/5	Public works and site selection decisions often require licensed professional geoscientists to sign off on findings due to liability and safety implications. Regulatory frameworks and client contractual requirements typically mandate human professional involvement and responsibility for the final determination.
Adoption barriers	claude-sonnet-5	4/5	Public works and regional development projects often require licensed professional geologists/engineers to sign off on site assessments, and liability for geological misjudgment (e.g., seismic or foundation risk) is high.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI inference for geophysical analysis is relatively cheap, but the task requires expensive human expertise in site assessment, regulatory compliance, and stakeholder communication. The overhead of oversight and integration with domain specialists keeps total cost comparable to or exceeding a geoscientist's labor.
Cost vs. human wage	claude-sonnet-5	2/5	Fieldwork, instrumentation, and expert judgment dominate costs; AI reduces some data-processing time but the human expert and physical survey costs remain, so overall savings are modest.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	AI products exist for specific subtasks (seismic interpretation, geological mapping from satellite imagery) but no deployed system reliably performs the full task of conducting studies and providing integrated recommendations for regional development or site selection independently. Tools are supplementary, not autonomous.
Technical feasibility today	claude-sonnet-5	2/5	AI tools assist with subsurface data interpretation and modeling (e.g., seismic interpretation software with ML), but no deployed product performs full geological/geophysical studies for public works siting autonomously.

Design geological mine maps, monitor mine structural integrity, or advise and monitor mining crews.

CI 25–25 · exposure 25 · 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	Mining is a capital-intensive, physically dispersed sector with strong institutional reliance on licensed professionals and slow digital transformation compared to information and finance sectors. Adoption of AI-driven geological analysis remains in the pilot phase; most major mining operations still rely on traditional expert-led structural monitoring and mine design.
Sector adoption velocity	claude-sonnet-5	2/5	Mining is a physically-oriented, moderately digitized sector with slower AI adoption compared to information/finance industries, though remote sensing and monitoring tech adoption is growing steadily.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can meaningfully assist geoscientists by accelerating data processing, pattern recognition in seismic or core samples, and visualization of 3D mine models, which raises productivity on analytical aspects of the work. However, the human expert remains essential for site-specific judgment, regulatory compliance, and safety decisions.
Augmentation potential	claude-sonnet-5	4/5	AI-driven GIS tools, predictive analytics for structural monitoring, and sensor data interpretation significantly boost geoscientists' productivity in mapping and risk assessment while humans retain final judgment and field oversight.
Task automatability	claude-haiku-4-5-20251001	2/5	Geological mine mapping and structural monitoring require field data collection, spatial reasoning, and contextual interpretation of subsurface conditions that remain difficult for current AI systems. While AI can assist with data analysis and visualization of existing maps, the full end-to-end task of designing new mine maps or monitoring structural integrity in changing conditions involves domain expertise and site-specific judgment that current systems cannot reliably replicate at scale.
Task automatability	claude-sonnet-5	2/5	This is a composite task combining spatial CAD-like mapping, physical structural monitoring, and on-site crew supervision, only the mapping component has strong AI leverage while the rest requires physical presence and judgment.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Mining operations face regulatory requirements (mine safety, environmental compliance) and liability concerns that mandate professional geoscientist oversight and sign-off. Many jurisdictions legally require a licensed geoscientist or mining engineer to certify structural assessments and design mine plans, creating hard barriers to full automation.
Adoption barriers	claude-sonnet-5	4/5	Mine safety is heavily regulated, structural integrity assessments often require licensed professional engineers/geologists to sign off, and crew advising requires human authority and accountability for safety-critical decisions.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI inference for geological analysis is relatively inexpensive, but integration with field data collection, domain expertise validation, and the high cost of failure in mining make total cost-per-task comparable to or higher than employing a geoscientist, especially when accounting for liability and oversight.
Cost vs. human wage	claude-sonnet-5	2/5	Sensor-based monitoring and mapping software reduce some labor costs, but the safety-critical judgment and physical crew interaction still require paid geoscientists, keeping AI substitution costs comparable or only modestly cheaper.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	No deployed product reliably performs the full task of designing mine maps or monitoring structural integrity autonomously. While AI tools exist for image analysis and data visualization in mining contexts, actual production systems still require heavy human oversight and depend on expert geoscientists to interpret findings, validate recommendations, and ensure safety compliance.
Technical feasibility today	claude-sonnet-5	2/5	GIS and mine-planning software exist with AI-assisted features, but no deployed product autonomously designs mine maps, monitors structural integrity via sensors and interprets results, and advises crews without heavy human oversight.

Research geomechanical or geochemical processes to be used in carbon sequestration projects.

CI 20–30 · exposure 20 · 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	Geoscience research, particularly in carbon sequestration, remains knowledge-intensive and slow to adopt fully autonomous AI workflows; pilots exist but production deployment of AI-driven research remains limited, with human researchers still leading investigations.
Sector adoption velocity	claude-sonnet-5	2/5	Geoscience and energy/environmental research sectors are relatively slow adopters of AI agents for core scientific discovery work compared to fast-moving digital-native sectors like finance or software.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI excels at augmenting geoscientists by accelerating data processing, running simulations, literature synthesis, and identifying patterns in large datasets; a human researcher leveraging AI tools for these tasks can substantially increase research velocity and explore more scenarios.
Augmentation potential	claude-sonnet-5	4/5	AI can meaningfully assist geoscientists by analyzing large geochemical/geomechanical datasets, running simulations, reviewing literature, and generating hypotheses, substantially boosting research productivity while humans retain interpretive control.
Task automatability	claude-haiku-4-5-20251001	2/5	AI can assist with literature review, data analysis, and modeling of geomechanical/geochemical processes, but the task requires novel experimental design, hypothesis generation, and integration of domain-specific scientific judgment that current systems cannot reliably perform end-to-end without expert oversight.
Task automatability	claude-sonnet-5	2/5	This is complex primary research requiring lab work, field data collection, modeling, and novel scientific interpretation that AI cannot yet perform end-to-end; AI can assist with literature review, data analysis, and modeling components but not conduct the full research process.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Carbon sequestration research is heavily regulated and typically requires peer review, publication in credible venues, and regulatory approval; organizations and funding bodies require human researchers to be accountable for novel findings, creating both liability and institutional friction against full automation.
Adoption barriers	claude-sonnet-5	3/5	No strict licensing requires a human to do this specific research, but organizational and scientific rigor norms, peer review, and the need for verifiable field/lab data create moderate friction against full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI modeling and analysis tools reduce some computational costs, but the specialized software, integration, and substantial human expert review required for research-grade geomechanical/geochemical work keeps total cost comparable to or higher than direct expert time.
Cost vs. human wage	claude-sonnet-5	2/5	While AI tools can cheaply assist with literature synthesis and some modeling, the core research still requires expensive specialized human expertise, lab equipment, and field work, so overall cost savings are limited.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	While AI tools exist for data analysis and simulation in geoscience, no deployed product reliably performs the full research cycle (hypothesis formation, experimental design validation, interpretation of complex geochemical interactions) for carbon sequestration projects at production quality.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product autonomously conducts geomechanical or geochemical research for carbon sequestration; this remains squarely in specialized scientific research territory requiring domain expertise and physical experimentation.

Review work plans to determine the effectiveness of activities for mitigating soil or groundwater contamination.

CI 20–25 · exposure 20 · augmentation 50 · importance 3.7/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Geoscience consulting and environmental firms adopt AI slowly for core technical reviews; most use remains in data processing and visualization. Regulatory and liability concerns, combined with the specialized expertise required, limit production deployment of AI for autonomously reviewing remediation effectiveness.
Sector adoption velocity	claude-sonnet-5	2/5	Environmental consulting and geoscience remain a moderately low-digitization sector with slow AI adoption for regulatory/technical review tasks compared to finance or software industries.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can assist geoscientists by automatically organizing and visualizing contamination datasets, flagging anomalies in monitoring data, and summarizing regulatory requirements, reducing review time. However, the core judgment task of assessing remediation effectiveness remains primarily human-directed.
Augmentation potential	claude-sonnet-5	3/5	AI tools can help summarize plans, cross-check data against regulations, and flag potential gaps, providing useful augmentation to a geoscientist's review process without replacing their judgment.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can extract and summarize information from work plans and contamination data, reviewing effectiveness requires integrating complex site-specific hydrogeological knowledge, regulatory context, and professional judgment about remediation strategies. Current systems lack the domain expertise and contextual reasoning to autonomously validate mitigation effectiveness at equal quality.
Task automatability	claude-sonnet-5	2/5	Reviewing remediation work plans requires site-specific geological judgment, regulatory context, and risk assessment that current AI cannot reliably replicate end-to-end, though it can assist with document review and summarization.The core evaluative judgment remains human-driven.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Regulatory frameworks (CERCLA, RCRA, state environmental agencies) often require licensed environmental professionals to review and sign off on remediation plans, creating a legal barrier to full AI substitution. Professional liability and environmental compliance obligations further protect human decision-making authority.
Adoption barriers	claude-sonnet-5	4/5	Environmental remediation plans often require sign-off by licensed professional geologists/engineers and are subject to regulatory review (e.g., EPA, state agencies), creating strong professional and legal barriers to full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	The cost of integrating and overseeing AI systems to review technical remediation plans, plus the liability of errors, likely exceeds the loaded wage of a geoscientist who performs this task directly. High-stakes environmental decisions make AI cost-efficiency marginal.
Cost vs. human wage	claude-sonnet-5	2/5	AI could cheaply assist with drafting summaries or flagging inconsistencies, but the liability and expertise required for actual plan approval means a licensed geoscientist's oversight cost dominates, keeping overall cost comparable to or only modestly cheaper than human review.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	No deployed AI product reliably performs independent reviews of soil/groundwater remediation work plans in production. Tools exist for data analysis and document summarization, but the task requires specialized geoscience judgment that remains human-dependent in practice.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product autonomously reviews and validates contamination mitigation work plans in production; this remains a specialized expert task with no mature commercial AI substitute.

Develop strategies for more environmentally friendly resource extraction and reclamation.

CI 20–25 · exposure 20 · 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	Mining and resource extraction sectors show modest AI adoption overall, concentrated in predictive analytics and equipment optimization rather than strategic planning. Reclamation strategy development remains largely human-driven, with slow organizational shift toward AI-assisted approaches.
Sector adoption velocity	claude-sonnet-5	2/5	Mining, oil & gas, and extraction industries are traditionally slower 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 usefully assist by analyzing geological data, simulating environmental outcomes, retrieving relevant case studies, and identifying regulatory constraints. These tools enhance a geoscientist's productivity in research and analysis, though strategic synthesis still requires human judgment and accountability.
Augmentation potential	claude-sonnet-5	4/5	AI can meaningfully assist by synthesizing environmental regulations, modeling scenarios, drafting reports, and surfacing best practices, significantly speeding up the strategy development process while humans retain decision authority.
Task automatability	claude-haiku-4-5-20251001	2/5	Strategy development for environmental reclamation requires complex tradeoff analysis, stakeholder input, site-specific geological knowledge, and novel problem-solving. While AI can assist with literature review and scenario modeling, end-to-end strategy creation with equivalent quality remains beyond current capabilities without substantial human direction.
Task automatability	claude-sonnet-5	2/5	This requires synthesizing site-specific geology, regulatory constraints, and environmental impact tradeoffs into novel strategies, which involves judgment and field knowledge that AI cannot yet fully replicate end-to-end., so time savings are partial at best.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Regulatory requirements, permitting obligations, and liability for environmental impact create substantial barriers. Jurisdictions typically require licensed geoscientists or engineers to sign off on extraction and reclamation plans, and stakeholders often demand human accountability for environmental decisions.
Adoption barriers	claude-sonnet-5	4/5	Environmental and reclamation plans are often subject to regulatory review and require professional geoscientist sign-off, creating liability and licensing barriers to full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Current AI systems require significant setup, domain expert validation, and iterative refinement to produce usable strategies. The labor cost of human geoscientists remains lower than the total cost of AI infrastructure, oversight, and rework for this specialized task.
Cost vs. human wage	claude-sonnet-5	2/5	AI can cheaply generate literature summaries or draft options, but the core strategic work still requires expensive expert geoscientist time for validation and site-specific judgment, keeping overall costs comparable to human-only work.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	No deployed product reliably generates standalone extraction and reclamation strategies at production scale. Research tools exist for environmental modeling and data analysis, but generating integrated, implementable strategies requires human geoscientists to synthesize and validate outputs.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product autonomously develops environmentally-friendly extraction/reclamation strategies; this remains expert-driven consulting work with AI only as a research aid.

Determine ways to mitigate the negative consequences of mineral dust dispersion.

CI 11–30 · 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	2/5	Mining and mineral extraction sectors are moderate in digital adoption; while some use AI-assisted modeling, actual mitigation strategy decisions remain human-driven and sector adoption of autonomous AI recommendation systems is still in pilot phases.
Sector adoption velocity	claude-sonnet-5	2/5	Geoscience and environmental engineering sectors show slower, more cautious AI adoption compared to information/finance sectors, with AI mostly used for data processing rather than strategic decision-making.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can meaningfully assist by automating dust dispersion simulations, analyzing monitoring data, and suggesting mitigation options, allowing geoscientists to focus on strategy selection and site-specific feasibility—a useful productivity boost but not transformative without human oversight.
Augmentation potential	claude-sonnet-5	4/5	AI can significantly help geoscientists by modeling dust dispersion patterns, analyzing sensor/satellite data, and summarizing mitigation literature, meaningfully speeding up the research and planning process while the scientist retains final judgment.
Task automatability	claude-haiku-4-5-20251001	2/5	AI can assist with data analysis and modeling of dust dispersion patterns, but determining effective mitigation strategies requires domain expertise, field assessment, and judgment about trade-offs between environmental, economic, and operational constraints that current systems cannot reliably do end-to-end.
Task automatability	claude-sonnet-5	1/5	This requires original scientific reasoning, field-specific judgment, and synthesis of environmental, geological, and engineering constraints that current AI cannot reliably perform end-to-end without extensive human expert oversight.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Environmental regulations often require qualified geoscientists to sign off on dust mitigation plans, and liability for inadequate mitigation (air quality, health impacts) creates strong legal and organizational barriers to full automation without expert human judgment.
Adoption barriers	claude-sonnet-5	3/5	No strict licensing requirement mandates a geoscientist sign off, but environmental and safety liability, regulatory compliance, and organizational trust in expert judgment create moderate friction against pure AI automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Geoscientists' loaded salaries are substantial, and current AI inference costs plus integration and expert oversight would not yet achieve order-of-magnitude savings; the task requires specialized domain knowledge validation that offsets computational efficiency.
Cost vs. human wage	claude-sonnet-5	2/5	While AI can assist with literature review and data analysis cheaply, the core mitigation strategy development still requires expensive expert geoscientist time and field validation, keeping overall costs comparable to human-led work.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	While AI tools exist for atmospheric modeling and dust simulation, no deployed product autonomously determines mitigation strategies for mineral dust. Products that exist are narrow-scope research or analytical aids, not end-to-end decision systems trusted in production by mining or mineral companies.
Technical feasibility today	claude-sonnet-5	1/5	No deployed products autonomously design mitigation strategies for mineral dust dispersion; this remains a specialized scientific/engineering problem requiring expert judgment and site-specific data collection.

Advise construction firms or government agencies on dam or road construction, foundation design, land use, or resource management.

CI 15–25 · exposure 20 · 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 is moderate and slow: large infrastructure and government agencies are early movers with computational modeling, but advice remains delivered by licensed professionals. Small and mid-sized construction firms lag significantly, and regulatory frameworks continue to require human expert sign-off, limiting displacement.
Sector adoption velocity	claude-sonnet-5	2/5	Construction, geotechnical engineering, and government infrastructure sectors are slow, conservative adopters of AI for core technical judgments, with pilots emerging mainly around data processing rather than advisory replacement.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI strongly augments geoscientist productivity: machine learning accelerates site modeling, remote-sensing data analysis, and regulatory compliance checks, allowing experts to focus on judgment and stakeholder engagement. AI transforms the analytical phase while the human remains the decision authority.
Augmentation potential	claude-sonnet-5	4/5	AI can meaningfully assist by analyzing geological/geospatial data, modeling scenarios, summarizing regulations, and drafting reports, substantially boosting geoscientist productivity while they retain final judgment.
Task automatability	claude-haiku-4-5-20251001	2/5	AI can assist with data analysis, site assessment, and preliminary design recommendations, but the task requires significant expert judgment, stakeholder consultation, and legal/regulatory navigation that current systems cannot reliably perform end-to-end. Site-specific conditions, safety-critical decisions, and client interaction remain fundamentally human-led.
Task automatability	claude-sonnet-5	2/5	This involves synthesizing site-specific geological data, regulatory context, and professional judgment for high-stakes infrastructure decisions, which current AI cannot do end-to-end reliably.dedication It could assist with data analysis but not replace the advisory judgment itself.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Strong legal and liability barriers exist: professional engineering licenses, liability for foundation failure or environmental damage, regulatory approval requirements, and client accountability all require a human geoscientist or licensed engineer to sign off. Liability asymmetry is high—errors in dam or road design have catastrophic consequences.
Adoption barriers	claude-sonnet-5	5/5	Professional geoscientist licensure, engineering liability, and regulatory sign-off requirements for infrastructure projects like dams and roads mean a licensed professional must legally advise and stamp such work.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI modeling and analysis tools reduce some analytical overhead, but the loaded cost of a licensed geoscientist's advice (including site visits, liability, and consultation hours) vastly exceeds current AI inference and integration costs. AI remains a support tool, not a cost replacement.
Cost vs. human wage	claude-sonnet-5	2/5	Human geoscientist judgment and liability exposure make AI substitution costly to insure against, though AI can cheaply handle some data-crunching subcomponents, keeping overall cost comparable to human expertise.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Some AI tools exist for geological modeling and site analysis, but no deployed product reliably performs the full advisory task (design recommendation, regulatory sign-off, stakeholder communication) at production quality. Tools are primarily analytical aids rather than autonomous advisors.
Technical feasibility today	claude-sonnet-5	1/5	No deployed AI product independently advises on dam, road, or foundation design decisions in production; this remains firmly in human expert territory with AI as a data-analysis aid at best.

Measure characteristics of the Earth, such as gravity or magnetic fields, using equipment such as seismographs, gravimeters, torsion balances, or magnetometers.

CI 7–30 · exposure 13 · augmentation 50 · importance 3.8/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	While data analysis tools see adoption, field-based geoscience remains physical-work intensive with slow digitization in smaller firms and less accessible automation. Most organizations still rely on human-led field campaigns, with AI adoption concentrated in post-collection data processing rather than measurement operations themselves.
Sector adoption velocity	claude-sonnet-5	2/5	Geoscience fieldwork sectors are adopting AI mainly for data analysis, not for physical measurement collection, so adoption in this specific task is slow.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI tools meaningfully assist in interpreting large volumes of seismic, gravity, or magnetic data, flagging anomalies and modeling subsurface features. However, the assistance is primarily in analysis phase; field measurement itself sees limited augmentation beyond automated logging and visualization of real-time sensor feeds.
Augmentation potential	claude-sonnet-5	3/5	AI can assist with planning survey routes, calibrating equipment settings, and interpreting readings in real time, but does not perform the physical measurement task.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can analyze and interpret seismic or magnetic field data post-collection, the physical act of deploying, calibrating, and positioning specialized instruments like seismographs and gravimeters in the field requires human presence and real-time judgment. AI cannot currently operate this equipment end-to-end without human technicians managing setup and quality control.
Task automatability	claude-sonnet-5	1/5	This is a physical field/lab measurement task requiring specialized equipment deployment, calibration, and site work; AI cannot perform the physical measurement itself.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Professional licensing (geologist/geoscientist credentials in many jurisdictions), liability concerns for environmental or structural assessment decisions, and regulatory oversight of subsurface investigations create meaningful barriers. Client contracts and industry standards often mandate human professional sign-off on field measurement protocols and interpretation.
Adoption barriers	claude-sonnet-5	3/5	No licensing barrier specifically bars AI, but physical access, equipment operation, and field safety protocols create strong practical barriers to automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Specialized field instruments are capital-intensive and require trained geoscientists to operate and interpret; data analysis software is relatively cheap, but the total system cost per measurement remains comparable to or higher than human technician wages due to equipment maintenance, calibration, and expertise requirements.
Cost vs. human wage	claude-sonnet-5	1/5	AI cannot substitute for the physical equipment operation, so there is no viable cost comparison—human plus instrument cost is required regardless.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Commercial data analysis tools exist for interpreting seismic and magnetic measurements, but no deployed product autonomously performs the full task of measuring field characteristics from equipment operation through interpretation. Data loggers and automated collection systems exist, but they still require human calibration, placement decisions, and field validation.
Technical feasibility today	claude-sonnet-5	1/5	No deployed AI product performs physical geophysical instrument operation; this remains entirely a human/hardware task.

Develop ways to capture or use gases burned off as waste during oil production processes.

CI 7–30 · exposure 13 · augmentation 50 · importance 3.0/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	While oil and gas companies increasingly use AI for reservoir modeling and production optimization, the upstream R&D for novel gas-capture methods remains a slower-adopting, traditionally human-expert-driven process with long regulatory cycles.
Sector adoption velocity	claude-sonnet-5	2/5	Oil and gas engineering is a slower-adopting, capital-intensive physical sector where AI is used for data analysis and modeling but not for autonomous invention of capture technologies.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can meaningfully augment geoscientist work via literature synthesis, thermodynamic simulation, design space exploration, and constraint optimization, raising their productivity in the design loop without replacing the core creative and validation role.
Augmentation potential	claude-sonnet-5	3/5	AI can assist geoscientists with literature review, simulation, data analysis of flare volumes, and design optimization, meaningfully supporting but not replacing the core innovation work.
Task automatability	claude-haiku-4-5-20251001	2/5	This task requires creative engineering design, feasibility analysis, and integration with complex industrial systems. While AI can assist with literature review, modeling, and optimization of existing capture methods, developing novel approaches demands domain expertise, intuition, and iterative physical/chemical reasoning that current AI systems cannot fully execute end-to-end.
Task automatability	claude-sonnet-5	1/5	This is an open-ended engineering R&D and process innovation task requiring novel physical/chemical solutions, field testing, and creative problem-solving that current AI cannot execute end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Oil and gas operations are heavily regulated; any new capture method must comply with environmental, safety, and equipment standards and typically requires engineer sign-off and regulatory approval. Liability and performance-assurance requirements mean human geoscientists must legally own and certify the design.
Adoption barriers	claude-sonnet-5	3/5	While not licensed in the way medicine or law is, engineering judgments tied to safety, environmental regulation, and capital investment decisions typically require professional accountability and sign-off, creating moderate friction.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	The all-in cost of current AI models with sufficient domain context, simulation infrastructure, and human oversight and iteration is not yet substantially cheaper than the cost of a geoscientist's time spent on design research, prototyping, and validation.
Cost vs. human wage	claude-sonnet-5	1/5	AI cannot substitute for the engineering design, prototyping, and field validation work involved, so there is no meaningful cost comparison favoring AI at present.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	No deployed product reliably generates novel gas-capture solutions for oil production on its own. AI tools can support design via simulation and parameter optimization, but the task as stated—developing new ways—requires human geoscientists to conceive, test, and validate solutions; products exist only for supporting subcomponents.
Technical feasibility today	claude-sonnet-5	1/5	No deployed AI product autonomously develops flare gas capture technologies or processes; this remains an engineering design task done by human specialists.

Determine methods to incorporate geomethane or methane hydrates into global energy production or evaluate the potential environmental impacts of such incorporation.

CI 11–25 · exposure 13 · augmentation 63 · importance 2.5/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Methane hydrate energy production remains largely in research and pilot phases globally, with minimal commercial deployment; adoption velocity in geoscience firms is slow, driven by regulatory uncertainty and nascent technology maturity rather than rapid AI-driven transformation.
Sector adoption velocity	claude-sonnet-5	2/5	Energy and geoscience research sectors are only beginning to adopt AI tools for data analysis and modeling; adoption for this specific complex research task is nascent and slow-moving.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can usefully assist geoscientists by automating literature synthesis, running scenario simulations, analyzing seismic or drilling data, and surfacing environmental trade-off analyses, but the human expert must still direct strategy, interpret context-dependent results, and make final judgments on feasibility and risk.
Augmentation potential	claude-sonnet-5	4/5	AI can meaningfully assist by synthesizing literature, running geochemical/reservoir models, and drafting environmental impact sections, significantly speeding up parts of the research process while humans retain final judgment.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can assist with literature review, data analysis, and scenario modeling for methane hydrate feasibility studies, the task requires domain-specific integration of geological, thermodynamic, and environmental knowledge, creative problem-solving on novel technical pathways, and expert judgment on trade-offs that current AI cannot synthesize end-to-end at the 50% time-saving threshold.
Task automatability	claude-sonnet-5	1/5	This requires original scientific research, novel methodological design, and complex environmental/geological judgment that current AI cannot perform end-to-end; it is a discovery and evaluation task, not a routine information task.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Methane hydrate exploitation is heavily regulated at national and international levels (climate commitments, environmental protection) and requires sign-off from licensed geoscientists, environmental agencies, and regulatory bodies before any incorporation into energy policy or production—human expertise and authorization are legally and institutionally mandated.
Adoption barriers	claude-sonnet-5	4/5	Environmental impact assessments often require credentialed geoscientists and are subject to regulatory review, and energy infrastructure decisions carry high liability, creating substantial barriers to full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	The task requires specialized geoscientist expertise with high-value output; while AI can reduce some intermediate analytical work, the integrated analysis and judgment needed still command high-wage human input, making the all-in cost of AI-assisted work comparable to or exceeding solo human effort.
Cost vs. human wage	claude-sonnet-5	2/5	AI can cheaply assist with literature review and data analysis, but the core scientific judgment and fieldwork/interpretation still require expensive specialist geoscientists, so overall cost savings are limited.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	No deployed products reliably perform independent method determination or environmental impact evaluation for methane hydrate integration at production scale; research papers and simulations exist, but end-to-end decision support or recommendation systems are not mature in industry practice.
Technical feasibility today	claude-sonnet-5	1/5	No deployed AI product autonomously determines methane hydrate extraction methods or conducts environmental impact evaluations; this remains firmly in expert research territory.

Plan or conduct geological, geochemical, or geophysical field studies or surveys, sample collection, or drilling and testing programs used to collect data for research or application.

CI 7–25 · exposure 13 · augmentation 50 · importance 4.4/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Geoscience remains heavily field-dependent and human-centric; adoption of AI automation is limited to backend data processing and modeling. Field operations in oil, gas, mining, and environmental sectors still rely on human teams and are slower to digitize than information-sector work.
Sector adoption velocity	claude-sonnet-5	2/5	Geosciences and extractive/field industries are relatively slow to adopt AI for physical operations, though data analysis components see some uptake.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can meaningfully assist by optimizing survey designs, predicting drill-site locations, automating core-log analysis, and processing real-time sensor data from the field, but the human geoscientist remains the primary decision-maker in planning and on-site execution.
Augmentation potential	claude-sonnet-5	3/5	AI can assist in planning survey designs, analyzing geochemical/geophysical data, and optimizing drilling programs, meaningfully supporting the human-led fieldwork.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can assist with planning surveys and analyzing data remotely, the core task requires on-site fieldwork, sample collection, drilling operations, and adaptive decision-making in variable geological conditions that demand physical presence and human judgment. Current systems cannot perform end-to-end field studies with 50% time savings.
Task automatability	claude-sonnet-5	1/5	This requires physical fieldwork, site travel, sample collection, and operating drilling/testing equipment in real-world terrain, none of which current AI systems can perform end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Professional licensing (many jurisdictions require Professional Geoscientist credentials), liability for drilling/sampling errors, safety regulations, and client/regulatory requirements for human geoscientist sign-off on field programs create significant legal and organizational barriers to full automation.
Adoption barriers	claude-sonnet-5	4/5	Field safety, equipment operation permits, environmental regulations, and liability for drilling/testing programs create substantial barriers, though not always strict professional licensure for every step.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Field geoscientists command substantial loaded wages ($75k–$120k+), and the equipment, travel, and human oversight costs of any autonomous field system would remain high relative to deploying experienced personnel for complex geological work.
Cost vs. human wage	claude-sonnet-5	1/5	AI cannot substitute for the physical labor and equipment operation involved, so there is no meaningful AI cost basis to compare against human fieldwork costs.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	AI tools exist for data analysis and survey planning optimization, but no deployed products can autonomously conduct geological field studies, manage drilling programs, or perform physical sampling at scale. Robotics for drilling exist in limited contexts but lack the generality and reliability required across diverse geological environments.
Technical feasibility today	claude-sonnet-5	1/5	No deployed AI product plans or conducts physical field surveys, drilling, or sample collection; this remains firmly human-executed fieldwork.

Inspect construction projects to analyze engineering problems, using test equipment or drilling machinery.

CI 5–25 · exposure 13 · augmentation 50 · importance 4.0/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Construction and geoscience sectors are slower adopters of automation. While some firms use remote sensors and data analytics, the actual on-site inspection with specialized equipment remains human-driven; adoption of autonomous inspection robots in production is minimal.
Sector adoption velocity	claude-sonnet-5	1/5	Construction and field geoscience are physical, low-digitization sectors with minimal AI agent deployment for on-site inspection and drilling work.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can meaningfully assist by processing drilling core data, analyzing sensor readings, or flagging anomalies in subsurface images before the geoscientist reviews them in the field. These augmentations improve efficiency but require human interpretation and on-site equipment operation.
Augmentation potential	claude-sonnet-5	3/5	AI can assist with analyzing test data, modeling subsurface conditions, and generating reports from inspection results, but the physical inspection and machinery operation itself sees little AI augmentation.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can process sensor data and images from construction sites, the task requires hands-on inspection using specialized equipment (test equipment, drilling machinery) and judgment about complex engineering problems that demand real-time, on-site decision-making. Current AI cannot independently operate drilling machinery or perform the physical field inspection component.
Task automatability	claude-sonnet-5	1/5	This requires physical presence at construction sites, hands-on operation of drilling machinery and test equipment, and real-time engineering judgment based on physical inspection—none of which current AI systems can perform.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Strong barriers exist: liability for engineering decisions affects asset safety and public welfare, regulatory requirements often mandate licensed professionals sign off on inspection findings, and client contracts typically require a credentialed geoscientist to certify results. These create high friction against full automation.
Adoption barriers	claude-sonnet-5	4/5	Engineering geology work often requires licensed professional judgment and sign-off, liability for construction safety is high, and physical site access with specialized equipment operation creates strong practical barriers.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	An AI system capable of autonomous site inspection with specialized equipment would require significant hardware integration, on-site deployment, and human oversight. The total cost per inspection would likely exceed or barely match a geoscientist's loaded wage, especially given customization needs for different sites.
Cost vs. human wage	claude-sonnet-5	1/5	There is no AI substitute performing this physical, equipment-operating task, so cost comparison favors the human by default since no viable AI alternative exists.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	AI systems can assist with data analysis from sensors and images, but no deployed product reliably performs independent construction inspection with test/drilling equipment. Real-world deployment requires physical presence, equipment operation, and contextual judgment that remains immature in production systems.
Technical feasibility today	claude-sonnet-5	1/5	No deployed AI product operates drilling machinery or physically inspects construction sites; this remains firmly in the domain of human field geoscientists.

Collaborate with medical or health researchers to address health problems related to geological materials or processes.

CI 7–11 · exposure 0 · 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	Interdisciplinary health-geology research occurs in specialized academic and government institutions that adopt new tools slowly. This work is not digitized like routine tasks, involves small numbers of experts, and lacks the commercial incentives driving AI adoption in other professional sectors.
Sector adoption velocity	claude-sonnet-5	2/5	Academic and geoscience research sectors adopt AI tools slowly for core collaborative research work, though AI use for literature review or data analysis is growing.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI can assist by analyzing geological and health datasets, mining literature for relevant connections, and summarizing existing research, thereby reducing time on information gathering. However, the core task of genuine collaboration and hypothesis generation remains distinctly human-led.
Augmentation potential	claude-sonnet-5	3/5	AI can assist with literature synthesis, data analysis, and drafting communications between disciplines, providing moderate productivity support during collaboration.
Task automatability	claude-haiku-4-5-20251001	1/5	This task requires cross-disciplinary expertise, deep scientific judgment, and sustained collaboration between geoscientists and health researchers to understand complex causal relationships. Current AI systems cannot independently establish novel health-geology connections or lead meaningful scientific partnerships.
Task automatability	claude-sonnet-5	1/5	This is an interdisciplinary collaborative research task requiring novel scientific judgment, field expertise, and human-to-human coordination that current AI cannot perform end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Strong barriers include the requirement for licensed expertise in both geosciences and health fields, institutional credentialing for research collaboration, funding agency requirements for qualified personnel, and liability considerations when health outcomes are involved. Research publication and peer review also require human accountability.
Adoption barriers	claude-sonnet-5	4/5	Scientific research collaboration typically requires credentialed expertise, institutional accountability, and professional judgment, creating strong organizational and epistemic barriers to AI substitution.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	While AI tools for analysis are inexpensive, the high value of expert geoscientist-health researcher collaboration and the specialized oversight required make the total cost-benefit unfavorable compared to retaining skilled professionals for this integrative work.
Cost vs. human wage	claude-sonnet-5	1/5	There is no AI substitute performing this collaborative research role, so no meaningful cost comparison favors AI; human expert collaboration is required.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed product reliably performs interdisciplinary scientific collaboration or serves as a substantive research partner between distinct fields. AI can assist with literature review or data analysis, but cannot replace the judgment and creative problem-solving required to address novel health-geology problems.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product performs cross-disciplinary scientific collaboration between geoscientists and health researchers; this remains firmly in the human research domain.

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.