Chemists
19-2031.00Median wage $91,240/yr82,770 employed (US)Rank #326 of 923 scored · top 35% by substitution
Conduct qualitative and quantitative chemical analyses or experiments in laboratories for quality or process control or to develop new products or knowledge.
Sub-scores
0–100 · band = confidence interval from rater disagreement
Substitution33
Exposure31
Augmentation65
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
12 rated tasks, binned by substitution score.
Position among all scored occupations
Distribution of 923 occupation scores; the marker is this occupation.
Tasks with substitution ≥ 70
0%
Run 1.0.0-draft.1 · computed 2026-08-05 · rater panel: claude-sonnet-5, claude-haiku-4-5-20251001 · intervals span rater disagreement.
Why this score
The five weighted dimensions of the composite, averaged across this occupation's tasks (importance-weighted, panel mean). Exact weights and formulas: /api/v1/methodology.
Task automatabilityw 35%31
panel mean rating 2.2/5 → substitution pressure 31/100
Technical feasibility todayw 20%31
panel mean rating 2.2/5 → substitution pressure 31/100
Cost vs. human wagew 15%30
panel mean rating 2.2/5 → substitution pressure 30/100
Adoption barriersw 20%inverted — strong barriers lower the score40
panel mean rating 3.4/5 (barrier strength) → substitution pressure 40/100
Sector adoption velocityw 10%33
panel mean rating 2.3/5 → substitution pressure 33/100
Task breakdown (12 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.
Purchase laboratory supplies, such as chemicals, when supplies are low or near their expiration date.
63CI 51–75 · exposure 62 · augmentation 75 · importance 3.4/5 · click for rater detail
Purchase laboratory supplies, such as chemicals, when supplies are low or near their expiration date.
63| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Laboratory and research organizations in the information/professional services sectors have broadly adopted digital inventory systems and automated reordering; chemical suppliers themselves now offer API-driven ordering platforms. Adoption is well beyond pilot stage in large institutions and increasingly in mid-size labs. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Lab environments and chemical R&D settings are generally slower adopters of AI automation compared to purely digital sectors, though inventory management systems are gradually being adopted in larger labs. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI-driven inventory alerts and supplier-cost comparison tools meaningfully enhance a chemist's ability to manage supplies efficiently, reducing time spent manually checking stock and negotiating orders while maintaining oversight over compliance and special requirements. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-enabled inventory and procurement systems can flag low stock and expiring chemicals, significantly reducing the manual tracking burden on chemists while they retain final purchasing decisions. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | AI systems can monitor inventory levels and expiration dates via data integration, generate purchase orders, and route them to suppliers with minimal human intervention. However, some discretion around supplier selection and budget approval typically requires human sign-off, limiting the achievable time savings to near but not quite 50% end-to-end without manual oversight. |
| Task automatability | claude-sonnet-5 | 3/5 | Ordering supplies based on inventory levels and expiration tracking is a rule-based procurement task that software can largely handle, though selecting correct specifications and vendor qualification may need human review. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Some organizational policies require human sign-off on purchases above certain thresholds or from new suppliers; regulatory traceability in chemical handling adds modest friction. However, no strict licensing requirement prevents a chemist or administrator from fully automating routine reorder workflows, making barriers moderate rather than high. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement for purchasing itself, but institutional procurement policies, budget approval chains, and safety/compliance considerations for chemical purchases create moderate organizational friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | Automated procurement via AI systems costs significantly less than employing a person to manually track inventory, check expiration dates, compare suppliers, and submit orders. The operational cost of automation is typically one-tenth to one-fifth the loaded labor cost for this routine task. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated inventory/reorder systems are cheap to run compared to a chemist's time spent manually checking stock and expiration dates, though integration and initial setup cost some investment. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Mature inventory management systems and procurement platforms (many with built-in AI/automation) already perform purchase ordering at scale in chemical and laboratory environments. Integration with existing lab management systems is well-established, though exception handling (urgent shortages, regulatory compliance checks) may require human review. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Inventory management and procurement software with automated reorder triggers exist and are deployed in labs, but full automation of chemical-specific purchasing (correct grades, hazard compliance, vendor selection) still typically involves human confirmation. |
Compile and analyze test information to determine process or equipment operating efficiency or to diagnose malfunctions.
52CI 50–55 · exposure 50 · augmentation 75 · importance 3.9/5 · click for rater detail
Compile and analyze test information to determine process or equipment operating efficiency or to diagnose malfunctions.
52| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Process analytics and condition monitoring tools are in pilot and limited production use across chemicals, pharma, and manufacturing, but full automation of diagnostic decision-making remains uncommon; most deployments augment rather than replace chemist analysis. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Process industries have moderate digitization and are increasingly using predictive maintenance and process analytics tools, but adoption is slower and more pilot-stage compared to fully digital sectors like finance. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI excels at rapid compilation of large datasets, trend spotting, and suggesting hypotheses for malfunction diagnosis. A chemist using these tools can dramatically accelerate the analytical phase and narrow the search space for root cause, maintaining their judgment while being substantially more productive. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-driven analytics and visualization tools significantly speed up compiling and interpreting test data, helping chemists identify trends and potential issues faster, even though final diagnosis often remains human-driven. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI can automate data compilation and pattern recognition in test results, and can flag anomalies or suggest likely equipment issues, but diagnosis of complex malfunctions often requires domain expertise, physical intuition, and contextual knowledge that current systems struggle with. A human chemist could save roughly 40-60% of time on the analytical portions, particularly routine data processing and trend identification. |
| Task automatability | claude-sonnet-5 | 3/5 | AI can compile and statistically analyze structured test data and flag anomalies, but diagnosing equipment malfunctions often requires physical inspection, tacit process knowledge, and judgment that current systems cannot fully replicate end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Some regulatory oversight applies in GxP environments (pharma, food, chemicals) where process validation and equipment maintenance decisions require qualified human sign-off, though diagnosis assistance itself is not legally restricted. Many organizations prefer human accountability in safety-critical process decisions, creating adoption friction. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No strict licensing requirement mandates a human chemist sign off on internal process diagnostics, though quality/safety protocols and organizational accountability create some friction to full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 3/5 | The all-in cost of AI tools (sensor integration, software, model training, human oversight) is roughly comparable to a chemist's loaded wage when applied at scale to routine monitoring and analysis tasks, but integration overhead and required human validation reduce the advantage. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Automated data analysis tools reduce time spent on compilation and trend analysis, but the need for human oversight, domain expertise, and system integration keeps costs roughly comparable to a chemist's efforts for full diagnostic tasks. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Products exist (data analytics platforms, anomaly detection tools, some process analytics software) that can compile test data and identify patterns, but they often require significant configuration, human validation, and domain expertise to interpret correctly. Reliability remains materially below human chemists in complex or novel failure modes. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Data analytics and anomaly-detection products (e.g., process historians with ML modules) are deployed in chemical plants, but they typically require significant customization and human interpretation, with narrow reliability for diagnosing novel malfunctions. |
Write technical papers or reports or prepare standards and specifications for processes, facilities, products, or tests.
40CI 30–50 · exposure 42 · augmentation 63 · importance 3.9/5 · click for rater detail
Write technical papers or reports or prepare standards and specifications for processes, facilities, products, or tests.
40| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Chemistry and pharma sectors are conservative; while some labs use AI drafting aids, widespread production deployment of AI-authored standards and peer-reviewed papers is rare—adoption remains primarily in lower-stakes internal documentation. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Chemical and scientific industries are adopting AI writing tools for documentation and reports, but adoption in regulated technical/safety documentation lags behind faster-moving sectors like finance or generic knowledge work. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can usefully assist by suggesting structures, filling in boilerplate, and generating first drafts, but chemists retain critical roles in interpreting experiments, ensuring correctness, and making judgments; augmentation is real but partial, not transformative. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI substantially speeds up drafting, formatting, literature summarization, and language polishing for technical papers and specifications, while chemists retain responsibility for scientific accuracy and final content. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Current AI can draft sections of technical papers and suggest report structures, but chemists must verify experimental details, interpret results, ensure accuracy, and make critical judgments about what findings mean—tasks requiring domain expertise and accountability that AI cannot fully own. Time savings rarely reach 50% for publication-ready work. |
| Task automatability | claude-sonnet-5 | 3/5 | AI can draft substantial portions of technical writing from provided data and outlines, but accurate scientific content, correct standards citation, and interpretation of experimental results still require significant human expertise and verification. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Standards, specifications, and published technical papers carry legal and professional liability; chemists and engineers often must personally attest to accuracy and sign off on standards—strong regulatory and organizational barriers prevent full AI autonomy in many contexts. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing requirement mandates a human chemist sign every report, but organizational liability, quality control, and professional accountability for technical accuracy create meaningful friction against pure AI authorship. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI inference is cheap, but the overhead of expert review, revision, and liability management means the total cost per publication-ready report remains comparable to or exceeds hiring a junior chemist or technical writer to draft and refine the work. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI drafting reduces time spent on structuring and initial writing, but the need for expert chemist review and validation of technical content keeps overall costs only moderately below fully human-authored output. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | AI writing assistants (ChatGPT, Claude) exist and can generate text, but deployed chemistry-specific tools for standards and specifications are limited; chemistry papers require precise terminology, accurate citations, and error-checking that current AI still struggles with at scale in production environments. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | LLM-based writing assistants are widely deployed for drafting technical documents, but for specification-grade chemistry documents with regulatory or safety implications, error rates in technical accuracy remain a real concern requiring expert review. |
Conduct quality control tests.
33CI 25–41 · exposure 38 · augmentation 75 · importance 3.9/5 · click for rater detail
Conduct quality control tests.
33| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Pharmaceutical and manufacturing sectors are piloting AI-driven lab automation and data analytics, but widespread production deployment of autonomous QC remains limited; adoption is faster in large enterprises with high-volume standardized testing than in smaller or more bespoke operations. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Chemical and pharmaceutical manufacturing sectors adopt automation steadily but cautiously due to validation and regulatory requirements, with AI-driven QC still largely pilot-stage. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI systems demonstrably assist chemists by automating data reduction, flagging anomalies, and generating preliminary reports, significantly accelerating review cycles and reducing manual data entry while chemists focus on interpretation, investigation, and decision-making. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools assist chemists significantly in data analysis, anomaly detection, and trend spotting from QC results, improving throughput while humans retain testing and judgment responsibilities. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | Much of the analytical component (spectroscopy, chromatography interpretation, data analysis) can be automated by AI-assisted laboratory systems and software, but physical sample preparation, instrument calibration, and decision-making on out-of-spec results still require human oversight, preventing full end-to-end automation at the 50% time-saving threshold. |
| Task automatability | claude-sonnet-5 | 2/5 | Quality control testing involves physical sample handling, instrument operation, and calibrated judgment about anomalies that current AI cannot perform end-to-end; only data analysis/interpretation portions are automatable. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Regulatory frameworks (FDA, ISO, GMP) often require documented human review and sign-off on quality test results, and liability for failed batches or safety incidents creates strong pressure to retain human accountability, making full substitution legally and organizationally difficult. |
| Adoption barriers | claude-sonnet-5 | 4/5 | QC testing in regulated industries (pharma, food, chemicals) often requires certified analysts and documented sign-off under GMP/ISO standards, creating strong regulatory and liability barriers to full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Initial investment in lab automation software and integration overhead is substantial, and quality assurance liability means human review still adds cost; AI-assisted QC may eventually be cost-neutral or favorable, but is not yet an order of magnitude cheaper than skilled chemist oversight. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Automated analyzers and robotics can reduce per-sample cost at scale, but capital investment, calibration, and human oversight keep overall costs comparable to skilled technician labor for most labs. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Deployed products exist for automated data analysis, trend detection, and report generation in lab settings, but most QC workflows still rely on technicians for hands-on execution and human verification; no single mature product fully automates the entire QC test cycle reliably at scale. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Lab automation and LIMS software with AI-assisted analytics exist, but they handle discrete sub-steps (instrument control, data flagging) rather than the full QC test workflow reliably in production. |
Direct, coordinate, or advise personnel in test procedures for analyzing components or physical properties of materials.
33CI 7–59 · exposure 33 · augmentation 63 · importance 3.6/5 · click for rater detail
Direct, coordinate, or advise personnel in test procedures for analyzing components or physical properties of materials.
33| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Pharmaceutical, chemical, and materials labs have pilot AI systems for data analysis and procedure recommendation, but widespread autonomous coordination of test personnel remains nascent; adoption is accelerating in large organizations but has not achieved mainstream production deployment. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Chemical and materials testing labs adopt digital tools slowly for physical procedures and personnel management, with AI adoption mostly limited to data analysis, not supervision. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can significantly assist chemists by automating data interpretation, flagging anomalies in test results, and recommending next steps, leaving the chemist to make judgment calls on complex materials and personnel decisions. This substantially raises chemist productivity while maintaining human authority. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help draft protocols, generate training materials, or analyze test data to support decisions, but the core direction and coordination of personnel remains human-led. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | AI can autonomously direct and coordinate standard test procedures, generate analysis reports, and advise on component characterization using established protocols and data. However, novel material properties or complex troubleshooting still benefit from human chemist oversight, preventing a full 5 rating. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a supervisory and advisory role requiring in-person direction, judgment about lab safety, and hands-on troubleshooting that current AI cannot perform end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Lab protocols often require professional credentialing and accountability; liability for incorrect test procedures or material characterization recommendations creates oversight friction. Regulatory bodies (FDA, EPA) may mandate human sign-off on certain analyses, creating moderate adoption friction. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Directing lab personnel often involves safety oversight, quality certification, and accountability structures tied to credentialed chemists, creating strong organizational and regulatory barriers. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 3/5 | AI-driven lab management and analysis cost roughly equivalent to employing a mid-level chemist for routine direction and coordination; specialized chemistry AI and human oversight remain necessary, offsetting pure labor cost savings. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI substitute performing this supervisory task, so cost comparison favors the human entirely; AI cannot replace the managerial function. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Deployed chemistry AI tools (spectroscopy interpretation, HPLC data analysis, lab management software) can handle routine test coordination and basic advisory functions, but production systems typically require human validation and are narrower in scope than the full task statement implies. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product manages personnel or directs physical lab test procedures; this remains a human management function with no production AI analog. |
Confer with scientists or engineers to conduct analyses of research projects, interpret test results, or develop nonstandard tests.
29CI 25–32 · exposure 25 · augmentation 63 · importance 3.8/5 · click for rater detail
Confer with scientists or engineers to conduct analyses of research projects, interpret test results, or develop nonstandard tests.
29| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Chemistry and materials science research remain relatively conservative in automation adoption. While AI aids data processing in some labs, the core task of interpersonal scientific conferencing and nonstandard test design is still primarily human-driven; pilot programs exist but production displacement is minimal. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | R&D-intensive sectors (pharma, chemicals, materials) are piloting AI for data analysis and hypothesis generation, but deep production-level adoption for collaborative scientific judgment tasks remains limited. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can meaningfully assist chemists by rapidly summarizing literature, preprocessing data, and suggesting statistical or analytical approaches for test result interpretation. However, the collaborative and novel aspects of the task limit how much AI augmentation can transform productivity without human judgment. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can significantly assist chemists by rapidly analyzing test data, suggesting statistical methods, and helping draft nonstandard test protocols, boosting productivity while humans retain interpretive and collaborative roles. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can help interpret some test results and suggest analyses, the core task requires domain expertise, judgment about novel or nonstandard tests, and back-and-forth professional collaboration that current systems cannot reliably conduct end-to-end. The task explicitly involves problem-solving for nonstandard situations, which falls outside typical AI automation capability. |
| Task automatability | claude-sonnet-5 | 2/5 | This task centers on collaborative discussion, expert judgment, and novel test design, which current AI cannot conduct end-to-end; AI can support analysis and literature synthesis but not replace the conferring and creative test-development process.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Professional licensing (chemists and engineers often credentialed), liability for incorrect test interpretation, regulatory oversight of test validity in regulated industries, and organizational requirements that human experts sign off on novel methods create strong adoption barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 3/5 | There's no strict licensing requirement for this specific task, but scientific credibility, peer accountability, and organizational reliance on expert judgment create moderate friction against full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Current AI assistance (language models, data analysis tools) reduces time on routine analysis interpretation but cannot eliminate the need for expert review and decision-making. Integration costs and human oversight remain substantial, making all-in cost comparable to or exceeding that of senior scientists performing these tasks directly. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Because the task requires expert human interaction and novel experimental design, AI only reduces some analytical labor while human oversight and consultation costs remain high, keeping overall cost comparable to or only modestly better than human-only work. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed product reliably conducts full scientific collaboration or nonstandard test development. AI can assist with data interpretation and literature synthesis, but coordinated scientific conference between chemists and engineers to decide on novel approaches remains largely research-stage rather than production-ready. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | AI tools exist for data analysis and literature search, but no deployed product autonomously confers with scientists or designs nonstandard tests reliably in production settings. |
Develop, improve, or customize products, equipment, formulas, processes, or analytical methods.
28CI 25–30 · exposure 25 · augmentation 75 · importance 4.2/5 · click for rater detail
Develop, improve, or customize products, equipment, formulas, processes, or analytical methods.
28| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Adoption of AI in chemistry is emerging but slow in production settings; most use is confined to early-stage research and academic labs or to specific cheminformatics tasks. Pharma and materials companies pilot AI but retain human chemists as decision-makers due to regulatory and risk constraints. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Chemical and materials R&D sectors are adopting AI tools (e.g., generative chemistry, ML property prediction) but adoption remains pilot-stage rather than deeply embedded in production workflows compared to fully digitized sectors. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI substantially augments chemist productivity: retrosynthesis engines, molecular property predictors, and literature mining tools assist in hypothesis generation, experimental design, and optimization. Chemists remain in the loop but can explore vastly larger solution spaces and reduce trial-and-error cycles. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI significantly aids chemists via literature synthesis, predictive modeling, formulation optimization, and data analysis, meaningfully accelerating parts of the R&D cycle while humans retain overall control and judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | AI can assist with literature review, formula ideation, and process optimization simulations, but development and customization require hands-on experimentation, tacit domain judgment, and iterative physical testing that current systems cannot perform end-to-end. Meaningful automation would require closed-loop lab integration that is not yet standard. |
| Task automatability | claude-sonnet-5 | 2/5 | Product/process development requires experimentation, physical validation, and creative synthesis of tacit and domain knowledge that current AI cannot execute end-to-end, though AI can assist with ideation and data analysis subcomponents. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Strong barriers exist: regulatory approval (FDA, EPA, consumer safety standards) typically mandate human expert sign-off on product formulations and processes. Liability for product failure, patent strategy, and GxP compliance requirements create hard friction against full automation. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No strict licensing requirement blocks AI use, but liability, IP protection, safety validation, and regulatory approval processes for new chemical products create meaningful friction against pure AI-driven development. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI inference costs are low, but integration with lab equipment, validation against regulatory standards, and mandatory human expert oversight add significant overhead. The human chemist's cost remains competitive because the task genuinely requires skilled judgment and experimental iteration. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI can cheaply generate candidate ideas or simulations, but the actual R&D cycle still requires costly physical testing, equipment, and expert oversight, keeping overall cost comparable to or only modestly below human-driven R&D. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | AI tools exist for molecular property prediction and reaction planning (e.g., Chematica, retrosynthesis models), but they remain research-oriented or narrowly scoped; production systems in chemistry labs typically use AI for data analysis or literature mining rather than autonomous product development. Reliability gaps and liability concerns prevent widespread deployment for critical formulation work. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some AI tools exist for molecule/material generation and lab automation, but they operate narrowly and require significant human expert validation and iteration; no deployed system autonomously develops full formulas or processes. |
Prepare test solutions, compounds, or reagents for laboratory personnel to conduct tests.
28CI 25–30 · exposure 25 · augmentation 50 · importance 3.9/5 · click for rater detail
Prepare test solutions, compounds, or reagents for laboratory personnel to conduct tests.
28| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Laboratory chemistry remains a relatively low-digitization, physical domain. While larger pharmaceutical firms deploy robotic handlers, most chemistry labs rely on human technicians. Adoption of AI or automation for routine solution preparation is limited to well-funded research and pharmaceutical settings, not mainstream industry practice. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Lab automation adoption is growing in pharma and biotech but remains slow and uneven across the broader chemical industry, with most labs still relying on manual reagent prep. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI could assist chemists by calculating precise formulations, flagging potential safety concerns, or automating documentation of preparation steps. Augmentation tools for recipe optimization and compliance checking are feasible, though they represent partial rather than transformative productivity gains for this fundamentally hands-on task. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can assist with calculating concentrations, generating protocols, and flagging safety issues, but the physical preparation itself sees limited direct AI augmentation. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Preparing test solutions involves precise measurement, mixing, and procedural steps that current AI systems struggle with in physical laboratory settings. While recipe-following and calculation aspects could be automated, the hands-on manipulation, quality verification, and real-time adjustments for chemical reactions require human oversight and physical dexterity that AI cannot reliably handle end-to-end today. |
| Task automatability | claude-sonnet-5 | 2/5 | Physical preparation of solutions and reagents requires manual manipulation of chemicals, glassware, and equipment that current AI cannot perform without robotic embodiment, which is not standard in most labs.hard core physical work.rating remains low. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | This task faces substantial regulatory and safety barriers: laboratory work is often covered by OSHA regulations, GLP standards, and pharmaceutical/chemical industry compliance requirements that mandate documented human responsibility and sign-off. Liability for contamination or incorrect preparation creates high error-cost asymmetry, and institutional requirements for trained personnel approval are significant. |
| Adoption barriers | claude-sonnet-5 | 3/5 | Safety protocols, chemical handling regulations, and quality control requirements create moderate friction, though not a strict licensing requirement for the physical act itself. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Specialized laboratory robots and AI systems for compound preparation are capital-intensive and require significant integration costs. The loaded cost of deploying and maintaining such systems exceeds the wage cost of trained chemistry technicians for most organizations, especially for lower-volume preparation tasks. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Lab automation equipment (liquid handlers, robotic synthesis) is costly to purchase and integrate, often exceeding or matching human technician costs except in very high-volume settings. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed AI system can reliably prepare chemical test solutions without human intervention. While robotic liquid handlers exist in some labs, they require extensive programming and human oversight; general-purpose AI systems cannot autonomously execute this task in production chemistry environments with consistent quality. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Automated liquid handling robots exist for high-throughput labs, but general reagent/compound preparation across varied chemist tasks is not reliably handled by deployed AI products at scale. |
Evaluate laboratory safety procedures to ensure compliance with standards or to make improvements as needed.
27CI 25–29 · exposure 25 · augmentation 63 · importance 3.7/5 · click for rater detail
Evaluate laboratory safety procedures to ensure compliance with standards or to make improvements as needed.
27| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Adoption of AI for safety evaluation remains limited; most labs rely on traditional human audits and checklists, with slow movement toward digitized safety management despite sector digitization, reflecting conservative risk posture. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Lab safety compliance in chemistry sectors (academia, manufacturing, pharma) is not a fast AI-adoption area; most safety audits remain manual and human-led with limited digitization. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist by flagging procedural gaps, cross-referencing against standard databases, and organizing safety data for human review, meaningfully improving the efficiency of human-led evaluations without replacing the expert judgment. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can meaningfully assist by summarizing regulations, flagging document inconsistencies, generating checklists, and tracking compliance history, improving efficiency while a human chemist retains inspection and sign-off responsibility. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can analyze written safety procedures and flag deviations from standards through text comparison and document review, the task fundamentally requires human judgment about practical lab conditions, risk assessment, and contextual safety considerations that current systems cannot reliably evaluate end-to-end. |
| Task automatability | claude-sonnet-5 | 2/5 | AI can help review written safety documents and cross-check against regulatory checklists, but true evaluation requires physical inspection of lab conditions, equipment, and practices that current AI cannot perform.the task demands on-site judgment and hands-on verification. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Laboratory safety compliance is heavily regulated and typically requires a licensed chemist or safety officer to sign off on evaluations; legal liability for safety failures and industry standards mandate human professional responsibility, creating strong legal barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Safety compliance often requires sign-off by qualified personnel accountable for regulatory adherence (OSHA, EPA, institutional biosafety), creating liability and authorization barriers against full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 3/5 | AI-assisted document review could reduce manual review costs to near-parity with human labor, but complete integration with inspection workflows and expert oversight still requires significant human involvement, keeping costs roughly comparable. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI could cheaply assist with document review portions, but the physical inspection and expert judgment components still require a qualified chemist's time, keeping overall cost savings modest. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed products reliably evaluate lab safety procedures holistically; AI systems can assist with document analysis and standard-matching but lack the embodied understanding and real-world lab inspection capabilities that compliance evaluations demand. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some compliance-checking software and LLM-based document review tools exist, but no deployed product autonomously conducts full lab safety evaluations including physical inspection and improvement recommendations at scale. |
Analyze organic or inorganic compounds to determine chemical or physical properties, composition, structure, relationships, or reactions, using chromatography, spectroscopy, or spectrophotometry techniques.
26CI 21–30 · exposure 25 · augmentation 75 · importance 4.1/5 · click for rater detail
Analyze organic or inorganic compounds to determine chemical or physical properties, composition, structure, relationships, or reactions, using chromatography, spectroscopy, or spectrophotometry techniques.
26| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Adoption remains slow in most chemistry sectors. Academic labs and some pharmaceutical firms pilot AI-assisted spectral interpretation, but production deployment is limited. Physical constraints (need for chemists on-site) and regulatory caution have kept adoption below mainstream levels in the broader chemical workforce. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Chemistry and materials science labs are moderate adopters of AI for data analysis but physical lab work and regulatory environments make deep, fast adoption slower than in purely digital fields. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI significantly augments chemist productivity by rapidly suggesting peak assignments, predicting likely structures, and flagging anomalies in spectra, allowing chemists to focus on novel interpretation and experimental design. Tools like structure prediction from spectral data meaningfully accelerate the analytical cycle while chemists retain critical decision-making. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools substantially speed up spectral pattern recognition, structure elucidation, and data analysis, meaningfully boosting chemist productivity while the chemist still designs experiments and validates results. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can analyze spectroscopy and chromatography data (interpreting peaks, suggesting structures), the full task requires physical sample handling, instrument operation, and interpretation of complex chemical behavior that depends on real-world experimental context. Current AI cannot autonomously run the instruments or troubleshoot failures, limiting time savings to perhaps 20-30% of the overall workflow. |
| Task automatability | claude-sonnet-5 | 2/5 | The physical execution of chromatography/spectroscopy requires lab instrumentation and sample handling that AI cannot perform end-to-end; AI can assist with data interpretation but not the hands-on analytical workflow.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Regulatory requirements (FDA, EPA compliance in pharmaceuticals and environmental work), liability for accuracy in published results, and the professional judgment required for novel compound analysis create strong barriers. Most organizations require a licensed chemist to sign off on analytical results, preventing pure substitution. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing mandate requires a human specifically, but liability for erroneous analytical conclusions (e.g., in pharma, forensics, safety testing) and organizational quality-control protocols create meaningful friction against full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | AI inference for spectral analysis is cheap, but integration into laboratory workflows, validation by chemists, and the retained need for skilled human chemists to design experiments and interpret edge cases means the all-in cost per task remains higher than or comparable to employing a chemist directly. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Instrumentation, sample prep, and calibration costs dominate; AI software adds only marginal savings on interpretation, so overall cost is still comparable to or higher than a trained chemist's labor for full task completion. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Products exist for spectral interpretation (e.g., ChemDraw predictors, some vendor software for peak identification), but they operate on pre-processed data and require expert validation. No deployed system reliably performs end-to-end structure determination or reaction analysis without substantial human oversight and manual verification. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Deployed AI tools exist for spectral interpretation and pattern matching (e.g., NMR/MS prediction software) but no product autonomously runs and interprets full analytical workflows reliably in production without chemist oversight. |
Maintain laboratory instruments to ensure proper working order and troubleshoot malfunctions when needed.
21CI 7–35 · exposure 13 · augmentation 50 · importance 3.9/5 · click for rater detail
Maintain laboratory instruments to ensure proper working order and troubleshoot malfunctions when needed.
21| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Adoption of AI-assisted maintenance monitoring is slow in academic and small-to-mid laboratory settings; larger pharmaceutical and biotech companies are early in deploying predictive maintenance, but human technicians remain essential and displacement is minimal to date. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Chemistry labs and scientific R&D settings adopt AI for data analysis and literature review faster than physical maintenance tasks, which remain largely manual and slow to change. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI-driven diagnostic suggestions and maintenance scheduling can meaningfully assist technicians by flagging potential failures and optimizing service intervals, but the human expert remains central to repair execution, decision-making, and regulatory sign-off. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-based predictive maintenance software and diagnostic decision-support tools can help identify likely causes of malfunction or predict failure timing, aiding the human who still performs physical troubleshooting and repair. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can assist with diagnostic logic and procedural documentation, maintaining and troubleshooting laboratory instruments requires hands-on physical work, real-time sensor interpretation, and judgment calls about part replacement that current AI cannot perform autonomously. The task is heavily dependent on direct manipulation and on-site assessment. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a physical, hands-on task involving equipment calibration, cleaning, part replacement, and hardware diagnosis that current AI cannot perform end-to-end; it requires physical manipulation of instruments in a lab.Software-based AI has no direct means of executing this. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Strong regulatory and compliance barriers exist: laboratory instruments in regulated settings (pharma, clinical) require certified technicians and documented maintenance procedures; equipment calibration and validation must often meet ISO/FDA standards and cannot be delegated to AI systems. Liability for instrument failure and data integrity creates legal friction against full automation. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement specifically bars AI from this task, but the inherent physical nature of instrument repair and troubleshooting creates a strong practical barrier to automation rather than a regulatory one. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | The cost of AI-powered monitoring systems plus human technician labor for physical repair work typically exceeds or matches the cost of direct human technician maintenance, given the current state of robotics and remote diagnostics in lab settings. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | AI cannot replace the physical labor and hands-on diagnostic skill required, so there is no viable AI cost comparison—humans remain the only option for physical maintenance. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Some deployed diagnostic and monitoring software exists, but no current AI system reliably performs end-to-end instrument maintenance or troubleshooting without human intervention. Predictive maintenance tools exist in narrow domains, but they still require human technicians for repair execution and validation. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product performs physical instrument maintenance and troubleshooting; at best, IoT sensors or diagnostic software can flag anomalies, but a human must still physically inspect and repair equipment. |
Induce changes in composition of substances by introducing heat, light, energy, or chemical catalysts for quantitative or qualitative analysis.
19CI 7–30 · exposure 13 · augmentation 50 · importance 4.0/5 · click for rater detail
Induce changes in composition of substances by introducing heat, light, energy, or chemical catalysts for quantitative or qualitative analysis.
19| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Adoption of AI-driven reaction automation remains concentrated in large pharmaceutical and materials companies; most academic and small-to-mid-size labs continue manual chemical induction, indicating slow and narrow uptake relative to the breadth of chemistry practice. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Chemistry R&D and lab work adopt AI mainly for data analysis and design suggestions, but physical experimentation automation (lab robotics) remains a slow-moving, capital-intensive niche in this sector. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI tools assist chemists through automated data collection, reaction parameter optimization suggestions, and spectroscopic interpretation, but the chemist must still make decisions about reaction conditions, safety precautions, and when to stop or modify the process. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can assist with reaction planning, predicting outcomes, literature review, and analyzing resulting data, meaningfully aiding the chemist even though it cannot perform the physical manipulation itself. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can control reaction parameters and interpret spectroscopic data, inducing chemical changes requires physical manipulation of laboratory equipment, precise timing decisions, and real-time observation of reaction progress that current AI systems cannot reliably execute end-to-end without human oversight. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a physical laboratory manipulation involving actual equipment, reagents, and hands-on experimental technique that AI cannot perform without robotic embodiment; current AI cannot physically induce or execute chemical reactions. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Strong regulatory oversight of chemical procedures (OSHA, EPA, GLP compliance), safety liability for exothermic or hazardous reactions, and institutional requirements that licensed chemists validate analytical methods create substantial legal and organizational barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 3/5 | While no licensing law mandates a human perform bench chemistry, safety protocols, equipment access, and physical dexterity requirements create substantial organizational and practical barriers to any automated substitution today. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Specialized automated reaction equipment is capital-intensive and requires significant setup; integrating AI to control these systems and interpret results adds cost that often exceeds the labor of an experienced chemist performing the task manually. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | AI cannot substitute for the physical execution of this task, so cost comparison favors human chemists or specialized lab automation hardware, which is expensive and task-specific, not general AI. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Deployed lab automation exists for narrow reaction sequences (e.g., high-throughput synthesis platforms), but general-purpose AI systems cannot reliably induce compositional changes across diverse chemical scenarios with the safety and accuracy required in production chemistry labs. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product performs physical chemical synthesis or reaction induction end-to-end; automated lab robotics exist only in narrow research settings, not as generally available products for this task. |
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.