Materials Engineers

17-2131.00
Median wage $112,860/yr22,770 employed (US)Rank #471 of 923 scored · top 51% by substitution

Evaluate materials and develop machinery and processes to manufacture materials for use in products that must meet specialized design and performance specifications. Develop new uses for known materials. Includes those engineers working with composite materials or specializing in one type of material, such as graphite, metal and metal alloys, ceramics and glass, plastics and polymers, and naturally occurring materials. Includes metallurgists and metallurgical engineers, ceramic engineers, and welding engineers.

Sub-scores

0–100 · band = confidence interval from rater disagreement

Substitution28
Exposure25
Augmentation70

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

21 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%26

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

Technical feasibility todayw 20%25

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

Cost vs. human wagew 15%24

panel mean rating 2.0/5 → substitution pressure 24/100

Adoption barriersw 20%inverted — strong barriers lower the score39

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

Sector adoption velocityw 10%27

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

Task breakdown (21 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.

Replicate the characteristics of materials and their components, using computers.

61

CI 46–75 · exposure 62 · augmentation 88 · importance 3.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510014/5Materials engineering and product development sectors show active and accelerating adoption of AI-assisted simulation, digital twins, and automated optimization in large firms and advanced manufacturing, though adoption remains uneven across company sizes.
Sector adoption velocityclaude-sonnet-53/5Materials engineering and manufacturing sectors are adopting computational modeling and AI-assisted simulation steadily, but adoption is uneven and often pilot-stage compared to fully digitized industries.
Augmentation potentialclaude-haiku-4-5-202510015/5AI tools dramatically augment engineer productivity by automating parameter sweeps, generating design variants, and accelerating convergence on material specifications, while the engineer interprets results and directs exploration toward novel solutions.
Augmentation potentialclaude-sonnet-54/5AI and advanced simulation substantially speed up materials characterization and predictive modeling, allowing engineers to explore more design variations and reduce physical testing cycles.
Task automatabilityclaude-haiku-4-5-202510014/5Computational modeling and simulation of material properties is largely automatable today using FEA, molecular dynamics, and AI-driven parameter optimization. However, full end-to-end automation from requirement specification through validation against real-world testing still requires significant human expertise for setup, interpretation, and refinement.
Task automatabilityclaude-sonnet-53/5AI/computational tools can perform significant portions of materials simulation and property replication, but complex multi-physics modeling still requires substantial human setup, validation, and domain expertise to ensure accuracy.'
Adoption barriersclaude-haiku-4-5-202510012/5Few regulatory or legal barriers exist to automating computational modeling itself, though critical applications (aerospace, medical devices) may require human engineer sign-off on results. Organizational practices and data access remain the primary friction points.
Adoption barriersclaude-sonnet-53/5No strict licensing requirement for simulation work itself, but engineering sign-off and validation are often required for safety-critical materials applications, creating moderate friction.
Cost vs. human wageclaude-haiku-4-5-202510014/5Computational infrastructure and software licenses are substantial but typically cost far less than the equivalent human engineer labor time for extensive iterative modeling and simulation cycles, especially for routine materials analysis.
Cost vs. human wageclaude-sonnet-52/5Specialized simulation software and computational infrastructure remain costly, and skilled engineers are still needed to set up and validate models, keeping the cost advantage over human labor modest.
Technical feasibility todayclaude-haiku-4-5-202510014/5Mature tools (ANSYS, COMSOL, LAMMPS, and commercial ML frameworks) reliably perform material characteristic replication in production at scale, though material complexity and novel compositions may require manual calibration and validation by experts.
Technical feasibility todayclaude-sonnet-53/5Materials simulation software (e.g., DFT, FEA, molecular dynamics tools) is mature and widely deployed, but AI-driven automation of full characteristic replication still requires expert configuration and interpretation, limiting full reliability.

Write for technical magazines, journals, and trade association publications.

61

CI 45–76 · exposure 58 · augmentation 88 · importance 3.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510013/5Materials engineering and technical publishing occupy a mixed-adoption space: some firms and journals actively pilot AI drafting tools, but many high-prestige publications and traditional materials science journals maintain conservative stances on AI authorship, slowing industry-wide adoption.
Sector adoption velocityclaude-sonnet-53/5Engineering and technical writing fields have moderate AI tool adoption for drafting and editing assistance, though full publication authorship by AI remains uncommon and this is a small, infrequent task within the occupation.
Augmentation potentialclaude-haiku-4-5-202510015/5AI offers transformative productivity gains for this task: engineers can dictate ideas or outlines, receive polished drafts for editing, and iteratively refine content while maintaining full control and subject-matter judgment. This is a textbook augmentation case where human expertise is elevated rather than replaced.
Augmentation potentialclaude-sonnet-54/5AI is highly useful for drafting outlines, improving clarity, summarizing technical data, and polishing prose, meaningfully speeding up the writing process while the engineer retains responsibility for technical accuracy.
Task automatabilityclaude-haiku-4-5-202510014/5Current AI can generate technical content with minimal human input, conducting research, drafting articles, and structuring complex materials science concepts into publication-ready prose. However, the requirement for domain-specific accuracy, peer review credibility, and the need for human verification on cutting-edge claims prevents a full 5 rating.
Task automatabilityclaude-sonnet-53/5LLMs can draft technical articles from provided data and outlines, but subject-matter accuracy, novel insights, and domain-specific claims for materials engineering still require substantial expert review and revision.
Adoption barriersclaude-haiku-4-5-202510013/5While there are no hard licensing barriers preventing AI authorship, professional norms around byline attribution, peer-review credibility, and journal editorial policies requiring human accountability create meaningful friction. Some publications and professional associations restrict or flag AI-assisted content.
Adoption barriersclaude-sonnet-52/5No licensing requirement to write these articles, but journals/trade publications expect author expertise and credibility, and factual errors carry reputational risk, creating moderate friction against pure AI authorship.
Cost vs. human wageclaude-haiku-4-5-202510015/5AI inference and integration costs for generating technical articles are negligible relative to the loaded wage of a materials engineer spending hours drafting, researching, and revising manuscripts. A single AI pass can reduce human authoring time by 60–80%, making the cost-per-output ratio heavily favoring automation.
Cost vs. human wageclaude-sonnet-53/5AI drafting reduces time spent on structuring and initial prose, lowering cost somewhat, but expert fact-checking and technical validation still require engineer time, keeping costs roughly comparable to fully human effort for quality output.
Technical feasibility todayclaude-haiku-4-5-202510014/5Large language models and specialized technical writing tools are deployed in production environments and demonstrably produce draft technical articles and journal submissions. Many engineering firms use AI for initial drafting and outlining; however, final editorial oversight and byline responsibility remain human-handled in most production settings.
Technical feasibility todayclaude-sonnet-52/5General-purpose AI writing tools exist and are used for drafting assistance, but no deployed product reliably produces publication-ready technical/trade articles in materials science without heavy expert editing.

Perform managerial functions, such as preparing proposals and budgets, analyzing labor costs, and writing reports.

41

CI 28–55 · exposure 38 · augmentation 75 · importance 3.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510013/5Mid-sized organizations are piloting AI for report generation and budget template assistance, but widespread production adoption for managerial functions remains limited; many enterprises still prefer human-authored proposals for stakeholder confidence and accountability reasons.
Sector adoption velocityclaude-sonnet-53/5Engineering and manufacturing sectors have moderate AI adoption for office/administrative tasks like reporting and budgeting, with pilots common but full production deployment less mature than in finance or software.
Augmentation potentialclaude-haiku-4-5-202510014/5AI excels at accelerating proposal and budget drafting through template completion, data synthesis, and iteration, materially increasing manager productivity while keeping the manager in control of strategy, accuracy, and sign-off. This is one of the more natural augmentation use cases.
Augmentation potentialclaude-sonnet-54/5AI writing assistants and data analysis tools significantly speed up drafting proposals, reports, and cost breakdowns, letting engineers focus on judgment and strategic content while staying in the loop.
Task automatabilityclaude-haiku-4-5-202510013/5AI can assist with drafting proposals, budgets, and reports by processing data and templates, but managerial functions require contextual business judgment, stakeholder alignment, and sign-off that necessitate human oversight. Roughly half the work—data compilation, initial drafting, formatting—can be automated; the other half requires human decision-making.
Task automatabilityclaude-sonnet-52/5AI can draft portions of proposals, reports, and budget analyses, but integrating engineering judgment, negotiating priorities, and final decision-making requires human oversight, so full end-to-end automation is not yet achievable at equal quality.
Adoption barriersclaude-haiku-4-5-202510014/5Managerial proposals and budgets typically require sign-off by authorized human managers; liability for errors, regulatory compliance (especially in regulated industries), and organizational governance structures create strong friction against full automation. Human accountability is often legally or contractually mandated.
Adoption barriersclaude-sonnet-52/5There's no licensing requirement for these managerial writing/analysis tasks, but organizational approval processes and accountability for budget accuracy create moderate friction against full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI inference is cheap, but integration with organizational systems, data validation, and mandatory human review oversight add labor costs that approach or sometimes exceed the cost of a human drafting from scratch for complex, bespoke proposals and budgets.
Cost vs. human wageclaude-sonnet-53/5AI tools can reduce drafting and analysis time cheaply, but the need for human review, domain-specific data integration, and accountability keeps overall cost roughly comparable to a human doing the full task well.
Technical feasibility todayclaude-haiku-4-5-202510013/5Products like ChatGPT, Claude, and business intelligence platforms can generate proposal drafts and budget summaries in production, but accuracy and appropriateness still require human review. Error rates in financial figures or strategic framing remain material, particularly for high-stakes budgets.
Technical feasibility todayclaude-sonnet-52/5Generic AI writing and spreadsheet tools assist with drafting and calculations, but no deployed product reliably performs the full managerial workflow of proposal/budget preparation and labor cost analysis for engineering contexts without heavy human editing.

Analyze product failure data and laboratory test results to determine causes of problems and develop solutions.

37

CI 30–44 · exposure 38 · augmentation 75 · importance 4.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Materials engineering remains relatively traditional; digitization and AI adoption in this sector lag finance and software, with most organizations still using spreadsheets and domain expertise rather than deployed AI failure analysis systems at scale.
Sector adoption velocityclaude-sonnet-52/5Manufacturing and materials engineering sectors are slower AI adopters compared to information/finance, though some analytics tools are being piloted.
Augmentation potentialclaude-haiku-4-5-202510014/5AI excels at rapidly surfacing patterns, anomalies, and correlations in large failure datasets, enabling materials engineers to focus on causal reasoning and solution design; this assistive role substantially amplifies human productivity without removing human judgment from critical decisions.
Augmentation potentialclaude-sonnet-54/5AI can strongly assist by rapidly correlating failure data, flagging patterns in test results, and suggesting hypotheses, significantly speeding up the engineer's analysis process.
Task automatabilityclaude-haiku-4-5-202510013/5AI can automate significant portions of data analysis, pattern detection, and hypothesis generation from failure data and test results, but developing and validating solutions typically requires materials science expertise and domain judgment that current systems cannot reliably provide end-to-end without human oversight.
Task automatabilityclaude-sonnet-52/5Root-cause analysis of material failures requires synthesizing physical test data, domain expertise, and often physical inspection/experimentation that current AI cannot fully replicate end-to-end.assumptions.
Adoption barriersclaude-haiku-4-5-202510013/5Professional liability and safety concerns (failures can affect product safety) create moderate friction; organizations typically require engineering sign-off on root-cause determinations and solution feasibility, limiting full automation despite technical capability.
Adoption barriersclaude-sonnet-53/5No licensing requirement specifically for this task, but liability for faulty engineering conclusions and reliance on physical evidence create meaningful oversight requirements.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI inference for data processing and pattern analysis is cheap, but the overhead of integration into domain-specific workflows, validation of causal inferences, and expert human review means total cost per solved failure case remains high relative to a materials engineer's time on routine cases.
Cost vs. human wageclaude-sonnet-52/5Engineers must validate AI-assisted analysis with physical testing and domain judgment, so AI reduces but does not eliminate the substantial human cost, keeping cost savings modest.
Technical feasibility todayclaude-haiku-4-5-202510013/5ML systems can perform data analysis and anomaly detection on laboratory results with reasonable accuracy, and some products exist for failure analysis dashboards, but no deployed product reliably performs the full causal analysis and solution development without substantial human expert validation and guidance.
Technical feasibility todayclaude-sonnet-52/5AI tools exist for data analysis and anomaly detection in failure data, but no deployed product reliably performs full failure-analysis-to-solution workflows in production engineering settings.

Determine appropriate methods for fabricating and joining materials.

30

CI 30–30 · exposure 25 · augmentation 63 · importance 4.0/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Materials engineering remains a traditional domain with strong craft expertise; while simulation and data tools are growing, full automation of method determination is still in pilot and research phases. Adoption is slow relative to information-sector tasks.
Sector adoption velocityclaude-sonnet-52/5Manufacturing and materials engineering sectors adopt digital tools more slowly than software/finance, with AI use concentrated in pilot programs rather than widespread production decision-making.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist engineers by rapidly screening literature, suggesting candidate methods, predicting material behavior via models, and accelerating design exploration—meaningfully raising productivity. However, the human engineer must validate and make final judgment calls, placing this in mid-range augmentation.
Augmentation potentialclaude-sonnet-54/5AI tools (materials databases, simulation software, generative design) can significantly speed up option generation and trade-off analysis, greatly aiding engineers who retain final judgment.
Task automatabilityclaude-haiku-4-5-202510012/5AI can assist in literature review and suggest fabrication methods based on material properties, but determining *appropriate* methods requires deep physical intuition, tacit knowledge of equipment constraints, cost-benefit tradeoffs, and iterative experimentation that current AI cannot fully perform end-to-end with consistent quality at 50% time savings.
Task automatabilityclaude-sonnet-52/5This requires deep engineering judgment integrating material properties, manufacturing constraints, and application requirements; AI can suggest options but cannot reliably finalize decisions without expert validation.'
Adoption barriersclaude-haiku-4-5-202510013/5Materials engineers must often certify or validate fabrication methods for safety and quality; liability and regulatory requirements (especially in aerospace, medical, automotive) create moderate friction. However, no strict legal barrier prevents AI-assisted exploration provided humans verify outputs.
Adoption barriersclaude-sonnet-53/5While no formal licensing mandates a human for this specific decision, liability for structural/material failures and organizational engineering sign-off processes create real friction against pure automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI inference is cheap, but the task demands expert human judgment to validate recommendations, run experiments, and sign off on appropriateness—making the human-in-the-loop cost comparable to or higher than purely human deliberation. Meaningful cost savings are minimal.
Cost vs. human wageclaude-sonnet-52/5Specialized simulation and materials databases still require expensive engineering oversight and validation, so AI assistance doesn't yet dramatically undercut the cost of skilled engineers.
Technical feasibility todayclaude-haiku-4-5-202510012/5While AI tools can generate method recommendations and query databases of fabrication techniques, no deployed product reliably *determines* appropriate methods for novel material combinations without significant human expert review and lab validation. Existing tools are narrow-scope and require heavy expert oversight.
Technical feasibility todayclaude-sonnet-52/5Some AI-assisted materials selection and process simulation tools exist, but they support rather than replace engineer decision-making in production settings today.

Review new product plans, and make recommendations for material selection, based on design objectives such as strength, weight, heat resistance, electrical conductivity, and cost.

30

CI 30–30 · exposure 25 · augmentation 63 · importance 4.0/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Materials engineering and product development are moderate-digitization sectors; adoption of AI-driven tools is pilot-stage rather than production-at-scale. Most firms still rely on CAD, FEA, and manual materials databases rather than autonomous recommendation systems.
Sector adoption velocityclaude-sonnet-52/5Manufacturing and materials engineering sectors are moderate adopters of AI/ML tools for materials discovery but production-scale autonomous decision-making in design review remains uncommon.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can meaningfully assist engineers by rapidly surfacing candidate materials, comparing property matrices, and highlighting trade-offs—improving the speed of literature review and initial screening. However, final recommendation still demands human judgment on manufacturability, cost-risk trade-offs, and organizational constraints.
Augmentation potentialclaude-sonnet-54/5AI-powered materials databases, property prediction models, and generative design tools meaningfully speed up the screening and comparison of candidate materials for engineers.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can retrieve material properties, compare specifications, and flag trade-offs between parameters, the task requires integrating design context, manufacturing constraints, and novel materials expertise that current systems handle inconsistently. Reliable end-to-end automation meeting the 50% time-saving threshold is not yet demonstrated.
Task automatabilityclaude-sonnet-52/5Material selection requires integrating multiple engineering constraints, design context, and judgment calls that current AI can assist with but not fully replace end-to-end at production quality.
Adoption barriersclaude-haiku-4-5-202510013/5Material selection is often embedded in design review processes with organizational and regulatory touchpoints (safety, compliance), and companies typically require engineer sign-off. However, no strict legal licensing barrier prevents AI recommendation, creating moderate but not insurmountable friction.
Adoption barriersclaude-sonnet-53/5While not formally licensed like some engineering disciplines, material selection failures carry significant liability and safety implications, creating strong organizational incentive to keep humans accountable.
Cost vs. human wageclaude-haiku-4-5-202510012/5Current AI systems (language models, materials databases, optimization tools) require substantial setup, validation, and oversight by expert engineers to catch errors. The all-in cost per recommendation remains comparable to or higher than a junior engineer's time, especially when factoring liability and rework risk.
Cost vs. human wageclaude-sonnet-52/5Engineer review still requires substantial human validation and liability acceptance, so AI tools reduce but don't eliminate the cost of a qualified materials engineer's judgment.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed product reliably performs the full recommendation task; materials selection tools exist as databases and calculators, but they do not autonomously review plans and make contextual recommendations at production quality. Most real-world use remains expert-human-in-the-loop or research-stage AI.
Technical feasibility todayclaude-sonnet-52/5Materials informatics tools and AI-assisted databases exist for property lookup and screening, but no deployed product autonomously reviews product plans and issues reliable material recommendations without engineer oversight.

Conduct training sessions on new material products, applications, or manufacturing methods for customers and their employees.

30

CI 30–30 · exposure 25 · augmentation 75 · importance 3.6/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Materials engineering firms are moderately digitized and risk-averse in customer-facing roles. Adoption of AI-driven training is slow; most companies still rely on in-person or synchronous expert-led sessions, with only early pilots of AI content support.
Sector adoption velocityclaude-sonnet-52/5Manufacturing and materials engineering sectors have historically slower AI adoption for customer-facing technical training compared to software or finance sectors.
Augmentation potentialclaude-haiku-4-5-202510014/5AI excels at augmenting trainers by generating customized slides, answering routine technical questions in real time, providing translated materials, and creating follow-up resources, significantly improving trainer productivity and session quality while the engineer remains central.
Augmentation potentialclaude-sonnet-54/5AI can significantly help engineers prepare training content, generate visual aids, answer common questions, and create documentation, boosting productivity even though the human still delivers the training.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can generate training slides, video content, and documentation, conducting live interactive training sessions that respond to customer questions, adapt to audience needs, and build relationships remains largely dependent on human facilitation. Current systems cannot reliably manage the full dynamic session end-to-end.
Task automatabilityclaude-sonnet-52/5Delivering live, interactive technical training tailored to specific customer materials and manufacturing contexts requires expert judgment, hands-on demonstration, and real-time Q&A that current AI cannot fully replicate end-to-end.
Adoption barriersclaude-haiku-4-5-202510013/5Customers often expect training from qualified engineers, and product liability concerns create preference for human experts who can be held accountable. However, no strict regulatory barrier prevents AI-assisted or hybrid delivery, creating moderate but surmountable adoption friction.
Adoption barriersclaude-sonnet-53/5No formal licensing requirement, but customer expectations for direct expert engagement, liability for misapplied technical guidance, and need for hands-on demonstration create moderate friction.
Cost vs. human wageclaude-haiku-4-5-202510012/5Training content generation via AI is relatively cheap, but human experts are still needed to deliver, customize, and maintain credibility with customers. The all-in cost of AI-assisted training (infrastructure, content curation, human oversight) remains comparable to or exceeds direct human trainer costs.
Cost vs. human wageclaude-sonnet-52/5Content generation is cheap, but the actual training delivery still requires an engineer's time and expertise, so overall cost savings versus a human trainer are limited.
Technical feasibility todayclaude-haiku-4-5-202510012/5AI can support training delivery through content generation and virtual assistants, but no mature product reliably conducts independent customer training sessions at scale. Chatbot and video-based systems exist but lack the contextual expertise and adaptive interaction required for technical material engineering training.
Technical feasibility todayclaude-sonnet-52/5AI tools can generate training materials or slide decks, but no deployed product reliably conducts full technical training sessions for engineering customers without a human expert leading them.

Conduct or supervise tests on raw materials or finished products to ensure their quality.

29

CI 25–32 · exposure 25 · augmentation 75 · importance 4.1/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510013/5Manufacturing and materials sectors show moderate AI adoption in quality control (computer vision for defect detection, predictive analytics), but widespread autonomous material testing is still nascent. Pilots and early implementations are common, but full displacement of testing workflows remains rare in production.
Sector adoption velocityclaude-sonnet-52/5Materials engineering and manufacturing QA are moderately digitized but physical-testing-heavy sectors adopt AI slower than pure information-work sectors; pilots for predictive analytics exist but production-scale autonomous testing supervision is rare.
Augmentation potentialclaude-haiku-4-5-202510014/5AI substantially assists materials engineers by automating data analysis, detecting patterns in test results, predicting material failure, and generating quality reports—freeing engineers to focus on interpretation, troubleshooting, and decision-making. This augmentation is actively deployed and materially raises engineer productivity.
Augmentation potentialclaude-sonnet-54/5AI substantially aids data analysis, anomaly detection, predictive modeling, and report generation from test results, meaningfully boosting engineer productivity while the human still conducts/supervises physical tests.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can analyze test data and flag anomalies, the task requires hands-on material testing, sample preparation, and physical equipment operation that remain largely manual. Current AI cannot independently execute the full testing workflow without significant human intervention in specimen handling and instrument calibration.
Task automatabilityclaude-sonnet-52/5Physical testing of materials (tensile, hardness, chemical composition) requires lab equipment and physical handling that AI cannot perform end-to-end; AI can assist with data analysis and interpretation but not the core physical testing or supervision.4
Adoption barriersclaude-haiku-4-5-202510013/5Quality assurance and material testing are often subject to industry standards (ASTM, ISO) and regulatory oversight (FDA for pharmaceuticals, aerospace regulations) that may require human sign-off or certified technicians, creating moderate friction. However, these are not absolute legal prohibitions on automation.
Adoption barriersclaude-sonnet-54/5Quality assurance testing often falls under regulatory/certification standards (ASTM, ISO, industry-specific compliance) requiring qualified engineers to supervise and sign off on results, creating liability-driven barriers.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI analytics and supervision tools reduce labor costs on data interpretation, but the capital cost of testing equipment, human technicians for specimen handling, and integration overhead keep total costs comparable to or higher than direct human testing in many contexts.
Cost vs. human wageclaude-sonnet-52/5Physical testing still requires human technicians and equipment; AI reduces analysis time but does not eliminate the dominant cost of physical test execution and supervision.
Technical feasibility todayclaude-haiku-4-5-202510012/5Deployed AI systems exist for data analysis and quality flagging, but production-grade end-to-end autonomous material testing remains limited. Most real-world implementations require human technicians to conduct tests while AI assists in interpretation; no mature product independently performs comprehensive material testing at scale.
Technical feasibility todayclaude-sonnet-52/5Deployed AI products exist for data analysis and anomaly detection in test results, but no product autonomously conducts or supervises physical material testing in production today.

Design and direct the testing or control of processing procedures.

28

CI 25–30 · exposure 25 · augmentation 63 · importance 4.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Materials engineering and manufacturing remain relatively traditional sectors with slow AI adoption; while digital transformation is underway, autonomous testing design direction remains rare in production, with most AI use limited to analytics.
Sector adoption velocityclaude-sonnet-52/5Manufacturing and materials engineering sectors have historically slower AI adoption compared to information-centric industries, with AI use concentrated in simulation and analytics rather than full process control design.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can meaningfully assist engineers by generating candidate testing protocols, analyzing historical process data, and suggesting optimizations, enabling faster design iteration while the engineer retains critical judgment and sign-off authority.
Augmentation potentialclaude-sonnet-54/5AI and simulation tools significantly aid in designing test matrices, analyzing data, predicting material behavior, and optimizing processing parameters, meaningfully boosting engineer productivity while humans retain oversight.
Task automatabilityclaude-haiku-4-5-202510012/5AI can assist in generating testing protocols and analyzing data, but cannot independently design end-to-end testing procedures or direct control of manufacturing processes without human judgment on process variables, safety constraints, and real-time decision-making.
Task automatabilityclaude-sonnet-52/5This task requires designing experimental protocols and exercising engineering judgment over physical processing systems, which AI cannot fully perform end-to-end; it can assist in planning and data analysis but not direct physical testing or control autonomously.
Adoption barriersclaude-haiku-4-5-202510014/5Materials testing and process control often require licensed engineers or PEs to certify designs and procedures; liability for material failures and safety issues creates strong legal and regulatory barriers to full automation.
Adoption barriersclaude-sonnet-53/5While no formal licensure typically gates this specific task, safety-critical materials testing (e.g., aerospace, automotive) often requires engineer sign-off and quality certification, creating moderate organizational and liability barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5Current AI solutions require significant integration, human oversight, and domain expert validation, making all-in costs comparable to or higher than the human engineer's loaded wage for autonomous reliable execution.
Cost vs. human wageclaude-sonnet-52/5AI can cheaply assist with data analysis or literature review, but the overall task still requires expert engineers overseeing physical experiments and equipment, keeping costs comparable to human-driven work.
Technical feasibility todayclaude-haiku-4-5-202510012/5Tools exist for data analysis and protocol suggestions, but no mature product reliably designs and directs testing procedures autonomously in production materials engineering environments where domain expertise and regulatory compliance are critical.
Technical feasibility todayclaude-sonnet-52/5No deployed product independently designs and directs materials processing testing regimes; AI tools support data analysis, simulation, and documentation but the core design/direction function remains human-led.

Monitor material performance, and evaluate its deterioration.

28

CI 25–30 · exposure 25 · augmentation 63 · importance 4.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Adoption of automated monitoring is slower in traditional heavy industries (aerospace, civil engineering) where materials failures carry catastrophic costs and regulatory scrutiny. While predictive maintenance is growing in manufacturing, material deterioration evaluation remains heavily human-driven in regulated sectors.
Sector adoption velocityclaude-sonnet-52/5Manufacturing and materials engineering sectors show slower AI adoption compared to information/professional services, with predictive analytics tools still in pilot or narrow deployment phases.
Augmentation potentialclaude-haiku-4-5-202510013/5AI-powered monitoring dashboards and anomaly detection can assist engineers by surfacing trends and flagging degradation signals that require investigation. However, the human engineer must still contextualize findings and make judgment calls on failure modes, so assistance is moderate and task-dependent rather than transformative.
Augmentation potentialclaude-sonnet-54/5AI-based predictive analytics, computer vision for defect detection, and data trend analysis significantly enhance an engineer's ability to monitor and interpret material performance data.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can analyze sensor data and imaging to detect anomalies in material performance, the task requires real-time decision-making about deterioration severity, environmental context, and predictive judgment that demands human expertise. Current systems can flag potential issues but cannot reliably evaluate complex deterioration patterns end-to-end without substantial human oversight.
Task automatabilityclaude-sonnet-52/5Monitoring involves physical inspection, sensor data collection, and interpretation of degradation mechanisms that require judgment and often hands-on testing, limiting full automation today.
Adoption barriersclaude-haiku-4-5-202510014/5Material performance evaluation often directly impacts safety-critical decisions (aircraft, infrastructure, medical devices), creating high liability and regulatory requirements. Industry standards, certification bodies, and liability frameworks typically require a licensed engineer to sign off on material fitness assessments, even if AI assists.
Adoption barriersclaude-sonnet-53/5No licensing requirement specifically for this task, but safety-critical applications (aerospace, infrastructure) impose liability and certification requirements that necessitate qualified engineer sign-off on deterioration assessments.
Cost vs. human wageclaude-haiku-4-5-202510012/5Sensor infrastructure, AI model development, integration with existing monitoring systems, and required human oversight add significant upfront and ongoing costs. For specialized materials engineering contexts, the all-in cost per evaluation often exceeds the loaded wage of an engineer performing spot checks.
Cost vs. human wageclaude-sonnet-52/5Deploying sensor networks, data pipelines, and specialized ML models for materials monitoring requires significant upfront investment often comparable to or exceeding engineer labor costs for many applications.
Technical feasibility todayclaude-haiku-4-5-202510012/5Computer vision and ML models can identify visible degradation in images/sensor feeds, and condition-monitoring software exists in production; however, these systems typically require careful tuning per material type and application, and they often generate false positives/negatives. No deployed product reliably evaluates material deterioration across diverse contexts without expert validation.
Technical feasibility todayclaude-sonnet-52/5AI/ML tools exist for predictive maintenance and materials degradation modeling but are narrow, domain-specific, and not broadly deployed as reliable end-to-end solutions across materials engineering contexts.

Evaluate technical specifications and economic factors relating to process or product design objectives.

28

CI 25–30 · exposure 25 · augmentation 75 · importance 4.1/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Materials engineering is a relatively specialized field with slower digitization and AI adoption compared to software or finance; pilots exist but production AI substitution for design evaluation remains limited.
Sector adoption velocityclaude-sonnet-52/5Manufacturing and materials engineering sectors show slower AI adoption than software/finance, with pilots for design assistance but limited deployment at scale.
Augmentation potentialclaude-haiku-4-5-202510014/5AI can meaningfully assist by rapidly synthesizing technical data, cost comparisons, and regulatory databases, substantially boosting engineer productivity in data gathering and initial screening phases while the engineer retains judgment on final design decisions.
Augmentation potentialclaude-sonnet-54/5AI can meaningfully assist by summarizing technical standards, running cost simulations, and flagging tradeoffs, boosting engineer productivity while decisions remain human-led.
Task automatabilityclaude-haiku-4-5-202510012/5AI can assist with extracting and comparing technical specifications and cost data, but evaluating tradeoffs between technical and economic factors requires domain expertise, contextual judgment, and understanding of unstated design constraints that AI struggles with reliably today.
Task automatabilityclaude-sonnet-52/5This requires integrating deep domain judgment, tacit knowledge of manufacturing constraints, and economic tradeoffs that current AI can partially inform but not autonomously resolve to equal quality.
Adoption barriersclaude-haiku-4-5-202510014/5Materials engineers bear professional liability for design decisions; regulatory standards (ASME, ISO) and client accountability typically require a licensed or experienced engineer to sign off on or approve critical technical-economic evaluations.
Adoption barriersclaude-sonnet-53/5No strict licensing requirement in most contexts, but engineering sign-off, safety liability, and organizational review processes create meaningful friction against full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI tools (LLMs, data extraction) are cheap per query, but the overhead of validation, error-checking, and engineer review to confirm recommendations means total integrated cost remains comparable to direct engineering work.
Cost vs. human wageclaude-sonnet-52/5AI tools can cheaply summarize specs or run cost models, but the human engineer's validation, liability, and judgment keep overall cost comparable to or only modestly below human-only work given required oversight.
Technical feasibility todayclaude-haiku-4-5-202510012/5No production systems reliably perform full evaluation of technical-economic design tradeoffs end-to-end; AI excels at data retrieval and analysis but lacks the integrative judgment and accountability materials engineers apply to complex design decisions.
Technical feasibility todayclaude-sonnet-52/5Some engineering copilots and analytics tools assist with data synthesis, but no deployed product reliably performs full technical-economic evaluation for materials/process design decisions in production.

Plan and evaluate new projects, consulting with other engineers and corporate executives, as necessary.

28

CI 25–30 · exposure 25 · augmentation 75 · importance 3.8/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Materials engineering and capital project planning remain conservative domains with slow digitization. While some firms pilot AI-assisted feasibility tools, actual autonomous planning adoption in production remains rare; sectors lag information/finance in AI deployment.
Sector adoption velocityclaude-sonnet-52/5Engineering and manufacturing sectors have historically been slower AI adopters compared to information/finance, and this task's collaborative planning nature further limits automation-driven production deployment.
Augmentation potentialclaude-haiku-4-5-202510014/5AI can meaningfully assist by generating preliminary cost/schedule analyses, summarizing technical literature, identifying design trade-offs, and organizing data for stakeholder review. These augmentations raise human productivity in evaluation significantly while the engineer and executive retain judgment and sign-off.
Augmentation potentialclaude-sonnet-54/5AI tools can meaningfully support this task by synthesizing technical data, drafting proposals, generating scenario analyses, and preparing materials for consultations, substantially boosting engineer productivity while humans retain decision authority.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist with data analysis, project feasibility studies, and documentation, the task fundamentally requires human judgment to evaluate trade-offs, consult stakeholders, and make strategic decisions. No current system can autonomously plan and evaluate novel engineering projects end-to-end at quality parity with 50% time savings.
Task automatabilityclaude-sonnet-52/5Project planning and evaluation involve strategic judgment, negotiation, and cross-functional consultation with executives that current AI cannot autonomously execute end-to-end; AI can assist with data synthesis but not replace the interpersonal and decision-making core.
Adoption barriersclaude-haiku-4-5-202510014/5This task involves strategic corporate decision-making, fiduciary responsibility, and coordination across engineering and management hierarchies. Organizations have strong institutional and liability reasons to retain human experts in planning and evaluation roles, and corporate culture favors human leadership in material decisions.
Adoption barriersclaude-sonnet-53/5No formal licensing barrier for this specific task, but organizational and trust-based friction is significant since executives and engineers expect human accountability and relational judgment in project decisions.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI inference for project analysis is inexpensive, but the consultation and decision-making roles require skilled human professionals whose loaded wages ($100K+/year) dwarf the marginal AI cost. Full task cost remains dominated by human labor.
Cost vs. human wageclaude-sonnet-52/5AI can cheaply generate supporting analysis, but the actual planning and consultation still requires paid engineer/executive time, so overall cost savings are modest relative to the full task.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed product reliably performs independent project planning and evaluation for materials engineering at scale. AI tools can support components (cost modeling, literature review) but cannot replace the iterative consultation with engineers and executives that defines this task.
Technical feasibility todayclaude-sonnet-52/5No deployed product independently plans and evaluates engineering projects while consulting executives; existing tools support drafting reports or summarizing data but do not perform the judgment-based interfacing task reliably in production.

Solve problems in a number of engineering fields, such as mechanical, chemical, electrical, civil, nuclear, and aerospace.

25

CI 25–25 · exposure 25 · augmentation 75 · importance 3.6/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Engineering sectors have digitized CAD and simulation tools but remain conservative about autonomous decision-making; adoption of AI in production workflows is primarily at the pilot and support level, not wholesale replacement, due to liability and verification requirements.
Sector adoption velocityclaude-sonnet-52/5Engineering sectors are adopting AI tools for specific tasks (simulation, drafting) but broad autonomous problem-solving adoption remains in early pilot stages, slower than software/finance sectors.
Augmentation potentialclaude-haiku-4-5-202510014/5AI tools (generative design, simulation acceleration, knowledge retrieval, failure analysis databases) meaningfully augment engineer productivity by compressing design iteration cycles and helping explore solution spaces; the human engineer retains final judgment and validation responsibility.
Augmentation potentialclaude-sonnet-54/5AI significantly aids engineers by accelerating literature review, generating hypotheses, running simulations, and drafting analyses, meaningfully boosting productivity while humans retain final judgment.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist with narrowly scoped engineering subproblems (e.g., simulation, calculation, literature review), end-to-end problem-solving across diverse engineering fields requires deep contextual judgment, integration of multiple disciplines, and validation against real-world constraints that current AI systems cannot reliably orchestrate without significant human oversight.
Task automatabilityclaude-sonnet-52/5Cross-disciplinary engineering problem-solving requires deep physical intuition, novel synthesis, and judgment that current AI cannot reliably replicate end-to-end, though it can assist with sub-components like calculations or literature review.
Adoption barriersclaude-haiku-4-5-202510014/5Professional engineering practice is heavily regulated (PE licensing, liability for design failures, safety standards, code compliance); any automation that signs off on designs or recommendations faces legal and regulatory barriers, and organizational risk aversion to AI-driven decisions in safety-critical domains remains high.
Adoption barriersclaude-sonnet-54/5Engineering work often requires licensed professional engineer sign-off and carries high liability for safety-critical failures (aerospace, nuclear, civil), creating strong barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI assists with computations and data processing but does not replace the specialized human engineer's involvement, and integration with design workflows, testing, and liability review add overhead; total cost savings remain modest compared to the loaded wage of a skilled materials engineer.
Cost vs. human wageclaude-sonnet-52/5Significant human oversight, validation, and domain expertise are still required, so AI assistance reduces but does not eliminate the dominant labor cost for this complex task.
Technical feasibility todayclaude-haiku-4-5-202510012/5Some narrow engineering tasks (FEA setup, calculation checks, design optimization within predefined parameters) have partial automation in research and practice, but no deployed product reliably solves novel cross-disciplinary engineering problems autonomously; human engineers remain essential for problem formulation and validation.
Technical feasibility todayclaude-sonnet-52/5AI copilots exist for narrow engineering calculations and simulations but no deployed product autonomously solves open-ended, multidisciplinary engineering problems in production.

Design processing plants and equipment.

23

CI 20–25 · exposure 20 · augmentation 75 · importance 3.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Capital-intensive, risk-averse industrial sectors adopt automation cautiously. While CAD/CAM tools are standard, integration of AI-driven design remains in pilot phases; most plants still rely on traditional engineering workflows and proven methodologies over novel AI suggestions.
Sector adoption velocityclaude-sonnet-52/5Engineering and heavy industry sectors are slower AI adopters compared to software/finance; AI-assisted design tools are being piloted but production-scale autonomous plant design is rare.
Augmentation potentialclaude-haiku-4-5-202510014/5AI excels at accelerating simulation, optimization, and code generation for plant design; engineers use these tools to explore design spaces faster and catch errors earlier. AI augmentation significantly boosts productivity when a human expert remains accountable for validation and final decisions.
Augmentation potentialclaude-sonnet-54/5AI-based simulation, generative design, and CAD assistance meaningfully speed up iteration, layout optimization, and preliminary calculations while engineers retain final design authority.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist with parametric modeling, simulations, and code generation for plant design, the task requires substantial human judgment on safety, process chemistry, and novel configurations that current AI struggles with end-to-end. Significant domain expertise and iterative refinement remain necessary to produce production-ready designs.
Task automatabilityclaude-sonnet-52/5Designing processing plants and equipment requires integrating physical constraints, safety codes, and novel engineering judgment that current AI cannot reliably perform end-to-end; AI can assist with sub-components but not the full design task at equal quality.
Adoption barriersclaude-haiku-4-5-202510014/5Engineering designs for processing plants face significant regulatory, safety, and liability requirements. Licensed Professional Engineers often must stamp designs, and clients typically require human accountability. Insurance and compliance frameworks strongly favor human sign-off, creating hard barriers to full automation.
Adoption barriersclaude-sonnet-54/5Plant and equipment designs typically require professional engineer stamping/certification and regulatory compliance (safety, environmental), creating strong legal and liability barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI tools (cloud simulations, CAD plugins) cost thousands monthly, but engineering salaries are substantial. The all-in cost of AI-assisted design—including oversight, validation, and rework—remains comparable to or higher than direct human design for complex plants.
Cost vs. human wageclaude-sonnet-52/5AI tools reduce some drafting and simulation time but the overall design process still requires expensive expert oversight, engineering validation, and liability sign-off, keeping costs comparable to or only modestly below human-only costs.
Technical feasibility todayclaude-haiku-4-5-202510012/5Some CAD and simulation tools integrate AI for component suggestions and optimization, but no mature production system reliably designs complete processing plants autonomously. Deployed products handle narrow subtasks (e.g., pipe routing) rather than the full design lifecycle, which requires validation against complex regulations and real-world constraints.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously designs full processing plants; CAD/simulation tools exist but require expert engineers to drive design decisions, making this research-stage for full automation.

Plan and implement laboratory operations to develop material and fabrication procedures that meet cost, product specification, and performance standards.

21

CI 16–25 · exposure 20 · augmentation 75 · importance 3.9/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5While AI-assisted data analysis is growing in materials science research, adoption of AI for autonomous laboratory planning and implementation remains limited. Most materials engineering firms use AI as a supplementary tool rather than autonomous execution systems.
Sector adoption velocityclaude-sonnet-52/5Materials science and manufacturing R&D sectors are slower to adopt AI at the operational level compared to information-heavy industries, though AI-assisted materials discovery tools are gaining traction in pilots.
Augmentation potentialclaude-haiku-4-5-202510014/5AI can meaningfully assist materials engineers by analyzing material properties data, suggesting parameter ranges, optimizing cost-specification tradeoffs, and accelerating simulation of fabrication procedures. This augmentation can substantially improve productivity while engineers retain critical decision-making authority.
Augmentation potentialclaude-sonnet-54/5AI significantly aids tasks like simulation, materials informatics, predictive modeling, and literature synthesis, helping engineers plan experiments and optimize fabrication procedures faster.
Task automatabilityclaude-haiku-4-5-202510012/5Planning and implementing laboratory operations for material development requires substantial experimentation, iterative judgment, and domain expertise. While AI can assist with documentation and analysis of results, the core task of designing procedures and making real-time decisions based on lab outcomes remains heavily dependent on human expertise and hands-on experimentation.
Task automatabilityclaude-sonnet-52/5This task combines hands-on laboratory management, physical experimentation, and iterative process development that cannot be executed end-to-end by AI; only planning/documentation sub-components are automatable today.dent
Adoption barriersclaude-haiku-4-5-202510014/5Significant barriers exist: laboratory operations often require direct human oversight for safety and equipment operation, professional licensing and accountability for quality assurance, regulatory compliance in manufacturing contexts, and organizational reliance on expert judgment for procedure approval.
Adoption barriersclaude-sonnet-54/5Engineering sign-off, safety protocols, and quality/performance certification for materials often require licensed or credentialed engineers, and liability for material failures creates strong barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510011/5The fully-loaded cost of a materials engineer remains substantially lower than the integration and oversight costs of AI systems that would need to coordinate lab equipment, run experiments, interpret nuanced results, and iterate on procedures—all requiring continuous human validation.
Cost vs. human wageclaude-sonnet-52/5AI can cheaply assist with data analysis or design-of-experiments planning, but the physical lab setup, equipment operation, and hands-on troubleshooting still require costly skilled engineers and technicians.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed AI systems can autonomously plan and implement full laboratory operations for materials development. Existing tools can help analyze data or suggest parameters, but reliable end-to-end execution of experimental procedures with cost and specification optimization requires human materials engineers.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously plans and runs laboratory operations for material fabrication; current AI tools are limited to simulation, literature review, or data analysis support, not lab execution.

Modify properties of metal alloys, using thermal and mechanical treatments.

19

CI 7–30 · exposure 13 · augmentation 75 · importance 4.1/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Materials engineering is relatively conservative; simulation and prediction tools are adopted for planning and optimization, but the physical execution and validation remain human-driven. Adoption is slower than in pure software/information sectors.
Sector adoption velocityclaude-sonnet-52/5Manufacturing and materials engineering sectors are slower AI adopters for physical process execution, though computational materials science tools are gaining traction for design and prediction.
Augmentation potentialclaude-haiku-4-5-202510014/5AI significantly augments materials engineers through predictive modeling, optimization of treatment parameters, real-time monitoring assistance, and simulation of outcomes before physical testing. These tools substantially raise engineer productivity in design and decision-making while the engineer remains responsible for execution and validation.
Augmentation potentialclaude-sonnet-54/5AI-driven materials modeling, simulation of phase transformations, and optimization of treatment parameters can significantly assist engineers in planning and predicting outcomes before physical execution.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can optimize parameters for thermal and mechanical treatments based on predictive models and historical data, the task requires hands-on physical intervention, real-time monitoring of complex material transformations, and judgment calls in response to unexpected outcomes. Current AI systems cannot autonomously perform the full end-to-end work of modifying alloy properties in a lab or production setting.
Task automatabilityclaude-sonnet-51/5This is a hands-on physical process involving furnaces, quenching, rolling, or forging equipment operated on real materials; no AI system can physically execute thermal or mechanical treatments today.
Adoption barriersclaude-haiku-4-5-202510013/5Materials testing and production modifications often require compliance with industry standards (ASTM, ISO) and documented human responsibility for material quality and safety. Many sectors expect a qualified engineer to certify or oversee the treatment, creating friction but not a total legal ban on AI support.
Adoption barriersclaude-sonnet-54/5Industrial safety regulations, equipment operation certifications, and quality/liability requirements in materials processing create strong barriers to any non-human execution of physical treatment steps.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI optimization tools and simulation software reduce engineering time for design and parameter selection, but the cost of specialized hardware, simulation software licensing, and continued human expertise remains substantial relative to a single engineer's wage. Integration is incomplete.
Cost vs. human wageclaude-sonnet-51/5The physical execution still requires human labor, equipment, and energy costs regardless of AI involvement, so AI does not reduce the core cost of performing the treatment.
Technical feasibility todayclaude-haiku-4-5-202510012/5Deployed AI tools exist for predicting alloy behavior and optimizing treatment parameters (simulation software, machine learning models), but they support rather than replace the engineer's decisions and physical actions. No production system autonomously executes thermal and mechanical treatments without human oversight.
Technical feasibility todayclaude-sonnet-51/5No deployed product performs physical alloy treatment; AI's role is limited to simulation/prediction support, not execution of the task itself.

Guide technical staff in developing materials for specific uses in projected products or devices.

18

CI 5–30 · exposure 13 · augmentation 63 · importance 4.0/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Adoption of AI for staff guidance in technical roles remains minimal; even digitized sectors view supervisory and mentoring roles as requiring human judgment and accountability. No measurable production displacement in this area.
Sector adoption velocityclaude-sonnet-52/5Materials engineering and manufacturing sectors show slower AI adoption for core R&D leadership tasks compared to information/finance sectors, though computational tools are gradually integrated.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can meaningfully assist by generating material property summaries, suggesting candidates for specific applications, or surfacing relevant prior projects—tools that help a human engineer guide staff more efficiently. However, the assistance is partial; human judgment remains central.
Augmentation potentialclaude-sonnet-54/5AI tools (materials informatics, simulation, generative design) significantly enhance an engineer's ability to guide staff by providing faster data analysis, predictive modeling, and literature synthesis.
Task automatabilityclaude-haiku-4-5-202510011/5Guiding technical staff requires judgment about product requirements, material trade-offs, and team coordination—fundamentally a human-to-human supervisory function that current AI cannot replicate end-to-end. While AI can assist in material property lookup or analysis, the core act of directing staff based on anticipated project needs remains beyond current capabilities.
Task automatabilityclaude-sonnet-52/5This is a leadership and technical judgment task involving mentoring, decision-making under uncertainty, and applying deep domain expertise; AI cannot autonomously guide human staff or take responsibility for material development decisions.
Adoption barriersclaude-haiku-4-5-202510014/5Organizations typically require a licensed or credentialed materials engineer or senior technician to direct technical development and mentor staff—a role with implicit professional accountability. Customer and regulatory expectations, plus team dynamics, create strong friction against AI takeover of this supervisory function.
Adoption barriersclaude-sonnet-53/5No strict licensing requirement, but organizational trust, liability for product failures, and the need for accountable human leadership create moderate friction against full automation.
Cost vs. human wageclaude-haiku-4-5-202510011/5Replacing the judgment and authority required to guide staff would require high-touch AI oversight and validation, making the all-in cost likely exceed hiring a qualified engineer or supervisor. The liability and quality risk of AI-driven staff guidance would be prohibitively expensive.
Cost vs. human wageclaude-sonnet-52/5AI can cheaply support literature review or simulations, but the managerial/technical guidance role still requires a paid, experienced engineer, so overall cost savings are limited.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed product reliably guides technical staff in this sense; AI systems can support material selection recommendations or generate technical content, but they cannot meaningfully lead or mentor a team in real time. Narrow tools exist for material property analysis, but not for the supervisory, interpersonal guidance this task demands.
Technical feasibility todayclaude-sonnet-51/5No deployed product provides supervisory/technical guidance to engineering staff on materials development; AI tools exist for simulation and literature search but not as autonomous guiders of teams.

Present technical information at conferences.

18

CI 5–30 · exposure 13 · augmentation 63 · importance 3.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Conference presentation delivery is an inherently human-facing professional activity in which material engineers maintain direct control; adoption of AI for full automation is minimal because the task is tied to expert visibility and professional standing.
Sector adoption velocityclaude-sonnet-52/5Engineering and R&D sectors show moderate AI adoption for drafting and analysis, but conference presentation itself is a niche, low-digitization activity with little AI displacement.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist by drafting slide content, generating speaker notes, and organizing data visualizations, helping engineers prepare more efficiently. However, the augmentation is limited to preparation; the delivery itself remains the human's responsibility.
Augmentation potentialclaude-sonnet-54/5AI tools can meaningfully help engineers prepare slides, summarize research, rehearse talking points, and anticipate audience questions, substantially boosting preparation productivity even though delivery remains human.
Task automatabilityclaude-haiku-4-5-202510011/5Presenting technical information at conferences requires real-time audience engagement, adaptive delivery, spontaneous handling of questions, and personal credibility—dimensions that current AI cannot handle end-to-end. AI cannot reliably substitute for the human speaker's presence, judgment, and interactive problem-solving.
Task automatabilityclaude-sonnet-52/5AI can help draft slides and scripts but the live presentation, audience interaction, and real-time Q&A require human presence and judgment, so end-to-end automation with equal quality is not achievable today.
Adoption barriersclaude-haiku-4-5-202510014/5Conference presentations require human expertise, professional credibility, and direct audience engagement; organizers and audiences expect a qualified human speaker, and liability for technical accuracy rests with the speaker. These reputational and professional norms create strong barriers to automation.
Adoption barriersclaude-sonnet-53/5No formal licensing requirement, but strong professional norms, credibility, and networking expectations mean human presence is effectively required at conferences.
Cost vs. human wageclaude-haiku-4-5-202510011/5The all-in cost of AI systems for generating slides and notes plus required human oversight and revision exceeds the cost of a materials engineer preparing and delivering their own presentation.
Cost vs. human wageclaude-sonnet-52/5While AI tools can cheaply generate slide content, an AI cannot substitute for the actual conference delivery, so the effective cost comparison for the full task favors the human presenter.
Technical feasibility todayclaude-haiku-4-5-202510012/5While AI can draft presentation slides and generate speaker notes, no deployed product reliably performs the full task of presenting at a conference. Live presentation requires real-time social interaction and adaptive response that current systems cannot replicate in production environments.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously presents technical materials science content at conferences on behalf of engineers; this remains outside current production use cases.

Supervise production and testing processes in industrial settings, such as metal refining facilities, smelting or foundry operations, or nonmetallic materials production operations.

16

CI 7–25 · exposure 13 · augmentation 63 · importance 3.8/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Industrial sectors adopting autonomous AI-driven supervision remain limited. While predictive maintenance and monitoring dashboards are emerging, true autonomous supervision of production in foundries and smelting operations is still primarily pilot-stage, with human supervisors retained as the decision-making authority.
Sector adoption velocityclaude-sonnet-52/5Heavy industry and manufacturing sectors are slower AI adopters compared to information/professional services, though sensor-based monitoring and predictive maintenance are gradually being introduced.
Augmentation potentialclaude-haiku-4-5-202510014/5AI-powered monitoring dashboards, anomaly detection, and predictive maintenance tools substantially augment engineer productivity by surfacing trends, automating routine alerts, and reducing time spent on manual sensor review. Engineers leveraging these tools can focus on complex decision-making and strategic optimization.
Augmentation potentialclaude-sonnet-53/5AI-driven sensors, predictive analytics, and quality control systems can meaningfully assist engineers in monitoring processes and flagging anomalies, improving their supervisory effectiveness without replacing them.
Task automatabilityclaude-haiku-4-5-202510012/5Supervising production and testing processes requires real-time assessment of complex, highly variable industrial systems with safety-critical decision-making. While AI can monitor sensor data and flag anomalies, current systems cannot reliably replace the judgment-intensive oversight of process deviations, equipment failures, and worker safety issues that demand immediate human intervention and contextual reasoning.
Task automatabilityclaude-sonnet-51/5Direct on-site supervision of physical production/testing processes requires physical presence, real-time judgment, and hands-on oversight of equipment and personnel that current AI cannot perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Strong regulatory barriers exist: OSHA, EPA, and industry safety standards typically require a licensed or certified engineer to be accountable for process safety, equipment integrity, and worker protection. Liability for failures in metal refining or smelting operations is high, and legal responsibility cannot be delegated to an automated system.
Adoption barriersclaude-sonnet-54/5Industrial safety regulations, liability for equipment failures/accidents, and often certification requirements for engineers overseeing hazardous processes create strong barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5Industrial supervision integration costs are significant (sensor infrastructure, AI platform maintenance, integration with legacy control systems, human oversight), and the loaded cost of a full-time materials engineer is moderate relative to the complexity of the installation. Automation does not yet approach an order-of-magnitude cost advantage.
Cost vs. human wageclaude-sonnet-51/5Since AI cannot perform the core supervisory function, there is no viable cost comparison—human supervisors remain necessary, making AI substitution infeasible rather than merely costly.
Technical feasibility todayclaude-haiku-4-5-202510012/5Deployed monitoring systems exist for industrial processes, but they typically operate as decision-support tools, not autonomous supervisors. No production systems today can independently manage the full scope of supervising metal refining or foundry operations—including troubleshooting unexpected failures, coordinating worker safety, and making dynamic process adjustments—without substantial human oversight.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product supervises industrial production floors autonomously; monitoring sensors and analytics tools exist but do not replace the supervisory role itself.

Teach in colleges and universities.

15

CI 5–25 · exposure 17 · augmentation 63 · importance 3.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Higher education has been slow to adopt AI in core teaching functions; while institutions pilot chatbots and automated grading tools, there is limited production use of AI replacing instructional delivery, and strong institutional and cultural resistance to full automation of teaching.
Sector adoption velocityclaude-sonnet-52/5Higher education adoption of AI for actual teaching delivery (versus administrative or research support) remains slow and mostly pilot-stage, constrained by academic norms and accreditation.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can usefully assist instructors by generating lecture outlines, summarizing papers, drafting feedback on student work, and creating practice problems, but the instructor must remain in the loop to evaluate appropriateness, accuracy, and pedagogical fit for their course.
Augmentation potentialclaude-sonnet-54/5AI tools substantially help materials engineers prepare course materials, generate examples, create quizzes, and explain concepts, meaningfully boosting teaching productivity while the instructor remains central.
Task automatabilityclaude-haiku-4-5-202510011/5Teaching in colleges and universities requires real-time interaction, dynamic adaptation to student questions, assessment of learning, and the exercise of judgment about pedagogical approaches—core human competencies that current AI cannot perform end-to-end at 50% time savings while maintaining teaching quality.
Task automatabilityclaude-sonnet-52/5Teaching involves live lecturing, mentoring, grading judgment calls, and adaptive interaction with students that current AI cannot fully replicate end-to-end, though some prep and grading sub-tasks can be assisted.atable content generation is partial.'
Adoption barriersclaude-haiku-4-5-202510015/5Universities require faculty to hold relevant degrees and professional credentials; teaching roles are deeply embedded in accreditation standards, employment law, and institutional governance, creating hard legal and regulatory barriers to substitution by AI systems.
Adoption barriersclaude-sonnet-54/5University teaching typically requires credentialed faculty, accreditation standards, and institutional/legal requirements for instructor qualifications, creating strong barriers to full AI substitution.
Cost vs. human wageclaude-haiku-4-5-202510011/5Delivering college instruction requires specialized human expertise, institutional credentialing, and sustained engagement that AI cannot yet match; the all-in cost of an AI system with oversight and quality control would exceed the cost of employing instructors at scale.
Cost vs. human wageclaude-sonnet-52/5While AI can cheaply produce lecture notes or slides, replicating the full teaching role (office hours, mentoring, accreditation-required instruction) still requires substantial human involvement, keeping costs comparable rather than dramatically lower.
Technical feasibility todayclaude-haiku-4-5-202510012/5While AI can assist with lecture preparation, grading, and content generation, no deployed product reliably performs the full task of college-level teaching as an instructor; production systems exist for narrow components (quiz generation, draft feedback) but not for authentic classroom instruction.
Technical feasibility todayclaude-sonnet-52/5AI tutoring and content-generation tools exist but no deployed product reliably conducts full college-level engineering courses including live instruction, advising, and assessment at scale in production.

Supervise the work of technologists, technicians, and other engineers and scientists.

13

CI 5–21 · exposure 13 · augmentation 50 · importance 3.9/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Adoption of AI for supervisory roles in engineering remains negligible; technical teams heavily prefer and depend on human management for accountability, career guidance, and contextual decision-making. No broad industry pattern of AI-driven supervision exists.
Sector adoption velocityclaude-sonnet-52/5While engineering firms adopt AI tools for technical work, replacing human supervisory roles with AI remains rare and not part of current adoption trends.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can meaningfully assist supervisors with performance analytics, workload balancing, documentation review, and meeting preparation, improving their productivity on administrative aspects while humans retain core leadership and developmental responsibilities.
Augmentation potentialclaude-sonnet-53/5AI can help supervisors track project status, generate reports, and summarize team performance data, aiding but not replacing the supervisory function.
Task automatabilityclaude-haiku-4-5-202510012/5Supervising technical work requires real-time judgment, interpersonal decision-making, and contextual understanding of complex engineering problems. While AI could assist with status tracking and documentation, end-to-end replacement with 50% time savings at equal quality is not achievable today—human leadership and mentorship remain essential.
Task automatabilityclaude-sonnet-51/5Supervision involves interpersonal leadership, performance management, mentoring, and judgment calls about people that current AI cannot perform end-to-end.'
Adoption barriersclaude-haiku-4-5-202510014/5Organizational hierarchy, legal accountability, professional liability, and implicit requirements for human judgment and decision authority create substantial barriers. Employees and organizations expect human supervisors; automated supervision would face strong cultural and structural resistance.
Adoption barriersclaude-sonnet-54/5Organizational structure, accountability, HR/legal responsibilities, and professional engineering oversight requirements mean a human must hold supervisory authority.
Cost vs. human wageclaude-haiku-4-5-202510011/5The all-in cost of AI systems that attempt to handle supervision tasks (plus required human oversight and integration) exceeds the loaded wage of a materials engineer supervisor, especially given the need for fallback human judgment and the liability of autonomous supervision failures.
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute performing this managerial function, so no meaningful cost comparison favors AI over a human supervisor.
Technical feasibility todayclaude-haiku-4-5-202510012/5No production systems reliably perform full supervisory and mentoring roles independently. AI tools can support scheduling, reporting, and performance tracking, but deployed products lack the social-emotional intelligence, accountability, and authority required to truly supervise technical teams.
Technical feasibility todayclaude-sonnet-51/5No deployed product manages or supervises technical staff autonomously; AI tools at best support scheduling or reporting but not supervisory responsibility.

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.