Chemical Engineers
17-2041.00Median wage $125,040/yr21,070 employed (US)Rank #304 of 923 scored · top 33% by substitution
Design chemical plant equipment and devise processes for manufacturing chemicals and products, such as gasoline, synthetic rubber, plastics, detergents, cement, paper, and pulp, by applying principles and technology of chemistry, physics, and engineering.
Sub-scores
0–100 · band = confidence interval from rater disagreement
Substitution34
Exposure33
Augmentation72
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
14 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%33
panel mean rating 2.3/5 → substitution pressure 33/100
Technical feasibility todayw 20%31
panel mean rating 2.3/5 → substitution pressure 31/100
Cost vs. human wagew 15%32
panel mean rating 2.3/5 → substitution pressure 32/100
Adoption barriersw 20%inverted — strong barriers lower the score38
panel mean rating 3.5/5 (barrier strength) → substitution pressure 38/100
Sector adoption velocityw 10%32
panel mean rating 2.3/5 → substitution pressure 32/100
Task breakdown (14 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.
Prepare estimate of production costs and production progress reports for management.
67CI 60–75 · exposure 70 · augmentation 88 · importance 3.7/5 · click for rater detail
Prepare estimate of production costs and production progress reports for management.
67| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Chemical and process industries have high digital maturity, with widespread ERP and data analytics adoption; report automation is increasingly common in manufacturing and engineering-heavy sectors as organizations modernize operations. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Chemical/process manufacturing is a slower-adopting sector for AI compared to finance or software, with pilots more common than full production deployment for cost/reporting workflows. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI significantly assists engineers by automating data aggregation and preliminary report structure, freeing them to focus on interpretation, anomaly flagging, and strategic recommendations rather than manual compilation. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI tools substantially speed up data compilation, trend analysis, and report drafting, letting engineers focus on interpretation and decision-making while staying in the loop. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Current AI systems can extract data from production logs, calculate cost estimates using standard formulas, and generate structured progress reports with minimal human input. While final validation and strategic interpretation may require human judgment, the core numerical and data-compilation work is highly automatable. |
| Task automatability | claude-sonnet-5 | 4/5 | Cost estimation and progress reporting are largely data aggregation and template-driven writing tasks that AI can perform well when given structured inputs, though verification of source data and domain judgment still requires human oversight.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Organizations typically require human sign-off and audit trails on cost estimates for financial and compliance reasons, and internal processes often mandate engineer review before reports reach management. These governance and oversight requirements create moderate friction but do not legally prevent automation. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement blocks AI from doing cost estimates, but internal governance, data sensitivity, and need for engineer sign-off on numbers used for management decisions create moderate friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | Automated data extraction, calculation, and report generation via AI incurs minimal inference and integration costs compared to the loaded wage of a chemical engineer or analyst spending hours on manual data collection and report assembly. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Once data pipelines are set up, AI-driven report generation and cost modeling is far cheaper per report than engineer hours, though initial integration with plant/ERP systems adds cost. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Mature ERP and business intelligence platforms already integrate with production systems to generate cost reports and dashboards; AI-powered report generation tools are deployed in manufacturing organizations. Some customization and domain-specific thresholds require setup, but the basic capability is production-ready. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Tools like spreadsheet AI add-ins, ERP-integrated analytics, and LLM-based report generators exist and are used, but full end-to-end automation of production cost estimation in chemical plants is not yet a mature, widely deployed product category. |
Monitor and analyze data from processes and experiments.
65CI 55–75 · exposure 62 · augmentation 88 · importance 4.2/5 · click for rater detail
Monitor and analyze data from processes and experiments.
65| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Chemical and petrochemical industries are heavy investors in Industry 4.0, IIoT, and predictive analytics. Real-time process monitoring is a core use case where AI-powered dashboards and anomaly detection are already in wide production deployment. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Chemical/process industries are adopting AI-based analytics and predictive tools at a moderate pace, driven by Industry 4.0 initiatives, but face slower adoption than pure information-sector work due to legacy infrastructure and safety-critical validation needs. |
| Augmentation potential | claude-haiku-4-5-20251001 | 5/5 | AI dashboards, auto-generated reports, and predictive alerts dramatically amplify engineer productivity by surfacing patterns in high-volume data and flagging issues in real time, freeing the engineer to focus on root-cause analysis and decision-making rather than manual data review. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI significantly enhances an engineer's ability to monitor large data streams, detect anomalies, and surface patterns for further human investigation, though the engineer remains essential for interpretation and decision-making. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | AI can automatically ingest sensor streams, detect anomalies, generate real-time alerts, and produce summary reports from experimental data with >50% time savings. However, interpretation of unexpected results and hypothesis formation typically require domain expertise, preventing full end-to-end automation. |
| Task automatability | claude-sonnet-5 | 3/5 | AI can automate significant portions of data monitoring and statistical analysis (anomaly detection, trend analysis, dashboards), but interpreting process chemistry, diagnosing root causes, and making engineering judgments often still requires human expertise and physical plant context. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | While safety-critical industries like chemicals have regulatory oversight and documentation requirements, there is no legal mandate that a licensed professional must personally monitor data—organizations routinely delegate this to technicians with AI support. Integration with existing SCADA and ERP systems creates some friction but is surmountable. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No strict licensing requirement for data monitoring itself, though engineering sign-off may be required for safety-critical process changes stemming from that analysis, creating moderate organizational and liability friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | Once industrial sensors and platforms are installed, AI inference and automated reporting are extremely cheap per monitoring cycle. Annual operating cost of AI monitoring is typically a small fraction of one engineer's salary for equivalent coverage. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Software-based monitoring tools can be cheaper than continuous manual review for routine data, but integration with proprietary sensors/instrumentation, validation, and engineer oversight for critical decisions keep costs comparable to human-augmented workflows. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Mature products (industrial IoT platforms, data analytics suites, and specialized process-monitoring software) reliably perform continuous monitoring and anomaly detection in production chemical plants today. Deployed systems can flag deviations and generate alerts, though human review remains standard. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Process analytics platforms and ML-based monitoring tools (e.g., predictive maintenance, statistical process control software) are deployed in industry, but they typically flag issues for human review rather than fully replacing engineer analysis, and reliability varies across plant types. |
Perform tests and monitor performance of processes throughout stages of production to determine degree of control over variables such as temperature, density, specific gravity, and pressure.
52CI 30–74 · exposure 50 · augmentation 88 · importance 3.6/5 · click for rater detail
Perform tests and monitor performance of processes throughout stages of production to determine degree of control over variables such as temperature, density, specific gravity, and pressure.
52| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Chemical manufacturing and petrochemicals are capital-intensive, highly digitized sectors where process automation and real-time monitoring systems have been standard practice for decades. Adoption of advanced process control and data analytics is rapid in these industries, particularly in large enterprises and continuous-process operations. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Chemical/process manufacturing is a physically intensive, moderately digitized sector where AI adoption for real-time process monitoring is growing but still lags behind information-sector adoption rates. |
| Augmentation potential | claude-haiku-4-5-20251001 | 5/5 | AI-powered dashboards, predictive analytics, and anomaly detection substantially augment engineer productivity by surfacing deviations, predicting failures, and prioritizing attention—allowing engineers to focus on root-cause analysis and optimization rather than routine observation. This is a clear case of transformative augmentation while the human remains in the loop for critical decisions. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-driven analytics, anomaly detection, and predictive maintenance tools significantly enhance an engineer's ability to monitor and interpret process variables, improving efficiency while the engineer retains decision authority. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Monitoring and testing processes at scale can be substantially automated through sensor networks, data logging systems, and anomaly detection algorithms that continuously track temperature, pressure, density, and other variables with minimal human intervention. While interpretation of anomalies and adjustments to control parameters may require human judgment, sensor-based monitoring and automated performance testing can deliver >50% time savings on the observational and data-collection portions of this task. |
| Task automatability | claude-sonnet-5 | 2/5 | Physical testing and monitoring require sensors, sampling, and physical presence at equipment; AI can analyze the resulting data but cannot perform the physical measurement and hands-on monitoring itself today. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Regulatory requirements (FDA, EPA, OSHA, industry standards like ISO 9001) often mandate documented testing and sign-off by qualified engineers, creating oversight requirements that prevent fully autonomous operation. However, these barriers mainly restrict *decision-making authority* rather than the *monitoring and testing* itself, which can be largely automated with human review. |
| Adoption barriers | claude-sonnet-5 | 3/5 | Process safety regulations and quality control requirements in chemical manufacturing often require engineer sign-off and accountability for critical variable control, creating moderate regulatory and liability barriers to full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Once instrumentation and SCADA systems are in place (capital investment amortized), the marginal cost of continuous automated monitoring and data logging is negligible compared to the loaded wage of a chemical engineer performing manual testing and observation rounds. AI-based anomaly detection adds minimal incremental cost over existing infrastructure. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Sensor networks, SCADA/DCS systems, and analytics platforms have significant capital and integration costs comparable to or exceeding engineer labor costs for equivalent monitoring scope, though incremental automation lowers marginal costs over time. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Industrial process control systems, SCADA (Supervisory Control and Data Acquisition) platforms, and real-time monitoring dashboards with automated alerts are mature, deployed products used in production across chemical manufacturing facilities worldwide. These systems reliably track and log process variables continuously, though final decisions on corrective action typically still involve human engineers. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Process control software and predictive analytics are deployed in industry, but full autonomous testing/monitoring across production stages still relies heavily on human oversight and instrumentation integration, not end-to-end AI products. |
Develop computer models of chemical processes.
42CI 39–46 · exposure 45 · augmentation 75 · importance 3.1/5 · click for rater detail
Develop computer models of chemical processes.
42| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Chemical engineering sectors (pharma, energy, materials) are moderately digitized but conservative in adopting AI for core modeling work. Adoption remains in the pilot and early-production phase; most organizations still rely on traditional simulation software (ASPEN, COMSOL) with human-led workflows rather than AI-driven model generation. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Chemical/process engineering firms are moderately digitized with some simulation-software AI features emerging, but adoption of AI-driven modeling remains at the pilot stage rather than widespread production use. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI significantly assists chemical engineers by auto-generating code scaffolds, suggesting reaction pathways, solving differential equations, and flagging dimensional inconsistencies, allowing engineers to focus on validation, domain constraints, and creative problem-solving. This assistive capability meaningfully raises productivity while engineers remain the decision-maker. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools meaningfully accelerate coding, equation derivation, literature review, and debugging of process models, significantly boosting engineer productivity while the engineer retains final modeling judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI can assist in generating boilerplate code, suggesting model structures, and validating equations, but developing robust chemical process models requires domain expertise, validation against empirical data, and iterative refinement that AI cannot reliably perform end-to-end. Significant human oversight and decision-making remain essential, so the time savings fall short of the 50% threshold at equal quality. |
| Task automatability | claude-sonnet-5 | 3/5 | AI can assist in generating process model code, simulation scripts, or kinetics equations, but validating and calibrating models against real plant behavior requires domain expertise and iterative human judgment that current AI cannot fully replace end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Chemical process models often feed into safety-critical and regulatory-approved systems (e.g., manufacturing scales, hazard analysis). Liability and regulatory sign-off requirements mean a qualified chemical engineer must validate and take responsibility for model outputs, creating friction against full automation. |
| Adoption barriers | claude-sonnet-5 | 3/5 | While no formal license mandates engineer sign-off on internal simulations, safety-critical process design often requires PE review and organizational validation processes that resist full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | API-based AI tools are cheap per inference, but integrating them into a chemical engineering workflow demands skilled engineers to validate outputs, tune models, and verify accuracy against experimental data. The total cost of AI + oversight often approaches or exceeds a direct engineer hour for high-stakes models. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI can speed up scripting and boilerplate model setup cheaply, but the overall cost is dominated by expert engineer time for validation, calibration, and safety review, keeping the AI-to-human cost ratio only modestly favorable. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Deployed AI tools (code generation, equation solvers, simulation frameworks) can accelerate parts of model development, but production systems for autonomous end-to-end chemical process modeling are immature. Tools like ChatGPT and GitHub Copilot handle coding scaffolding, yet specialized thermodynamic and kinetic validation still requires expert review. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some engineering copilots and code-generation tools assist with modeling in ASPEN, MATLAB, or Python, but no deployed product autonomously builds validated chemical process models in production without engineer oversight. |
Troubleshoot problems with chemical manufacturing processes.
30CI 25–35 · exposure 30 · augmentation 75 · importance 4.3/5 · click for rater detail
Troubleshoot problems with chemical manufacturing processes.
30| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | While large chemical manufacturers have invested in Industry 4.0 and predictive maintenance pilots, autonomous troubleshooting adoption remains limited. Most deployments are still in early pilot phases; production-level autonomous troubleshooting is rare, reflecting the conservative, safety-conscious nature of the chemical industry. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Chemical manufacturing is a physically-oriented, moderately digitized sector where AI adoption for process control exists but is slower and shallower than in software or finance sectors. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI significantly augments chemical engineers through real-time monitoring dashboards, anomaly alerts, pattern matching against historical faults, and recommended next steps. These tools materially accelerate problem identification and solution synthesis while keeping the engineer in the loop for final validation and safety-critical decisions. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-based analytics, digital twins, and anomaly detection significantly help engineers spot deviations and narrow root causes faster, meaningfully boosting productivity while humans remain responsible for diagnosis and fixes. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Troubleshooting chemical processes requires integrating real-time sensor data, process knowledge, and physical reasoning to diagnose root causes. While AI can assist in pattern matching and anomaly detection on historical data, end-to-end autonomous troubleshooting—identifying root cause and proposing solutions—across the full complexity of manufacturing systems remains beyond current capabilities without significant human oversight. |
| Task automatability | claude-sonnet-5 | 2/5 | Troubleshooting requires integrating sensor data, tacit process knowledge, physical inspection, and hypothesis-driven experimentation that current AI cannot fully replicate end-to-end, though it can assist with data analysis and pattern recognition. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Chemical manufacturing is heavily regulated (EPA, OSHA, industry safety standards), and troubleshooting decisions often carry liability and safety consequences. Process changes typically require sign-off by licensed engineers, and regulatory requirements mandate human responsibility for safe operation, creating strong legal and compliance barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No formal licensing requires a human for troubleshooting, but safety-critical decisions in chemical plants create liability concerns and organizational reluctance to let AI act autonomously without engineer sign-off. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Current industrial AI systems require substantial integration, domain-specific model training, sensor infrastructure, and continuous human oversight. The all-in cost (inference, customization, validation, liability) remains comparable to or higher than employing experienced chemical engineers for troubleshooting, especially given the high cost of errors in chemical manufacturing. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI tools for anomaly detection are cheap to run, but the overall troubleshooting task still requires expensive skilled engineer oversight and physical plant investigation, keeping costs comparable to human-only processes. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | AI-based anomaly detection and diagnostic support systems exist in production (e.g., predictive maintenance platforms, AI monitoring dashboards), but they typically flag issues rather than fully troubleshoot autonomously. Material error rates and the need for human engineers to interpret recommendations and validate solutions limit deployment to augmentation roles rather than independent execution. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some predictive-maintenance and anomaly-detection products are deployed in process industries, but full autonomous troubleshooting of novel manufacturing problems is not reliably performed by any commercial product today. |
Conduct research to develop new and improved chemical manufacturing processes.
29CI 25–32 · exposure 25 · augmentation 75 · importance 3.6/5 · click for rater detail
Conduct research to develop new and improved chemical manufacturing processes.
29| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Adoption of AI in chemical process research is uneven and cautious. While large pharma and petrochemical firms pilot machine learning for optimization, actual displacement of research chemists or process engineers remains minimal; most adoption is augmentative rather than substitutive. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Chemical and process industries are adopting AI for optimization and computational chemistry at a moderate pace, with pilots more common than full production-scale autonomous R&D. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI significantly assists this task through molecular modeling, reaction prediction, literature mining, and process optimization—tools like computational chemistry platforms and machine-learning-accelerated screening markedly enhance researcher productivity while the engineer retains critical judgment on feasibility and innovation direction. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI significantly aids literature review, molecular modeling, data analysis, and process simulation, substantially boosting engineer productivity while humans retain control of experimental design and validation. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can assist with literature review, data analysis, and simulation of chemical processes, developing genuinely novel manufacturing processes requires creative hypothesis generation, physical intuition, and iterative experimental design that current AI systems cannot perform end-to-end. The core inventive and synthesis work remains heavily human-dependent. |
| Task automatability | claude-sonnet-5 | 2/5 | Research design, hypothesis generation, and experimental interpretation for novel chemical processes require deep domain judgment, physical experimentation, and creativity that current AI cannot fully replicate end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Strong barriers exist: regulatory approval of new manufacturing processes requires documented human expert sign-off, intellectual property and patent strategy demand human judgment, and process safety validation typically mandates licensed engineer accountability. These legal and safety requirements protect the role. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No strict licensing requirement for the research itself, but safety, regulatory compliance, and corporate IP/liability concerns create meaningful friction against full automation of novel process development. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Computational chemistry tools and AI-assisted simulation can reduce costs for exploratory phases, but chemical engineers command high salaries, and the overhead of AI infrastructure, validation, and human oversight often approaches or exceeds the labor cost for this complex, custom R&D work. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI can cut some literature review and simulation costs, but expensive lab work, pilot plants, and expert oversight remain necessary, keeping overall cost comparable to human-led R&D teams. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Deployed AI tools exist for molecular simulation and data analysis, but no product reliably performs the full research-to-process-development cycle. Systems like AlphaFold aid protein design, but scaling to novel chemical manufacturing process innovation with production reliability remains largely at the research-demonstration stage. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | AI tools (e.g., molecular simulation, literature synthesis, process optimization software) assist parts of R&D but no deployed product autonomously conducts full process development research reliably. |
Develop safety procedures to be employed by workers operating equipment or working in close proximity to ongoing chemical reactions.
25CI 25–25 · exposure 25 · augmentation 63 · importance 4.4/5 · click for rater detail
Develop safety procedures to be employed by workers operating equipment or working in close proximity to ongoing chemical reactions.
25| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Chemical and petrochemical sectors adopt AI slowly for safety-critical tasks due to regulatory compliance burden, risk aversion, and the requirement for licensed engineering oversight. Pilot projects in process optimization are more common than AI-driven safety procedure automation in production. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Chemical/process engineering is a moderately digitized but safety-critical, heavily regulated sector where AI adoption for safety-critical documentation remains cautious and slow. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist by generating hazard checklists, searching relevant standards and past procedures, and flagging overlooked equipment or scenarios, meaningfully accelerating the chemical engineer's drafting and review process. However, the human expert must remain the decision-maker and validator. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can meaningfully speed up drafting, referencing regulations, and structuring safety documents, letting engineers focus on validation and site-specific judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can draft hazard analyses and generate procedure templates based on chemical data, the task requires domain expertise, legal/regulatory judgment, and accountability for worker safety that current systems cannot reliably provide end-to-end. Human chemical engineers must validate, contextualize, and sign off on final procedures. |
| Task automatability | claude-sonnet-5 | 2/5 | Drafting safety procedures requires synthesizing plant-specific hazards, chemistry, equipment configurations, and regulatory requirements, which AI can assist with but cannot reliably originate end-to-end without expert validation. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Strong regulatory and legal barriers exist: OSHA, EPA, and industry standards require that safety procedures be developed and approved by qualified professionals, with documented accountability. Liability exposure for inadequate procedures creates asymmetric error costs that deter full automation and require human sign-off. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Safety procedures in chemical plants are subject to OSHA/EPA regulations and liability concerns, typically requiring a qualified engineer's review and approval before implementation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI-assisted drafting may save 20–30% of engineering time on initial procedure writing, but human experts must still review, test, and validate everything. The loaded cost of a chemical engineer remains substantially higher than inference costs, and the human cannot be removed from the loop without unacceptable risk. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI drafting assistance is cheap, but the overall task still requires substantial paid engineer review, hazard analysis, and sign-off, keeping total cost comparable to human-led work. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | AI tools exist for hazard identification and procedure drafting (e.g., LLMs trained on safety literature), but no deployed product reliably generates complete, legally defensible safety procedures without substantial expert review and modification. Current systems lack the integration of equipment specs, site-specific conditions, and liability-aware compliance needed. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | No deployed product autonomously generates certified safety procedures for chemical operations; LLMs can draft templates but engineers must verify technical accuracy and site-specific hazards. |
Evaluate chemical equipment and processes to identify ways to optimize performance or to ensure compliance with safety and environmental regulations.
25CI 25–25 · exposure 25 · augmentation 75 · importance 4.2/5 · click for rater detail
Evaluate chemical equipment and processes to identify ways to optimize performance or to ensure compliance with safety and environmental regulations.
25| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Adoption of AI in chemical engineering remains modest relative to software or finance; most firms use monitoring tools but retain human engineers for evaluations. Pilots are common in large petrochemical firms, but production-scale displacement of evaluation tasks is rare; organizational friction and liability concerns slow deployment. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Chemical/process industries are historically slower to adopt AI at scale compared to information-sector fields, with digital transformation in physical plants proceeding cautiously due to safety-critical operations. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can significantly assist with real-time process monitoring, anomaly flagging, literature search for optimization strategies, and compliance checklist generation, allowing engineers to focus on judgment and design trade-offs. Augmentation is already substantial in leading firms using ML dashboards and simulation support. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-driven process simulation, anomaly detection, and predictive maintenance tools meaningfully boost engineers' ability to identify optimization opportunities and flag compliance issues, even though final decisions remain human-led. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can assist with data analysis, sensor monitoring, and literature review for optimization strategies, evaluating complex chemical processes requires domain expertise, risk assessment, and judgment about safety trade-offs that current AI systems cannot reliably perform end-to-end. Compliance verification often involves novel regulatory contexts and nuanced technical judgment beyond automated benchmarking. |
| Task automatability | claude-sonnet-5 | 2/5 | AI can assist with data analysis and identifying optimization opportunities using process data, but the physical inspection, judgment on safety compliance, and integration of tacit engineering knowledge require substantial human involvement, falling short of 50% end-to-end automation. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Safety and environmental compliance create hard barriers: engineers must document decisions, sign off on process changes, and bear liability for failures. Regulatory frameworks often explicitly require licensed engineer involvement, and error costs in chemical safety are extremely high, strongly limiting pure automation. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Safety and environmental compliance work is heavily regulated, often requiring licensed professional engineer sign-off and accountability, creating strong legal and liability barriers to full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI inference for monitoring is cheap, but integration into existing systems, tuning domain-specific models, and mandatory human review significantly raise the all-in cost. For comprehensive evaluation work, the overhead still approaches or exceeds the loaded cost of a chemical engineer's time on this task. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Specialized process optimization software and data analytics platforms carry significant licensing, integration, and validation costs, and still require expert oversight, so savings versus a chemical engineer's time are modest rather than order-of-magnitude. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Some narrow products exist for real-time monitoring dashboards and anomaly detection, but no deployed system reliably conducts the full evaluation task (design critique, regulatory compliance assessment, optimization pathway selection) without significant human oversight. Most deployed tools handle data collection and flagging, not decision-making. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Process simulation and predictive analytics tools exist and are used in some plants, but comprehensive evaluation of equipment and regulatory compliance is still largely performed by engineers with AI as a supplementary tool, not a reliable standalone product. |
Design and plan layout of equipment.
25CI 25–25 · exposure 25 · augmentation 63 · importance 3.8/5 · click for rater detail
Design and plan layout of equipment.
25| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Chemical engineering firms use CAD and simulation tools but have not broadly adopted autonomous or agentic design systems; adoption remains limited to pilot projects and vendor-specific parametric libraries rather than widespread AI-driven displacement. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Chemical/process engineering and heavy industry sectors are slower AI adopters compared to software or finance, with AI design tools still in pilot or assistive stages rather than widespread production use. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can usefully augment engineers by generating layout alternatives, checking compliance against rule libraries, and automating routine drafting, but the engineer remains responsible for validation, optimization, and final design decisions. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-assisted CAD, simulation, and optimization tools can meaningfully speed up iteration, clash detection, and preliminary layout generation, helping engineers explore options faster while retaining final judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Current AI can assist with 2D/3D layout visualization and suggest standard configurations, but equipment design requires iterative engineering judgment, constraint satisfaction across multiple domains (thermal, mechanical, safety), and validation against domain-specific standards that AI systems cannot reliably perform end-to-end with ≥50% time savings at equal quality. |
| Task automatability | claude-sonnet-5 | 2/5 | Equipment layout design requires integrating safety codes, spatial constraints, process flow, and site-specific factors that current AI cannot fully synthesize end-to-end; AI can assist with drafting and calculations but not autonomously produce a validated layout. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Equipment layout designs must be signed off by licensed Professional Engineers in most jurisdictions, and liability for thermal, pressure, and safety failures attaches to the responsible engineer, creating hard legal and accountability barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Process safety, environmental regulations, and professional engineering licensure requirements mean a qualified engineer typically must review and sign off on layouts involving hazardous materials or regulated processes. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI-assisted tools (CAD plugins, layout suggestions) still require significant specialized engineer oversight, review, and correction, making the all-in cost (inference + integration + expert validation) comparable to or exceeding unassisted engineer labor. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI tools can reduce some drafting/calculation time but licensed engineering review, safety verification, and iterative design still require significant human engineering hours, keeping costs comparable to human-led work. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | CAD and parametric design tools exist, but no deployed AI system independently produces validated chemical equipment layouts meeting regulatory and operational requirements; research prototypes and demos exist, but production-scale autonomous design remains absent. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | CAD and simulation tools have AI-assisted features (e.g., generative layout suggestions, clash detection) but no deployed product reliably performs full equipment layout planning without heavy engineer oversight and iteration. |
Determine most effective arrangement of operations such as mixing, crushing, heat transfer, distillation, and drying.
25CI 25–25 · exposure 25 · augmentation 63 · importance 3.4/5 · click for rater detail
Determine most effective arrangement of operations such as mixing, crushing, heat transfer, distillation, and drying.
25| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Chemical engineering and process design remain relatively conservative sectors with long project cycles and strong reliance on domain expertise. Adoption of AI-driven process synthesis is in pilot phases; most major decisions still rely on human engineers and established simulation tools. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Chemical/process engineering is a physically-grounded, capital-intensive sector with slower digitization and cautious adoption of AI for core design decisions compared to information-sector functions. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI-powered simulation and optimization tools can assist engineers by rapidly exploring design alternatives and flag trade-offs, but the human engineer remains central to interpreting results, managing constraints, and making final decisions. Productivity gain is meaningful but not transformative. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-driven simulation, optimization algorithms, and generative design tools meaningfully speed up scenario testing and identify promising configurations, substantially aiding engineers who retain final decision authority. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | AI can assist with optimization models and simulations of unit operations, but determining 'most effective arrangement' requires balancing competing constraints (safety, cost, throughput, product quality), domain expertise, and context-specific factors that current systems struggle with end-to-end. Significant human oversight and validation remain necessary. |
| Task automatability | claude-sonnet-5 | 2/5 | This requires deep process design judgment integrating physical chemistry, safety, and economics across a plant; current AI can assist with simulation and optimization but cannot autonomously determine full process arrangements reliably end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Professional licensing (P.E.) is common for chemical engineers, process safety regulations (OSHA PSM, EPA) govern equipment arrangement, and liability for process failures is significant. A licensed engineer must typically sign off on process designs; regulatory compliance and safety responsibility create hard barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Process design decisions affecting safety, environmental compliance, and capital investment typically require professional engineer sign-off and regulatory review, creating strong liability and licensing barriers. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Simulation licenses and AI tools are expensive; integrating recommendations into engineering workflows requires skilled interpretation. Current tools do not reduce the cost of hiring a chemical engineer to validate and refine process designs. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI tools can reduce iteration time in simulations, but the cost of validating safety-critical process designs and integrating engineering judgment keeps overall costs comparable to or only modestly cheaper than skilled engineer time. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Process simulation software (ASPEN, COMSOL) exists and can model individual operations, but no deployed product reliably recommends optimal process flow architecture end-to-end. Research systems show promise in narrow domains, but production deployment for process synthesis remains limited. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Process simulation software (Aspen, etc.) with AI-assisted optimization exists, but no deployed product autonomously designs full unit operation sequences without expert engineer oversight and validation. |
Design measurement and control systems for chemical plants based on data collected in laboratory experiments and in pilot plant operations.
25CI 25–25 · exposure 25 · augmentation 75 · importance 3.3/5 · click for rater detail
Design measurement and control systems for chemical plants based on data collected in laboratory experiments and in pilot plant operations.
25| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Chemical manufacturing is moderately digitized but remains conservative in process automation due to safety criticality. AI adoption in design roles is limited to assistive tools; few organizations deploy autonomous design systems in production environments. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Chemical engineering and process industries are historically slower adopters of AI-driven design automation compared to software or finance sectors, with pilots more common than production deployment. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can significantly augment engineers by rapidly analyzing laboratory and pilot-plant data, generating simulation scenarios, and proposing design variants. Engineers retain critical review and validation, making AI a powerful assistant for productivity gains on data-intensive portions of design work. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI and simulation tools can meaningfully assist in analyzing experimental/pilot data, generating design options, and modeling control systems, significantly boosting engineer productivity while humans retain final design authority. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can assist with data analysis and generate draft control system designs from experimental data, the task requires judgment about safety margins, equipment interactions, and plant-specific constraints that demand human expertise. End-to-end automation with 50% time savings at equal quality is not yet achievable with current systems. |
| Task automatability | claude-sonnet-5 | 2/5 | This requires synthesizing experimental data, engineering judgment, safety considerations, and physical plant constraints that current AI cannot autonomously integrate into a validated control system design end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Process safety regulations (OSHA PSM, IEC 61511 functional safety standards) typically require licensed or highly experienced engineers to certify control system designs. Liability for equipment failures and safety incidents creates strong legal barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Chemical plant control system design involves significant safety, regulatory, and liability considerations that typically require licensed professional engineer sign-off, creating strong barriers to full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Current AI tools (data analysis, simulation software) have meaningful licensing and integration costs. When combined with required human expert validation and oversight, the total cost per complete system design remains comparable to or exceeds the cost of a chemical engineer's time. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | The engineering judgment, validation, and liability involved mean AI can reduce some analysis time but still requires substantial human engineering effort, keeping costs comparable rather than dramatically lower. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | AI tools exist for data analysis and simulation-based design support, but no deployed product reliably performs the full integrated task of designing measurement and control systems for chemical plants independently. Commercial CAD and simulation tools require substantial human oversight and validation. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | AI tools can assist with data analysis and control system simulation, but no deployed product autonomously designs measurement and control systems for chemical plants from lab/pilot data without extensive engineer oversight. |
Develop processes to separate components of liquids or gases or generate electrical currents, using controlled chemical processes.
25CI 20–30 · exposure 20 · augmentation 75 · importance 3.4/5 · click for rater detail
Develop processes to separate components of liquids or gases or generate electrical currents, using controlled chemical processes.
25| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Chemical engineering is a traditional, highly regulated sector where adoption of automation is cautious and driven by regulatory constraints. While computational tools see steady adoption, end-to-end process automation and AI agents replacing process development remain nascent, with most firms in pilot or early evaluation phases. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Chemical/process engineering is a physical, moderately-digitized sector with slower AI adoption compared to software or finance; AI use is mostly in pilots for modeling and optimization rather than full process design. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI significantly augments chemical engineers through molecular simulation, reaction prediction, process optimization, and parameter sensitivity analysis, allowing faster exploration of design space. Engineers remain in the loop for validation, scale-up, and safety decisions, making this a high-value augmentation scenario. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI and simulation tools (process modeling, molecular design assistance, data analysis) meaningfully speed up parts of the design cycle like literature review, simulations, and optimization, even though humans still drive core innovation. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can optimize process parameters and simulate chemical reactions computationally, the task requires hands-on experimentation, physical pilot testing, and iterative refinement with real materials that current systems cannot perform end-to-end. AI assists in modeling but cannot autonomously design, build, and validate separation or electrical generation processes to meet the >50% time-saving threshold. |
| Task automatability | claude-sonnet-5 | 2/5 | This is fundamentally creative process engineering requiring novel chemistry design, experimentation, and physical validation that current AI cannot perform end-to-end; AI can assist with calculations and simulations but not replace the core R&D work. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Professional licensure (PE/Professional Engineer certification), regulatory compliance (EPA, OSHA for chemical processes), liability for process safety, and the requirement for documented engineering sign-off on designs create strong barriers. Process development must be certified by licensed engineers, preventing full automation. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing requirement specifically blocks AI use, but engineering sign-off, safety regulations, and liability for process design changes create meaningful organizational and regulatory friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI-assisted tools reduce modeling time but do not eliminate the cost of lab work, equipment, and senior engineer time for design, validation, and troubleshooting. Integration and oversight remain labor-intensive relative to the task's scale, making the overall cost comparable to or higher than traditional human engineering workflows. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | While AI tools can cheaply run simulations or literature reviews, the actual process development still requires expensive human expertise, lab work, and pilot testing, so overall cost savings are limited. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Deployed products exist for chemical simulation and process optimization (e.g., ASPEN, computational chemistry tools with ML modules), but these are narrow aids rather than end-to-end process developers. No production system autonomously designs and validates novel separation or electrochemical processes; human expertise remains essential for experimental design and interpretation. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product autonomously develops novel separation or electrochemical processes; this remains research-stage assistance (simulation tools, literature search) rather than a production capability. |
Perform laboratory studies of steps in manufacture of new products and test proposed processes in small-scale operation, such as a pilot plant.
16CI 7–25 · exposure 13 · augmentation 63 · importance 3.3/5 · click for rater detail
Perform laboratory studies of steps in manufacture of new products and test proposed processes in small-scale operation, such as a pilot plant.
16| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Chemical engineering and pilot plant operations remain low-velocity adopters of full automation due to the physical nature of the work, regulatory constraints, and the premium on human expertise for novel and high-risk processes. Adoption of AI tools for simulation and data analysis is growing, but replacement of the core experimental task is minimal. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Chemical manufacturing and process engineering are capital-intensive, physically-oriented sectors with historically slower AI adoption compared to information-based industries, though some digital modeling tools are emerging. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI substantially assists chemical engineers today through process modeling, literature mining, experimental design recommendations, data analysis, and failure prediction, meaningfully accelerating the design and iteration cycles of pilot plant work while humans retain critical control and safety decisions. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can assist with experimental design, data analysis, simulation of chemical processes, and predictive modeling to guide pilot plant testing, improving efficiency of the human-led process. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can model chemical processes and analyze data from lab reports, hands-on experimental design, physical manipulation of equipment, real-time troubleshooting of pilot plant failures, and safety-critical decision-making during novel synthesis require human judgment and physical presence that current systems cannot fully automate. Only preliminary hypothesis generation and some data analysis could be meaningfully automated. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires physical manipulation of equipment, hands-on lab work, and pilot plant operation—physical actions AI cannot perform; only ancillary data analysis or planning could be assisted. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Significant regulatory barriers exist: EPA, OSHA, and industry-specific regulations require licensed or formally trained personnel to design, conduct, and sign off on process safety and environmental compliance testing. Liability for failed processes and safety incidents creates strong legal requirements for human responsibility. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Chemical engineering work often requires licensed PE oversight, safety protocols, and regulatory compliance for pilot-scale chemical processes, creating substantial institutional and legal barriers to automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | The task requires expensive specialized equipment, controlled facilities, and hazmat handling that cannot be replaced by software. AI infrastructure for modeling assists but does not substitute for the capital and labor costs of running pilot-scale operations with trained chemists. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | AI cannot substitute for the physical experimentation and equipment operation involved, so there is no viable AI-driven cost alternative for the core task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed product reliably performs end-to-end laboratory testing of novel manufacturing processes in pilot plants. AI exists for process modeling and data analysis but cannot operate lab equipment, handle safety protocols, or make real-time adjustments to unexpected chemical behaviors in a production environment. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product operates pilot plants or conducts physical laboratory experiments autonomously; this remains firmly in the physical/human domain. |
Direct activities of workers who operate or are engaged in constructing and improving absorption, evaporation, or electromagnetic equipment.
9CI 5–14 · exposure 8 · augmentation 38 · importance 3.1/5 · click for rater detail
Direct activities of workers who operate or are engaged in constructing and improving absorption, evaporation, or electromagnetic equipment.
9| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Chemical manufacturing and construction remain heavily regulated, safety-critical sectors where human supervisors are mandated by law. Adoption of autonomous worker direction is negligible and legally constrained. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Chemical process/construction environments are physical, low-digitization settings with minimal AI-driven displacement of on-site supervisory roles. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could assist with equipment performance monitoring, predictive maintenance alerts, or scheduling optimization, but the core task of directing workers in real-time requires human judgment and presence; augmentation potential is limited to backend support functions. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can assist with scheduling, documentation, monitoring equipment data, and communication support, but the core in-person directing of workers is not significantly transformed. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Directing workers in physical construction and equipment operations requires real-time on-site oversight, safety judgment, and adaptive coordination that AI cannot perform end-to-end today. While AI could assist with scheduling and documentation, the core supervisory and safety-critical aspects remain human-dependent. |
| Task automatability | claude-sonnet-5 | 1/5 | This is in-person supervision and coordination of physical construction/equipment-building workers, which current AI cannot execute end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Legal and regulatory frameworks (OSHA, worker safety statutes) require a qualified human supervisor to direct construction and equipment operations. Liability for worker safety creates a hard requirement for licensed engineer sign-off and on-site authority. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Supervisory responsibility for safety-critical equipment construction typically requires engineering accountability, liability, and often licensure (PE oversight), creating strong barriers to automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The cost of AI systems capable of real-time site supervision, safety oversight, and worker coordination would exceed the loaded wage of a chemical engineer supervisor, particularly when factoring in liability and backup human oversight. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | AI cannot substitute for the human supervisory role here, so there is no viable AI cost comparison—human labor remains the only option. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI product reliably directs worker activities on construction sites or in equipment operation environments. This requires physical presence, dynamic problem-solving, and accountability that current systems cannot handle in production. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product directs on-site construction workers or manages physical equipment-building crews; this remains a human management function. |
Related occupations — Architecture & Engineering
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.