Museum Technicians and Conservators

25-4013.00
Median wage $51,440/yr12,310 employed (US)Rank #691 of 923 scored · top 75% by substitution

Restore, maintain, or prepare objects in museum collections for storage, research, or exhibit. May work with specimens such as fossils, skeletal parts, or botanicals; or artifacts, textiles, or art. May identify and record objects or install and arrange them in exhibits. Includes book or document conservators.

Sub-scores

0–100 · band = confidence interval from rater disagreement

Substitution19
Exposure17
Augmentation47

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

24 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%20

panel mean rating 1.8/5 → substitution pressure 20/100

Technical feasibility todayw 20%12

panel mean rating 1.5/5 → substitution pressure 12/100

Cost vs. human wagew 15%14

panel mean rating 1.6/5 → substitution pressure 14/100

Adoption barriersw 20%inverted — strong barriers lower the score33

panel mean rating 3.7/5 (barrier strength) → substitution pressure 33/100

Sector adoption velocityw 10%10

panel mean rating 1.4/5 → substitution pressure 10/100

Task breakdown (24 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.

Enter information about museum collections into computer databases.

69

CI 65–72 · exposure 70 · augmentation 75 · importance 4.0/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510013/5Museums and cultural institutions adopt digital tools at moderate pace; many are digitizing collections with AI assistance, but adoption varies widely by institution size and funding. Widespread production deployment is visible but not yet universal across sectors.
Sector adoption velocityclaude-sonnet-52/5Museums and cultural heritage institutions are historically slow adopters of AI tools due to budget constraints, legacy systems, and small technical staff, resulting in mostly pilot-level implementation.
Augmentation potentialclaude-haiku-4-5-202510014/5AI-powered data entry assistants (automatic field population, duplicate detection, metadata suggestions) significantly boost technician productivity by reducing manual keystrokes and catching errors, while the human retains verification and contextual judgment.
Augmentation potentialclaude-sonnet-54/5AI can substantially speed up data entry via auto-tagging, transcription, and metadata suggestion, letting technicians focus on verification and complex judgment calls.
Task automatabilityclaude-haiku-4-5-202510014/5Data entry from structured museum records into databases is largely routine and rule-based. Current AI with OCR and structured extraction can handle 70–80% of typical collection metadata (object ID, artist, date, dimensions, materials), though complex provenance, conservation history, and context-dependent fields may require human review.
Task automatabilityclaude-sonnet-54/5Data entry of collection metadata into structured databases is a repetitive, well-defined text/data task that AI (with OCR, NLP extraction, and integration) can perform with significant time savings, though some entries require domain judgment or physical inspection first.
Adoption barriersclaude-haiku-4-5-202510012/5No legal requirement mandates human data entry for museum collections; institutional policies and preference for accuracy-checked input provide modest friction, but nothing prevents substitution of automated solutions.
Adoption barriersclaude-sonnet-52/5No licensing requirement for data entry, but institutional standards, provenance accuracy needs, and curatorial oversight create some friction against fully unsupervised automation.
Cost vs. human wageclaude-haiku-4-5-202510014/5AI-assisted data entry (OCR + structured extraction) costs are typically 1/3 to 1/5 the cost of manual technician data entry at museum wages, accounting for infrastructure, oversight, and residual human correction of difficult records.
Cost vs. human wageclaude-sonnet-54/5Automated data extraction and entry tools are far cheaper per record than skilled technician time once set up, though initial integration with legacy museum database systems adds cost.
Technical feasibility todayclaude-haiku-4-5-202510014/5Multiple mature products (including specialized museum management software with AI-assisted data extraction, general OCR, and low-code database tools) are deployed in museum operations today. Error rates on standard fields are low; complex or handwritten records still see material mistakes requiring human correction.
Technical feasibility todayclaude-sonnet-53/5Products exist for automated cataloging, OCR of accession records, and database population, but museums still rely heavily on manual verification and many collections management systems lack seamless AI-driven entry pipelines in production.

Photograph objects for documentation.

48

CI 39–57 · exposure 53 · augmentation 75 · importance 4.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Adoption is uneven and slow; large museums and digital heritage projects pilot robotic/AI systems, but small and mid-size institutions rely on human photographers. The sector is traditionally conservative, digitization budgets are constrained, and artifact safety concerns limit aggressive automation rollout.
Sector adoption velocityclaude-sonnet-52/5Museums and cultural heritage institutions are generally slow adopters of AI/automation due to budget constraints, specialized equipment needs, and small-scale, artisanal workflows.
Augmentation potentialclaude-haiku-4-5-202510014/5AI excels at assisting human photographers: automated lighting suggestions, instant focus-stack merging, background removal, and multi-angle composition guidance significantly boost productivity. Museum technicians remain in control of artifact handling and final quality judgment while AI handles repetitive, compute-intensive tasks.
Augmentation potentialclaude-sonnet-54/5AI tools can assist with image enhancement, automatic tagging, metadata generation, and quality checks, meaningfully speeding up the documentation workflow while a human still handles object positioning and photography decisions.
Task automatabilityclaude-haiku-4-5-202510014/5AI-assisted photography workflows can now handle object positioning, lighting optimization, and multi-angle capture with minimal human intervention. Modern computer vision systems can automatically detect optimal framing and exposure, reducing setup and post-processing time by 50%+ for standard documentation tasks, though human judgment on artistic/archival quality still adds value.
Task automatabilityclaude-sonnet-53/5Basic photographic capture can be partially automated with camera automation and AI-assisted image processing, but proper conservation documentation requires careful staging, lighting, scale references, and condition-specific angles that still need human setup and judgment.
Adoption barriersclaude-haiku-4-5-202510013/5Museums face moderate friction: conservators prefer human judgment on sensitive/irreplaceable artifacts, institutional resistance to replacing skilled staff, and liability concerns if automated systems damage priceless objects during positioning or lighting. No legal requirement for human photography, but professional standards and artifact risk create practical barriers.
Adoption barriersclaude-sonnet-52/5No licensing requirement exists, but institutional standards, artifact fragility, insurance requirements, and the need for trained handling of fragile or valuable objects create moderate friction against full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5High-end robotic photography systems and AI-enhanced workflows carry significant capital and integration costs. While per-shot costs can be low at scale, total cost of ownership (equipment, maintenance, calibration, oversight) often exceeds trained human photographer wages for typical museum documentation volumes.
Cost vs. human wageclaude-sonnet-52/5Specialized imaging equipment, controlled lighting setups, and calibration still require human operators and expensive hardware, so cost savings versus a technician's time are modest rather than order-of-magnitude.
Technical feasibility todayclaude-haiku-4-5-202510013/5Deployed products (robotic camera systems, AI lighting adjustment tools, automated focus stacking) exist in museum and conservation settings, but typically require careful operator oversight and fail on unusual artifacts or complex lighting scenarios. Production use is growing but not yet universal or fully reliable across diverse object types.
Technical feasibility todayclaude-sonnet-52/5Automated imaging rigs and AI-assisted cataloging tools exist in some large museums, but most institutions still rely on manual photography by trained staff; no widely deployed product fully handles conservation-grade documentation photography end-to-end.

Classify and assign registration numbers to artifacts and supervise inventory control.

36

CI 34–39 · exposure 41 · augmentation 63 · importance 4.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Museum adoption of AI for curatorial tasks lags significantly; most institutions are small, risk-averse regarding artifact data, and operate with legacy systems and limited digitization; pilot projects exist but production deployment of AI-driven registration and classification remains sparse.
Sector adoption velocityclaude-sonnet-52/5Museums and cultural heritage institutions are generally slow, under-resourced, and cautious adopters of AI compared to fast-moving digital-native or finance/professional-service sectors.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist technicians by auto-suggesting classifications based on visual features and automating routine data entry, allowing humans to focus on complex judgments about provenance and condition, but the augmentation remains supplementary rather than transformative given the centrality of expert human judgment in conservation work.
Augmentation potentialclaude-sonnet-54/5AI can meaningfully assist with image-based artifact matching, drafting catalog descriptions, database organization, and flagging inventory discrepancies, significantly aiding conservators while they retain final judgment.
Task automatabilityclaude-haiku-4-5-202510013/5AI can automate parts of this task—visual classification of artifacts, number assignment based on predefined rules, and basic inventory tracking are feasible—but full end-to-end automation falls short of 50% time savings at equal quality due to the need for domain expertise, contextual judgment about artifact condition and provenance, and human verification of classification accuracy and registration integrity.
Task automatabilityclaude-sonnet-53/5Classification and registration numbering can be substantially aided by AI (image recognition, database matching, cataloging assistance), but final classification of unique artifacts often requires expert judgment and supervision of inventory control requires human oversight and accountability.
Adoption barriersclaude-haiku-4-5-202510014/5Museums face strong organizational and curatorial barriers: human expertise and accountability for artifact authenticity and proper stewardship are deeply embedded in professional standards, internal governance, and insurance requirements; many institutions require human sign-off on registration and classification decisions.
Adoption barriersclaude-sonnet-53/5No licensing requirement, but professional standards, provenance/legal documentation needs, institutional accountability, and the requirement for expert judgment on unique objects create meaningful friction against full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI systems for image recognition and database integration require substantial setup, domain training data, and ongoing human review to catch misclassifications, making the total cost per artifact registered comparable to or higher than a technician's loaded wage for the same artifact.
Cost vs. human wageclaude-sonnet-52/5AI tools can reduce time on data entry and cataloging support, but given the need for expert verification, integration with museum-specific databases, and low volume of specialized artifacts, cost savings are modest rather than order-of-magnitude.
Technical feasibility todayclaude-haiku-4-5-202510012/5While image recognition and database management tools exist, no mature production system reliably handles the full scope of artifact classification across diverse museums and collection types with the accuracy required for curatorial trust; most deployments remain at pilot stage with significant manual oversight.
Technical feasibility todayclaude-sonnet-52/5Some collection management software includes AI-assisted tagging/classification features, but production-grade systems reliably classifying diverse artifacts and independently supervising inventory control are not widely deployed.

Prepare reports on the operation of conservation laboratories, documenting the condition of artifacts, treatment options, and the methods of preservation and repair used.

36

CI 25–47 · exposure 33 · augmentation 63 · importance 4.0/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Museums and conservation labs are traditionally low-tech organizations with slow digitization and limited AI adoption. Most are small or mid-sized institutions with conservative practices; production deployment of AI in artifact documentation remains extremely rare in the sector.
Sector adoption velocityclaude-sonnet-52/5Museums and conservation labs are typically small, specialized, and slow to adopt AI tools compared to fast-moving digital-native sectors, though general AI writing tools are creeping in for administrative tasks.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist conservators by organizing data on artifact conditions, suggesting literature references, or drafting preliminary text summaries that the expert then refines. However, the core task of professional judgment and documentation requires sustained human expertise, so augmentation potential is moderate rather than transformative.
Augmentation potentialclaude-sonnet-54/5AI can meaningfully speed up drafting, organizing, and standardizing report language once a conservator provides observations and treatment details, improving efficiency while the human remains central to assessment.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can draft sections summarizing condition assessments or suggest treatment options from documentation, the task fundamentally requires expert conservator judgment about artifact condition, conservation ethics, and individualized treatment plans. Current AI cannot reliably perform the full task of preparing authoritative conservation reports without substantial human expert review and revision.
Task automatabilityclaude-sonnet-53/5AI can draft report text, summarize condition notes, and structure documentation from dictated or typed observations, but the underlying visual condition assessment and treatment decisions require human expertise that must feed into the report.'
Adoption barriersclaude-haiku-4-5-202510014/5Conservation work on artifacts of cultural, scientific, or historical significance is subject to institutional governance, professional ethics codes (ICOM, AIC), and often legal accountability for treatment decisions. Official reports on artifact condition and preservation methods carry institutional and sometimes legal liability, creating strong barriers to fully automated reporting.
Adoption barriersclaude-sonnet-52/5No strict licensing requirement mandates a human write these reports, but institutional standards, accountability for artifact condition records, and specialized terminology create moderate friction against full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5The task requires specialized expertise from trained conservators whose loaded wages are substantial. AI tools may reduce drafting time but cannot replace the high-value expert judgment required, making the cost ratio unfavorable for full automation relative to human conservator labor.
Cost vs. human wageclaude-sonnet-53/5Using AI writing assistance for report drafting is cheap relative to a technician's time, but the bulk of task value lies in expert observation and judgment that AI cannot substitute, keeping overall cost savings moderate.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed AI system reliably performs end-to-end conservation laboratory reporting. AI can assist with documentation organization or draft text, but production systems do not exist that independently generate authoritative conservation condition assessments and treatment recommendations that meet professional standards.
Technical feasibility todayclaude-sonnet-52/5General LLM tools can assist with drafting and formatting such reports, but no specialized production system reliably generates conservation condition reports from artifact examination without substantial human input and correction.

Recommend preservation procedures, such as control of temperature and humidity, to curatorial and building staff.

26

CI 23–30 · exposure 25 · augmentation 63 · importance 3.9/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Museums and cultural institutions are slower-moving sectors with high conservatism around artifact safety and curatorial authority. Adoption of AI for preservation decisions is negligible; institutions retain expert staff for this high-stakes function.
Sector adoption velocityclaude-sonnet-52/5Museums and cultural heritage institutions are generally slow adopters of AI in specialized preservation science, with pilots for monitoring systems but little production-level AI-driven recommendation practice.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist a conservator by drafting summaries of standard procedures, flagging relevant literature, or cross-checking environmental parameters against best practices—useful support that raises efficiency. However, the human expert must retain full decision authority and contextual judgment.
Augmentation potentialclaude-sonnet-54/5AI can help synthesize research literature, analyze sensor/environmental data trends, and draft recommendation reports, significantly aiding conservators while they retain final judgment and sign-off.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can retrieve and summarize standard preservation guidelines from literature, recommending procedures requires contextual judgment about specific artifacts, facility constraints, and institutional priorities. Current AI cannot reliably assess physical condition or environmental trade-offs at the level needed to replace human expert recommendation.
Task automatabilityclaude-sonnet-52/5This task requires expert judgment about specific artifact conditions, materials science, and institutional context that AI cannot independently assess or recommend end-to-end without human expertise driving the analysis.chercher.dr Only limited sub-parts (e.g., summarizing standard guidelines) could be automated.
Adoption barriersclaude-haiku-4-5-202510014/5Significant barriers exist: professional liability for preservation failures rests with certified conservators; many institutions require a licensed or certified individual to sign off on preservation procedures; regulatory compliance (e.g., museum accreditation standards) often mandates human expertise. Legal and reputational risk strongly favors human authority.
Adoption barriersclaude-sonnet-53/5No strict licensing mandate exists for this specific task, but institutional liability, insurance requirements, and reliance on professional conservation expertise create meaningful friction against pure AI substitution.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI inference cost is low, but the overhead of human expert review, institutional validation, and potential error correction makes the all-in cost comparable to a senior technician or conservator reviewing standards directly. No cost advantage materializes without removing human oversight.
Cost vs. human wageclaude-sonnet-52/5While AI-assisted research or drafting is cheap, the specialized expertise needed to interpret object-specific conservation needs still requires a trained conservator, keeping all-in costs comparable to human-led work.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed product reliably generates site-specific preservation recommendations that institutions depend on for decision-making. AI can draft generic guidelines or flag best practices, but production systems do not yet perform this task end-to-end with the curatorial authority and liability acceptance required.
Technical feasibility todayclaude-sonnet-52/5No deployed product autonomously issues preservation recommendations to museum staff; existing AI tools (chatbots, sensor dashboards) support but don't replace the conservator's judgment-driven recommendation process.

Estimate cost of restoration work.

24

CI 18–30 · exposure 20 · augmentation 50 · importance 3.1/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Museum and conservation sectors are digitization laggards with limited IT investment, small teams, and strong professional cultures favoring established expertise. Even large museums rarely automate cost estimation beyond basic templating; adoption remains minimal.
Sector adoption velocityclaude-sonnet-51/5Museums and conservation are a small, specialized, low-digitization sector with minimal AI adoption for such judgment-heavy estimating tasks.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist by retrieving historical pricing data, suggesting material costs, and organizing estimate templates, improving a conservator's productivity on data gathering and calculation portions. However, the core judgment task (assessing condition and selecting approach) remains human-led.
Augmentation potentialclaude-sonnet-53/5AI can help by referencing past cost data, generating estimate templates, or aiding documentation, providing moderate assistance while the conservator retains judgment.
Task automatabilityclaude-haiku-4-5-202510012/5Cost estimation for restoration requires domain expertise, assessment of material condition, rarity, and conservation standards. While AI can assist with price lookup and basic calculations, the judgment-intensive aspects—determining appropriate materials, techniques, and labor—remain heavily dependent on human conservator expertise and cannot meet the 50% time-saving bar end-to-end.
Task automatabilityclaude-sonnet-52/5Cost estimation for restoration requires physical inspection of artifact condition, material assessment, and specialized labor pricing that AI cannot directly observe or judge without human expert input.,
Adoption barriersclaude-haiku-4-5-202510014/5Museum conservation is subject to professional standards (AIC, ICOMOS), insurance and liability requirements, and institutional protocols requiring expert sign-off. Estimates drive funding and donor communication, creating reputational and financial risk that mandates human expert authority.
Adoption barriersclaude-sonnet-53/5No strict licensing mandates AI cannot do this, but institutional liability and reliance on trusted conservator judgment create moderate friction against automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5Current AI tools (basic data extraction, estimation templates) are cheaper per inference than a senior conservator hour, but integration, validation, and oversight overhead is substantial. The full system cost approaches or exceeds the human wage for meaningful output quality.
Cost vs. human wageclaude-sonnet-52/5Human conservators' expertise is essential for accurate estimates, and any AI assistance still requires substantial expert oversight, limiting cost savings.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed products reliably perform independent restoration cost estimation at the quality required by museums. Spreadsheet and database tools exist for inventory, but specialized conservation cost modeling is largely manual; research prototypes may exist but production systems are absent.
Technical feasibility todayclaude-sonnet-51/5No deployed products estimate conservation restoration costs reliably; this remains a niche expert judgment task not addressed by production AI tools.

Notify superior when restoration of artifacts requires outside experts.

23

CI 18–28 · exposure 20 · augmentation 38 · importance 3.7/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Museum conservation is a small, specialized, low-digitization sector with high reliance on in-person expertise and established professional networks. Adoption of automated escalation systems in this domain is minimal.
Sector adoption velocityclaude-sonnet-51/5Museum conservation is a small, specialized, low-digitization field with minimal AI deployment for hands-on judgment tasks like this; adoption is a laggard sector.
Augmentation potentialclaude-haiku-4-5-202510013/5AI could usefully assist by flagging potential damage patterns or suggesting expertise categories (e.g., 'textile conservation'), helping conservators organize their assessment—but the human expert must remain the decision-maker.
Augmentation potentialclaude-sonnet-52/5AI could help draft notification messages or track condition documentation, but it contributes little to the core judgment of when outside expertise is truly needed.
Task automatabilityclaude-haiku-4-5-202510012/5While AI could potentially flag when artifacts appear to need specialized expertise through image analysis or damage assessment, the task requires nuanced judgment about the scope and severity of conservation needs—which often hinges on subtle professional standards and organizational context. Current systems cannot reliably make this determination end-to-end without substantial human review.
Task automatabilityclaude-sonnet-52/5This is a brief judgment-and-communication task requiring assessment of artifact condition and organizational escalation, which AI cannot independently determine or execute reliably.expertise identification and internal reporting norms are context-specific.this task is largely embedded in professional judgment rather than a discrete automatable workflow.
Adoption barriersclaude-haiku-4-5-202510014/5Museum conservation is governed by professional standards, ethical codes, and organizational protocols that strongly favor human expert judgment in escalation decisions. Curators and conservators are expected to take professional responsibility for determining when outside expertise is needed, creating substantial friction against substitution.
Adoption barriersclaude-sonnet-53/5While no strict licensing law mandates a human notify superiors, professional norms, accountability for irreplaceable cultural artifacts, and institutional decision-making structures create meaningful friction against full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5The task is lightweight communication (~few minutes of expert judgment) with minimal cost; an AI system would need integration, training, and oversight that may exceed the cost of direct professional assessment. The human cost per instance is already low.
Cost vs. human wageclaude-sonnet-52/5Since AI cannot reliably perform the underlying judgment (assessing when expertise is needed), any AI involvement still requires a human conservator's assessment, so cost savings are minimal to none.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed product reliably performs this integrated task in museum conservation workflows today. Damage-detection systems exist but are narrow in scope and require museum professionals to interpret them and make the final escalation decision.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously assesses artifact restoration needs and escalates decisions to a supervisor; this remains a human judgment and communication task in practice.

Study object documentation or conduct standard chemical and physical tests to ascertain the object's age, composition, original appearance, need for treatment or restoration, and appropriate preservation method.

21

CI 18–25 · exposure 20 · augmentation 63 · importance 4.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Museums have digitized records and adopted imaging tools, but core conservation decisions remain human-led. Adoption of AI for autonomous decision-making in conservation is slow; tools are augmentative rather than substitutive, and sector digitization is uneven across institutions.
Sector adoption velocityclaude-sonnet-51/5Museum conservation is a small, specialized, non-digitized field with minimal AI production deployment; adoption of AI tools here lags far behind sectors like finance or general office work.
Augmentation potentialclaude-haiku-4-5-202510014/5AI augmentation is strong in this domain: image analysis for damage detection, spectroscopy data interpretation, composition matching against databases, and condition documentation all meaningfully assist conservators in making faster, better-informed decisions while the expert remains in the loop.
Augmentation potentialclaude-sonnet-53/5AI can help analyze historical documentation, cross-reference provenance records, or assist in image analysis of surface conditions, providing moderate assistance to conservators during research phases.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist with some analytical steps (e.g., image analysis, chemical database lookups), the task requires hands-on physical testing, contextual judgment about object condition, and integrating multiple data sources into a comprehensive conservation plan. Current AI lacks the sensorimotor capability and domain expertise to conduct standard chemical/physical tests end-to-end or to reliably determine preservation methods without expert human oversight.
Task automatabilityclaude-sonnet-52/5Interpreting chemical/physical test results and hands-on inspection of physical artifacts requires expert judgment and manual procedures that AI cannot perform end-to-end; AI can assist with documentation review but not the physical testing or final determination.
Adoption barriersclaude-haiku-4-5-202510014/5Museum conservation involves professional credentials (conservator certifications), liability for irreversible treatment decisions on irreplaceable objects, legal responsibility for institutional collections, and institutional trust in human expertise. These create substantial barriers to replacing human judgment with automated systems.
Adoption barriersclaude-sonnet-54/5Conservation of valuable/irreplaceable cultural objects carries high liability, and professional certification and ethical standards typically require a trained conservator's judgment and sign-off on treatment decisions.
Cost vs. human wageclaude-haiku-4-5-202510012/5Deploying AI analytical tools (imaging, spectroscopy software, databases) requires significant infrastructure and integration costs. The loaded cost of a conservator is high, but so is the overhead of maintaining reliable automated analysis pipelines, making the cost ratio unfavorable for full substitution.
Cost vs. human wageclaude-sonnet-52/5Physical testing equipment, sample handling, and expert interpretation still require paid conservator time and lab costs; AI might speed literature review but doesn't replace the core costly physical analysis steps.
Technical feasibility todayclaude-haiku-4-5-202510012/5Computer vision and chemical analysis tools exist in research settings, but no deployed production systems reliably perform the full task—ascertaining age, composition, original appearance, treatment needs, and preservation methods—independently. Museums still depend on trained conservators to interpret results and make critical decisions.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product independently conducts physical/chemical analysis of art objects or determines conservation treatment plans in production settings; this remains a specialist human task with only research-stage AI applications in materials analysis.

Coordinate exhibit installations, assisting with design, constructing displays, dioramas, display cases, and models, and ensuring the availability of necessary materials.

20

CI 5–35 · exposure 13 · augmentation 50 · importance 3.6/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Museums remain relatively low-digitization, conservative organizations with limited resources for AI infrastructure. Adoption of AI-assisted tools in exhibit design or logistics is slow and mostly limited to large institutions; production deployment of AI agents for installation coordination is rare.
Sector adoption velocityclaude-sonnet-51/5Museums and cultural institutions are a small, slow-digitizing sector with minimal AI deployment for physical exhibit construction and installation work.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can usefully assist with design visualization, material inventory tracking, supply scheduling, and budget forecasting, helping museum technicians plan and coordinate more efficiently. However, the high creative and physical demands of the task limit the transformative impact of augmentation.
Augmentation potentialclaude-sonnet-53/5AI can assist with design visualization, 3D modeling, layout planning, and inventory/material tracking, offering moderate productivity gains even though the physical construction itself remains manual.
Task automatabilityclaude-haiku-4-5-202510012/5While AI could assist with design visualization, supply chain coordination, and material planning documentation, the physical construction of displays, dioramas, and cases—requiring spatial judgment, hands-on assembly, and real-time problem-solving—cannot be automated end-to-end today. The task requires on-site coordination and creative judgment that falls well short of 50% time savings at equal quality.
Task automatabilityclaude-sonnet-51/5This is a hands-on physical coordination and construction task involving fabrication, spatial design, and material logistics that current AI cannot perform end-to-end; robotics and physical dexterity for custom exhibit builds are far from off-the-shelf capability.
Adoption barriersclaude-haiku-4-5-202510014/5Museum conservation and exhibit installation involve significant liability, aesthetic judgment, and preservation standards that create organizational friction. Curators and conservators often retain final authority over display decisions; regulatory and professional standards also protect human expertise in artifact handling and installation safety.
Adoption barriersclaude-sonnet-52/5No licensing requirement exists, but physical presence, custom fabrication skill, and hands-on coordination with vendors and staff create practical friction against any automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5Current AI systems capable of meaningful assistance on design or logistics are expensive relative to their actual time savings on this mixed physical-cognitive task. The cost of AI infrastructure, setup, and human oversight exceeds the efficiency gain for most museum operations.
Cost vs. human wageclaude-sonnet-51/5There is no AI system substituting for the physical labor, craftsmanship, and on-site coordination involved, so AI cost comparison is not meaningful—human labor remains the only viable option.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed product reliably performs end-to-end exhibit installation coordination. AI can support scheduling and inventory management in narrow applications, but the integration of design consultation, material logistics, and physical assembly oversight remains research-stage or prototype-only.
Technical feasibility todayclaude-sonnet-51/5No deployed product coordinates physical exhibit installation, builds display cases/dioramas, or manages on-site material logistics; this remains entirely human-executed skilled craft work.

Perform tests and examinations to establish storage and conservation requirements, policies, and procedures.

18

CI 14–23 · exposure 20 · augmentation 50 · importance 3.9/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Museums and cultural institutions are traditionally conservative, slow to digitize workflows, and strongly prefer human expert oversight for irreplaceable collections; automation adoption in this sector remains minimal and pilot-stage at best.
Sector adoption velocityclaude-sonnet-51/5Museum conservation is a small, specialized, low-digitization field with minimal AI production deployment for physical testing tasks.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can meaningfully assist conservators by analyzing test data, suggesting storage parameters based on condition reports, and helping draft documentation, thereby accelerating expert review cycles while the conservator retains decision-making authority.
Augmentation potentialclaude-sonnet-53/5AI can assist by analyzing spectroscopy data, researching material degradation literature, or drafting conservation policy documents, aiding but not replacing expert physical examination.
Task automatabilityclaude-haiku-4-5-202510012/5While AI could assist in analyzing test data and generating preliminary reports on storage conditions, the core task requires specialized domain knowledge, judgment about unique artifacts, and hands-on examination that current systems cannot reliably perform end-to-end. AI might automate 20–30% of documentation and analysis, falling well short of the 50% threshold.
Task automatabilityclaude-sonnet-52/5This requires physical scientific testing (material analysis, environmental sensing) and expert judgment on fragile, unique objects that AI cannot physically perform, though AI can help interpret some data.'
Adoption barriersclaude-haiku-4-5-202510014/5Significant barriers exist: conservation and curation decisions carry legal/insurance liability if automated carelessly, institutional boards typically require credentialed conservators to sign off on policies, and direct handling of valuable artifacts mandates in-person expert judgment that cannot be delegated to AI systems.
Adoption barriersclaude-sonnet-54/5Conservation decisions carry high liability for irreversible damage to valuable/cultural heritage objects, requiring trained human conservators with specialized certification and judgment.
Cost vs. human wageclaude-haiku-4-5-202510012/5Current AI tools for artifact analysis are specialized, require expert integration and validation, and cannot yet reduce the need for human conservators; the all-in cost remains comparable to or higher than employing trained staff.
Cost vs. human wageclaude-sonnet-51/5AI cannot replace the physical instrumentation, sampling, and expert hands-on examination required, so it offers no cost substitution for this labor.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed product reliably performs end-to-end conservation assessment and policy generation; existing AI tools lack the multi-modal sensory understanding and specialized conservator judgment needed to evaluate specific artifacts' requirements at production scale.
Technical feasibility todayclaude-sonnet-51/5No deployed product performs hands-on conservation testing and examination of physical artifacts; this remains a specialized, research-stage application at best.

Repair, restore, and reassemble artifacts, designing and fabricating missing or broken parts, to restore them to their original appearance and prevent deterioration.

15

CI 5–25 · exposure 13 · augmentation 50 · importance 4.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Museum and conservation sectors are digitization-laggard, with small teams, long tenures, and risk-averse cultures. AI adoption remains experimental in documentation and analysis; production deployment for actual restoration work is minimal.
Sector adoption velocityclaude-sonnet-51/5Museum conservation is a small, highly specialized, low-digitization craft sector with minimal AI or robotics adoption for physical restoration tasks.
Augmentation potentialclaude-haiku-4-5-202510013/5AI tools meaningfully assist conservators through high-resolution imaging analysis, 3D reconstruction for design reference, condition documentation, and CAD support for missing-part prototyping. These augment human decision-making and craftsmanship without replacing the conservator's expertise and judgment.
Augmentation potentialclaude-sonnet-53/5AI can assist with research, image analysis, 3D scanning/modeling for missing parts, and material identification, aiding planning even though execution remains manual.
Task automatabilityclaude-haiku-4-5-202510012/5Physical artifact restoration—design, fabrication, assembly—requires hands-on manipulation, material judgment, and spatial reasoning in unstructured environments. Current AI can assist with design visualization and documentation but cannot autonomously perform the mechanical work, reducing automation potential to modest parts like imaging and CAD proposals.
Task automatabilityclaude-sonnet-51/5This is fine-motor, hands-on physical restoration work requiring manual dexterity, material science judgment, and artistic skill that current AI cannot perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Museums and heritage institutions operate under strict ethical and legal standards for artifact handling; irreversible damage carries high liability. Professional conservators are often required by accreditation and insurance, and institutional policies strongly favor human expertise for irreplaceable objects.
Adoption barriersclaude-sonnet-54/5Conservation of valuable, often irreplaceable cultural artifacts requires specialized certified expertise and carries high liability for irreversible damage, though no formal licensing regime universally applies.
Cost vs. human wageclaude-haiku-4-5-202510012/5Current AI tooling for conservation (imaging, design assistance) costs overlap with or exceed specialized conservator labor when integration, oversight, and error-correction are factored in. The manual dexterity and domain expertise of human conservators remain comparatively cost-effective.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute for the physical labor involved, so any AI cost comparison is moot; human conservators remain the only option.
Technical feasibility todayclaude-haiku-4-5-202510012/5While AI can support documentation and design planning, no deployed product reliably performs end-to-end artifact repair and restoration autonomously. Existing systems are research-stage or narrow (e.g., crack detection); human conservators remain essential for quality and liability.
Technical feasibility todayclaude-sonnet-51/5No deployed AI or robotic product performs physical artifact repair, fabrication of replacement parts, and reassembly; this remains a specialized human craft.

Specialize in particular materials or types of object, such as documents and books, paintings, decorative arts, textiles, metals, or architectural materials.

15

CI 5–25 · exposure 13 · augmentation 50 · importance 3.9/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Museums and conservation labs are moderate-to-lagging adopters of AI; most remain in pilot phases with imaging and documentation tools. Institutional conservatism, small team sizes, low digital infrastructure maturity, and risk-averse governance slow deep production adoption of AI agents in actual conservation work.
Sector adoption velocityclaude-sonnet-51/5Museum conservation is a small, specialized, low-digitization field with minimal AI deployment or production use in the physical craft of conservation.
Augmentation potentialclaude-haiku-4-5-202510013/5AI tools (spectral imaging, 3D scanning, database systems, condition-tracking software) usefully assist conservators in documentation, analysis, and record-keeping, boosting productivity on information-heavy tasks while the expert remains in full control of treatment decisions.
Augmentation potentialclaude-sonnet-53/5AI can assist with research, material analysis via imaging/spectral data interpretation, and documentation, aiding but not replacing the specialist's judgment and manual skill.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist with material identification (imaging, analysis) and documentation, conservation decisions require physical judgment, hands-on manipulation, condition assessment, and ethical expertise in preservation that current systems cannot perform end-to-end. AI can automate only specific substeps like cataloging or initial triage, not the specialized decision-making that drives the work.
Task automatabilityclaude-sonnet-51/5Specializing in a material domain involves years of hands-on training, tacit sensory judgment, and physical conservation skill that current AI cannot replicate or substitute for.
Adoption barriersclaude-haiku-4-5-202510014/5Significant legal and professional barriers protect this work: conservation ethics codes, institutional liability (damage to irreplaceable artifacts), museum accreditation standards, and the requirement for licensed/credentialed conservators to sign off on interventions. Substituting AI for human specialists creates unacceptable legal and reputational risk.
Adoption barriersclaude-sonnet-54/5Conservation of valuable cultural artifacts typically requires certified expertise, institutional accreditation, and hands-on physical handling, creating strong professional and liability barriers.
Cost vs. human wageclaude-haiku-4-5-202510012/5The cost of AI systems (hardware, imaging, software, integration) plus the mandatory human specialist oversight and final decision-making remains comparable to or higher than direct human labor, especially for the rare, high-value objects that drive conservation demand.
Cost vs. human wageclaude-sonnet-51/5There is no AI system performing this specialization at all, so no cost comparison favors AI; human expertise remains the only viable option.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed product reliably performs material-specialized conservation work at production scale. Research prototypes exist for image analysis and condition assessment, but real conservation requires expert human judgment integrated with physical intervention that remains far beyond current automation.
Technical feasibility todayclaude-sonnet-51/5No deployed product performs material-specific conservation expertise; this is a professional specialization built through apprenticeship and practice, not a discrete automatable task.

Plan and conduct research to develop and improve methods of restoring and preserving specimens.

14

CI 5–23 · exposure 13 · augmentation 50 · importance 3.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Museums and cultural institutions are typically low-digitization, conservative sectors with limited AI adoption in core conservation activities. Adoption of AI for research planning is laggard, with most institutions relying on established professional practices and incremental improvements by trained staff.
Sector adoption velocityclaude-sonnet-51/5Museum conservation is a small, specialized, low-digitization field with minimal AI adoption for research tasks.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can usefully assist by analyzing conservation literature, identifying relevant research patterns, suggesting testing parameters, and organizing experimental data, moderately raising the efficiency of a human conservator's research planning. However, the core evaluation and decision-making remain human-driven.
Augmentation potentialclaude-sonnet-53/5AI can assist by summarizing literature, suggesting relevant prior studies, or analyzing data from experiments, aiding but not replacing the research process.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist with literature review, data analysis, and hypothesis generation for conservation methods, the core task requires hands-on experimentation, material assessment, and iterative physical testing that cannot be automated end-to-end. The experiential judgment and sensory evaluation needed to evaluate specimen condition and preservation outcomes remain primarily human domain.
Task automatabilityclaude-sonnet-51/5This is original scientific/experimental research requiring hands-on testing of materials on physical specimens, hypothesis generation, and expert judgment that current AI cannot execute end-to-end.atur
Adoption barriersclaude-haiku-4-5-202510014/5Strong barriers exist: conservation decisions carry high liability for irreversible damage to artifacts, professional standards and institutional protocols require human expert sign-off, and museum accreditation often mandates human conservators direct preservation research. Regulatory and professional-ethical requirements protect this role.
Adoption barriersclaude-sonnet-54/5Conservation research on irreplaceable specimens carries high liability and typically requires credentialed expertise and institutional oversight, creating strong barriers to automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5Current AI solutions offer limited cost advantage because the task requires specialized domain expertise, physical experimentation, and custom workflows that demand significant human oversight and integration effort. The all-in cost of AI assistance likely exceeds the loaded wage of a trained conservator for this specialized work.
Cost vs. human wageclaude-sonnet-51/5AI cannot substitute for the physical experimentation and specialized labor involved, so there is no meaningful cost displacement; human conservators remain necessary.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed products reliably conduct conservation research independently. AI tools exist for document analysis and pattern recognition in datasets, but conservation-specific research—involving physical specimens, chemical testing, and materials science—remains in the research/prototype stage with minimal production deployment in museums.
Technical feasibility todayclaude-sonnet-51/5No deployed product plans or conducts conservation research; AI is at best a literature-search aid, not a research-execution system.

Determine whether objects need repair and choose the safest and most effective method of repair.

13

CI 5–21 · exposure 13 · augmentation 50 · importance 4.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Museums have been slow to adopt AI for core conservation decisions; adoption remains experimental and limited to larger institutions with digital infrastructure, and the high stakes of error create organizational resistance to displacement of expert judgment.
Sector adoption velocityclaude-sonnet-51/5Museums and conservation labs are a small, specialized, low-digitization sector with minimal AI deployment for physical object assessment and repair decisions.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can meaningfully assist conservators by analyzing high-resolution images, flagging damage patterns, and suggesting repair options from a database, allowing faster triage and broader option consideration, though the final judgment remains with the human expert.
Augmentation potentialclaude-sonnet-53/5AI can assist with research, documentation, image analysis, and material identification to inform a conservator's decision, though the core judgment and hands-on assessment remain human-driven.
Task automatabilityclaude-haiku-4-5-202510012/5Assessing whether objects need repair requires visual inspection, material knowledge, and condition judgment that AI can partially support through image analysis, but the determination of the safest repair method demands deep expertise, contextual judgment about irreversibility, and legal/ethical accountability that current AI systems cannot reliably provide end-to-end at equal quality.
Task automatabilityclaude-sonnet-51/5This requires hands-on physical inspection, material science expertise, and judgment about irreplaceable objects that current AI cannot perform end-to-end.'
Adoption barriersclaude-haiku-4-5-202510014/5Museum conservators are typically trained specialists subject to professional standards (AAM, IIC), and decisions about irreversible conservation interventions carry high legal and reputational liability, creating substantial institutional and ethical barriers to full automation.
Adoption barriersclaude-sonnet-54/5While not formally licensed like medicine, conservation carries high liability for damaging irreplaceable cultural objects, strong professional norms, and institutional oversight requiring trained human judgment.
Cost vs. human wageclaude-haiku-4-5-202510011/5The cost of integrating AI for damage assessment plus required human expert oversight and validation still exceeds the cost of direct expert assessment, and errors in repair selection can result in catastrophic object loss far exceeding labor savings.
Cost vs. human wageclaude-sonnet-51/5AI cannot substitute for the physical examination and decision-making, so there is no meaningful AI cost basis for comparison against skilled conservator labor.
Technical feasibility todayclaude-haiku-4-5-202510012/5While AI systems can assist with object detection and damage classification in images, no deployed product reliably performs the full task of diagnosing repair necessity and selecting the safest method in production museum settings; conservation remains largely human-expert driven with AI in supporting roles only.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously assesses conservation needs and selects repair methods for physical artifacts; this remains a specialized human craft/science.

Prepare artifacts for storage and shipping.

11

CI 5–16 · exposure 8 · augmentation 38 · importance 4.1/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Museums are traditionally slower to adopt automation due to risk aversion, small budgets, and the specialized nature of conservation work; digital cataloging has seen adoption, but physical handling automation remains rare in practice.
Sector adoption velocityclaude-sonnet-51/5Museum conservation is a low-digitization, highly manual field with minimal AI adoption for physical artifact handling tasks.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can meaningfully assist with imaging, condition documentation, storage location optimization, and shipping logistics planning, helping conservators work more efficiently while they retain full responsibility for hands-on preparation.
Augmentation potentialclaude-sonnet-52/5AI can help with documentation, condition-report generation, or logistics planning around shipping, but offers little assistance with the physical preparation and handling itself.
Task automatabilityclaude-haiku-4-5-202510012/5While AI could assist with documentation, cataloging, and planning workflows, the task involves physical handling of fragile, irreplaceable items requiring tactile judgment, environmental assessment, and adaptive decision-making that current AI systems cannot perform end-to-end. Robotic arms exist but lack the dexterity and contextual reasoning for delicate artifact preparation.
Task automatabilityclaude-sonnet-51/5This task requires physical handling, packing, and stabilization of often fragile, unique artifacts, which demands fine motor skill, tactile judgment, and material-specific knowledge that current AI systems cannot perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Museum conservation carries high legal liability if artifacts are damaged; cultural institutions have strong regulatory oversight, insurance requirements, and professional licensing standards that mandate human expert judgment and accountability for irreplaceable items.
Adoption barriersclaude-sonnet-54/5Handling irreplaceable, high-value artifacts typically requires trained conservators due to liability, insurance requirements, and institutional protocols mandating human expertise and accountability.
Cost vs. human wageclaude-haiku-4-5-202510011/5AI-based documentation systems cost less than human labor for cataloging, but the physical preparation itself cannot be automated, so overall cost savings are minimal when a trained conservator must still perform the majority of the work.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing the physical labor involved, so AI cost comparison is moot; the human specialist remains the only cost-effective option.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed products perform the core physical task of artifact preparation, wrapping, crating, and condition assessment at production scale. Computer vision can document items, but the hands-on conservation work remains entirely human-dependent.
Technical feasibility todayclaude-sonnet-51/5No deployed AI or robotic product handles the physical preparation, cushioning, crating, or handling of museum artifacts for storage or shipping; this remains a manual conservation task.

Lead tours and teach educational courses to students and the general public.

11

CI 5–16 · exposure 0 · augmentation 50 · importance 3.1/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Museum and cultural institutions are relatively low-digitization, laggard sectors in AI adoption; tours remain primarily human-led, and educational programming remains centered on live instruction.
Sector adoption velocityclaude-sonnet-52/5Museums and cultural institutions are generally slow adopters of AI-driven guiding/teaching tools relative to information-sector benchmarks, with pilots (e.g., AI audio guides) still uncommon and narrow.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can usefully assist museum educators by generating tour scripts, providing real-time curatorial information, or offering translation support, but the core teaching and engagement function remains human-centered.
Augmentation potentialclaude-sonnet-53/5AI can help prepare educational materials, generate tour scripts, answer visitor questions via kiosks, or provide translation support, meaningfully aiding preparation and supplementary materials even though it doesn't replace the live task.
Task automatabilityclaude-haiku-4-5-202510011/5Leading tours and teaching require live interaction, real-time responsiveness to questions, physical presence in galleries, and nuanced engagement with diverse audiences—capabilities that current AI systems cannot deliver end-to-end in a meaningful way today.
Task automatabilityclaude-sonnet-51/5Leading live tours and teaching involves real-time physical presence, adapting to audience questions, and interpersonal engagement that current AI cannot replicate end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Strong adoption barriers exist: museums and educational institutions prioritize human instructors for credibility and compliance, liability concerns over autonomous educational delivery, and explicit customer preference for human expertise and mentorship in learning contexts.
Adoption barriersclaude-sonnet-53/5No licensing requirement, but strong customer preference for human interaction, safety supervision of groups, and organizational reliance on human educators create moderate friction.
Cost vs. human wageclaude-haiku-4-5-202510011/5AI systems cannot yet deliver comparable educational experiences at lower cost; the technology required for autonomous guided tours and teaching engagement would exceed the human wage it aims to replace.
Cost vs. human wageclaude-sonnet-52/5AI audio-guide or chatbot alternatives are cheap per use, but they don't substitute for the guided, interactive teaching task, so true cost comparison for equivalent output favors humans still needing to be present.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product reliably performs the full task of leading museum tours or teaching educational courses in production settings; isolated components like automated content delivery exist, but not the integrated, interactive experience.
Technical feasibility todayclaude-sonnet-51/5No deployed product leads in-person museum tours or teaches live courses autonomously; at best AI provides audio guides or chatbot supplements, not the actual instructional task.

Install, arrange, assemble, and prepare artifacts for exhibition, ensuring the artifacts' safety, reporting their status and condition, and identifying and correcting any problems with the set up.

9

CI 5–14 · exposure 8 · augmentation 38 · importance 4.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Museums are typically slower adopters of automation due to their conservationist culture, limited digitization of artifact-handling workflows, small team sizes, and institutional preference for human expertise. Few museums have the capital or risk tolerance for experimental robotic systems in exhibition work.
Sector adoption velocityclaude-sonnet-51/5Museums and conservation are a small, low-digitization, physically-oriented sector with minimal AI/robotics adoption for hands-on artifact handling.
Augmentation potentialclaude-haiku-4-5-202510013/5AI assists meaningfully with condition documentation (image capture and analysis), inventory tracking, and digital records of artifact placement, raising curatorial efficiency. However, the hands-on judgment and physical execution remain human-led, making this augmentative rather than transformative.
Augmentation potentialclaude-sonnet-52/5AI can assist with condition documentation, exhibit planning software, or digital mock-ups, but offers little help with the physical setup and handling itself.
Task automatabilityclaude-haiku-4-5-202510012/5While AI could assist with condition reporting (image analysis, documentation), the physical installation and arrangement of artifacts requires dexterous manipulation, spatial judgment, and expert handling of fragile objects—tasks beyond current robot capabilities. The safety-critical nature and need for real-time problem-solving limit end-to-end automation well below the 50% time-saving threshold.
Task automatabilityclaude-sonnet-51/5This is a physical, dexterous task requiring hands-on manipulation of fragile, valuable artifacts plus real-time judgment about safety and condition; no AI system can perform physical installation or handling.
Adoption barriersclaude-haiku-4-5-202510014/5Museum work is heavily regulated by conservation standards, liability frameworks (artifact damage carries severe institutional and legal consequences), and professional accreditation. Human conservators must sign off on artifact condition and handling; automation would require legal and curatorial sign-off at each step, creating substantial organizational and regulatory friction.
Adoption barriersclaude-sonnet-54/5Handling irreplaceable, high-value artifacts carries major liability and conservation risk, typically requiring trained, credentialed conservators and institutional sign-off, creating strong barriers to automation.
Cost vs. human wageclaude-haiku-4-5-202510011/5The specialized robotic hardware required for safe artifact handling, combined with high integration and oversight costs, far exceeds the cost of skilled conservators who can adapt to diverse objects, environments, and regulatory requirements in museum settings.
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute performing this physical work, so the human is the only cost-effective option; AI cannot replace the labor at any cost.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI systems reliably perform the full suite of physical installation, arrangement, artifact handling, and on-site problem correction. Computer vision can document condition, but embodied execution in museum contexts remains research-stage and site-specific.
Technical feasibility todayclaude-sonnet-51/5No deployed product installs, arranges, or physically prepares museum artifacts; this remains entirely a human physical and curatorial task.

Clean objects, such as paper, textiles, wood, metal, glass, rock, pottery, and furniture, using cleansers, solvents, soap solutions, and polishes.

9

CI 5–14 · exposure 8 · augmentation 38 · importance 4.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Museum and conservation sectors are low-digitization, small-scale, and risk-averse environments with deep attachment to human expertise and hands-on practice. Adoption of AI-driven cleaning automation is minimal and unlikely in the foreseeable future.
Sector adoption velocityclaude-sonnet-51/5Museum conservation is a small, low-digitization, physically-oriented field with minimal AI/robotic adoption for hands-on object treatment.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist by recommending cleaning protocols based on material analysis, predicting solvent compatibility, and documenting before/after via imaging—useful productivity aids for a human conservator. However, augmentation remains secondary to human judgment and hands-on control.
Augmentation potentialclaude-sonnet-52/5AI can help identify materials, suggest cleaning agents, or document condition via imaging, but offers little assistance in the actual physical cleaning process.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can guide cleaning protocols and analyze materials via imaging, the task requires delicate physical manipulation, tactile feedback, and real-time judgment about solvent application and surface damage risk—capabilities current robots lack reliably. The task is >50% supervisory only; hands-on execution remains firmly human.
Task automatabilityclaude-sonnet-51/5This is a hands-on physical conservation task requiring fine motor manipulation of delicate, often irreplaceable objects; no AI system can perform physical cleaning.
Adoption barriersclaude-haiku-4-5-202510014/5Museums face high liability and professional standards barriers: conservators typically require formal training/certification, and institutional risk aversion around artifact damage creates strong organizational and legal friction. Regulatory standards and professional accreditation further restrict automation substitution.
Adoption barriersclaude-sonnet-54/5Irreversible damage risk to valuable/irreplaceable artifacts creates strong professional and institutional requirements for trained, credentialed conservators to handle materials.
Cost vs. human wageclaude-haiku-4-5-202510011/5Specialized robotic systems capable of this work would require custom engineering, integration, and insurance—far exceeding the loaded hourly wage of a museum technician. Current AI solutions offer minimal labor displacement, so cost advantage is negligible or negative.
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute performing this physical task, so AI cost is not comparable; human conservators remain the only option.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product exists that can autonomously clean fragile museum artifacts with the precision, material discrimination, and damage-avoidance required. Robotic cleaning systems exist for industrial settings but not for heterogeneous cultural artifacts where error costs are catastrophic.
Technical feasibility todayclaude-sonnet-51/5No deployed AI or robotic product performs conservation cleaning of museum artifacts; this remains entirely a specialist manual craft.

Build, repair, and install wooden steps, scaffolds, and walkways to gain access to or permit improved view of exhibited equipment.

7

CI 5–10 · exposure 0 · augmentation 25 · importance 3.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Museums are typically lower-digitization organizations with specialized, infrequent carpentry needs. There is minimal economic pressure or infrastructure for roboticizing small-batch physical construction tasks in museum contexts.
Sector adoption velocityclaude-sonnet-51/5Museum conservation and physical facilities work is a low-digitization, low-AI-adoption sector with no meaningful robotic automation deployed for this kind of task.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could assist with design visualization, structural planning, or material specification, but the hands-on carpentry and installation components offer limited scope for meaningful AI augmentation of the human worker.
Augmentation potentialclaude-sonnet-52/5AI could assist with design planning, material calculations, or generating blueprints for the structures, but offers minimal help with the actual physical building and installation.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires physical construction, carpentry, spatial reasoning, and adaptation to unique museum environments. Current AI systems cannot physically build, repair, or install structures in the real world.
Task automatabilityclaude-sonnet-51/5This is physical carpentry and construction work requiring manual dexterity, tool use, and on-site fabrication that current AI systems cannot perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Museums have stringent safety codes, preservation standards, and liability requirements for structural work. Installation of scaffolds and walkways typically requires signed-off engineering and certified technician approval, creating legal and regulatory barriers to automation.
Adoption barriersclaude-sonnet-53/5No licensing requirement specifically bars automation, but the physical nature of the work, safety concerns around structural stability, and museum-specific handling needs create practical friction.
Cost vs. human wageclaude-haiku-4-5-202510011/5Robotic systems capable of carpentry are extremely expensive, require extensive customization per installation, and still need significant human oversight and intervention, making them far more costly than skilled human technicians.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute for physical carpentry, so any AI-based approach would be more expensive or simply infeasible compared to a human carpenter.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI product can autonomously perform carpentry or structural installation work. The task demands embodied physical manipulation that goes well beyond current robot capabilities in unstructured museum settings.
Technical feasibility todayclaude-sonnet-51/5No deployed product builds or installs physical wooden structures autonomously; this remains firmly in the domain of skilled human tradespeople.

Perform on-site field work which may involve interviewing people, inspecting and identifying artifacts, note-taking, viewing sites and collections, and repainting exhibition spaces.

6

CI 5–7 · exposure 0 · augmentation 50 · importance 3.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Museums and conservation are traditionally low-digitization sectors with slower tech adoption; field conservation work is inherently local and hands-on, limiting automation even where museums use digital cataloging and documentation tools.
Sector adoption velocityclaude-sonnet-51/5Museum conservation is a small, specialized, low-digitization field with minimal AI deployment for physical fieldwork tasks.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist with documentation (automated photo tagging, note organization, artifact database search), but the core field inspection, artifact assessment, and hands-on conservation remain human-dependent; assistive tools offer modest productivity gains on peripheral tasks.
Augmentation potentialclaude-sonnet-53/5AI can assist with note transcription, identification research, image analysis of artifacts, or drafting interview summaries, though the core physical and interpersonal activities remain human-driven.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires physical field presence, inspection of physical artifacts with expert judgment, direct interpersonal interaction (interviewing), and on-site painting. Current AI cannot be physically deployed to inspect artifacts, conduct interviews, or perform repainting; these activities fundamentally depend on embodied human presence and tactile expertise.
Task automatabilityclaude-sonnet-51/5This task bundles physical fieldwork—interviewing, hands-on artifact inspection, site visits, and manual repainting—none of which current AI can perform end-to-end; these require physical presence and manipulation.
Adoption barriersclaude-haiku-4-5-202510014/5Museum conservation involves significant liability for artifact damage, professional licensing/accreditation requirements in many jurisdictions, and organizational trust in human expertise for irreversible decisions on valuable collections. Regulatory and institutional frameworks strongly protect human conservator sign-off.
Adoption barriersclaude-sonnet-54/5Artifact identification and condition assessment often require professional expertise and judgment, and physical site work demands human presence, creating strong practical barriers to automation.
Cost vs. human wageclaude-haiku-4-5-202510011/5Field work, especially artifact inspection and conservation/repainting, remains labor-intensive and requires human expertise; AI systems cannot substitute for the full cost of a technician's site time, physical labor, and professional judgment even with current tools.
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute for the physical and interpersonal components, so any AI cost would be additive to, not a replacement for, human labor costs.
Technical feasibility todayclaude-haiku-4-5-202510011/5While AI can assist with note-taking and documentation, no deployed product can independently perform artifact identification and inspection in the field, conduct interviews, or execute physical conservation work such as repainting. The task's core components require human domain expertise and physical agency in situ.
Technical feasibility todayclaude-sonnet-51/5No deployed product conducts on-site interviews, physically inspects artifacts, or repaints exhibition spaces; this remains firmly outside current AI product capability.

Supervise and work with volunteers.

5

CI 5–5 · exposure 0 · augmentation 25 · importance 3.6/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Volunteer management and supervision occur in nonprofits, cultural institutions, and small organizations with low automation adoption rates. These sectors remain manual and human-centric, with minimal pressure or capability to automate people management.
Sector adoption velocityclaude-sonnet-51/5Museums are a small, low-digitization sector with minimal AI adoption for people-management tasks.
Augmentation potentialclaude-haiku-4-5-202510012/5While AI might assist with scheduling, communications templates, or volunteer record-keeping, it offers minimal support for the core supervision task—observing performance, providing feedback, and managing interpersonal dynamics. Augmentation is narrow and marginal.
Augmentation potentialclaude-sonnet-52/5AI can help with scheduling tools, communication drafts, or training materials for volunteers, but offers limited direct assistance to the supervisory act itself.
Task automatabilityclaude-haiku-4-5-202510011/5Supervising and working with volunteers requires human judgment, empathy, conflict resolution, and real-time adaptive management—core interpersonal skills that current AI cannot perform end-to-end. AI cannot meaningfully supervise, mentor, or resolve volunteer issues today.
Task automatabilityclaude-sonnet-51/5Supervising and working with volunteers requires interpersonal leadership, motivation, scheduling, and relationship management that AI cannot perform end-to-end today.4o AI cannot substitute for the human presence required in direct supervision.4o
Adoption barriersclaude-haiku-4-5-202510014/5Museums and nonprofit organizations typically require a human supervisor or manager for volunteer coordination due to legal liability, volunteer satisfaction, institutional culture, and accountability norms. Organizational structure and human-contact requirements are substantial barriers.
Adoption barriersclaude-sonnet-54/5Supervision involves interpersonal trust, accountability, and organizational hierarchy that strongly resist automation, though not formally licensed like some professions.
Cost vs. human wageclaude-haiku-4-5-202510011/5AI has no cost advantage when the task is fundamentally about human relationship management; oversight and human involvement remain necessary, making the all-in cost higher than direct human supervision.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing this supervisory task, so cost comparison favors the human by default.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product can reliably supervise or manage volunteers in a production museum setting. Task assignment, conflict resolution, and performance feedback all require human presence and accountability that AI systems do not provide.
Technical feasibility todayclaude-sonnet-51/5No deployed product manages or supervises volunteers in a museum/conservation setting; this remains a fully human management function.

Preserve or direct preservation of objects, using plaster, resin, sealants, hardeners, and shellac.

5

CI 5–5 · exposure 0 · augmentation 25 · importance 3.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Museums are among the slowest-adopting sectors for automation due to their small scale, physical/hands-on nature of work, and institutional conservatism around artifact handling. Adoption of AI in this context remains negligible in production deployments.
Sector adoption velocityclaude-sonnet-51/5Museum conservation is a small, highly specialized, physically-oriented field with minimal AI adoption for hands-on treatment work.
Augmentation potentialclaude-haiku-4-5-202510012/5AI can assist conservators in documentation, analysis (material identification via imaging), and condition assessment, but its role in augmenting the core manual preservation task itself is limited. The human remains fully in the loop, and AI support is indirect rather than transformative to preservation productivity.
Augmentation potentialclaude-sonnet-52/5AI could assist with research on materials, historical treatment records, or documentation, but offers little help with the actual physical preservation technique itself.
Task automatabilityclaude-haiku-4-5-202510011/5Preservation of museum objects requires hands-on physical manipulation, material judgment, and context-specific decision-making that current AI cannot perform end-to-end. The task involves selecting appropriate materials based on object properties and applying preservation techniques with precision—work that demands embodied expertise and real-time sensory feedback.
Task automatabilityclaude-sonnet-51/5This is a hands-on, physical conservation task requiring manual dexterity with materials like plaster, resin, and shellac; no AI system can physically apply or manipulate these materials.
Adoption barriersclaude-haiku-4-5-202510014/5Museum preservation is heavily regulated by professional standards (ICOM-CC codes of ethics, institutional policies) and often requires licensed or credentialed conservators, particularly for high-value objects. Institutional liability, risk of object damage, and the requirement for human expertise to make preservation judgments create substantial legal and organizational barriers to automation.
Adoption barriersclaude-sonnet-54/5Museum conservation of valuable, often irreplaceable objects requires specialized training, professional judgment, and accountability for irreversible material choices, creating strong practical and institutional barriers to automation.
Cost vs. human wageclaude-haiku-4-5-202510011/5The specialized labor of museum conservators commands significant hourly rates, and any AI system capable of physical object manipulation would require expensive robotic infrastructure far exceeding the cost of human conservator labor for this specialized, low-volume work.
Cost vs. human wageclaude-sonnet-51/5AI cannot perform the physical application work at all, so there is no viable AI cost basis to compare against human labor for this hands-on task.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI system can autonomously perform physical preservation work using plaster, resin, sealants, hardeners, and shellac. While AI can assist in documentation and analysis, the core task of actual material application and preservation direction remains a human-performed activity with no production-stage AI alternative.
Technical feasibility todayclaude-sonnet-51/5No deployed product performs physical conservation treatment on museum objects; this remains entirely a skilled human craft process.

Deliver artwork on courier trips.

5

CI 5–5 · exposure 0 · augmentation 25 · importance 3.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Museums and cultural institutions operate in sectors with low digitization pressure for this task; delivery logistics remain human-dependent and adoption of automation for high-value artwork is minimal.
Sector adoption velocityclaude-sonnet-51/5Museum logistics and physical art handling are a low-digitization, physically-bound sector with essentially no AI displacement occurring in courier travel.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could assist with route optimization or scheduling of courier trips, but the core task of hand-delivering and physically securing artwork offers limited augmentation opportunity while the human remains essential.
Augmentation potentialclaude-sonnet-52/5AI can assist with trip logistics, condition report documentation, or customs paperwork, but offers minimal help with the core physical act of accompanying and safeguarding the artwork.
Task automatabilityclaude-haiku-4-5-202510011/5Delivering artwork requires physical handling, navigation in real-world environments, and human judgment about placement and safety—core capabilities where current AI systems have no meaningful autonomous capacity today.
Task automatabilityclaude-sonnet-51/5Physically transporting and personally escorting valuable artwork requires a human body present with the object at all times; no AI system can perform this physical courier function.'
Adoption barriersclaude-haiku-4-5-202510014/5Insurance liability for artwork damage, legal custody-of-property requirements, and institutional preference for human accountability create substantial barriers to automated or autonomous delivery substitution.
Adoption barriersclaude-sonnet-54/5Insurance, chain-of-custody, and institutional loan agreements typically require a qualified human courier to physically accompany and verify the artwork's condition, creating strong contractual and liability barriers.
Cost vs. human wageclaude-haiku-4-5-202510011/5A human courier with vehicle, insurance, and wage costs remains far cheaper than any conceivable autonomous system capable of handling valuable artwork safely and reliably.
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute for the physical act of courier travel, so cost comparison is moot; humans remain the only option and thus cheaper by default.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI product can independently deliver physical objects to museum locations; this task remains entirely dependent on human couriers with vehicles and decision-making capability.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product performs physical courier escort of artwork; this is entirely a physical logistics and human-presence task.

Direct and supervise curatorial, technical, and student staff in the handling, mounting, care, and storage of art objects.

3

CI 0–5 · exposure 0 · augmentation 25 · importance 3.8/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Museums are traditionally low-digitization, high-human-judgment sectors with strong professional norms around in-person curation and hands-on supervision. Adoption of AI for supervisory roles in conservation remains negligible.
Sector adoption velocityclaude-sonnet-51/5Museums and cultural institutions are a small, slow-to-digitize sector with little evidence of AI displacing supervisory conservation roles.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could assist with documentation, scheduling, or compliance tracking, but supervisory direction and staff management fundamentally require human presence, trust, and accountability. Augmentation potential is limited to administrative overhead rather than core supervision.
Augmentation potentialclaude-sonnet-52/5AI could help with scheduling, documentation, or training materials, but offers limited direct assistance to the core act of supervising staff and physical object handling.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires hands-on supervision, judgment calls about individual staff performance, and real-time decision-making in response to personnel needs and object conditions. AI cannot meaningfully substitute for direct oversight and interpersonal management of human staff.
Task automatabilityclaude-sonnet-51/5This is a supervisory/managerial task requiring in-person leadership, judgment about staff performance, and physical oversight of delicate object handling that AI cannot perform end-to-end.'
Adoption barriersclaude-haiku-4-5-202510015/5Museums have legal and fiduciary responsibility for collection care; a qualified human curator or conservator must legally sign off on staff training, object handling protocols, and institutional practices. Liability and regulatory requirements create hard barriers to substitution.
Adoption barriersclaude-sonnet-54/5Supervision of staff handling valuable/fragile art objects involves institutional liability, professional expertise, and accountability structures that require a qualified human in charge.
Cost vs. human wageclaude-haiku-4-5-202510011/5Supervision of specialized staff in a museum setting requires a human with domain expertise, authority, and liability. The cost of a qualified museum supervisor far exceeds any AI assistance available today, and no AI can assume the role end-to-end.
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute performing this supervisory role, so any AI cost would be additive rather than substitutive, making it more expensive than simply having a human supervisor.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI system can direct or supervise human staff in handling sensitive art objects or make real-time curatorial and care decisions. This requires embodied presence, accountability, and contextual human judgment that current AI cannot provide.
Technical feasibility todayclaude-sonnet-51/5No deployed product directs and supervises human staff in physical conservation work; this remains entirely a human management function.

Related occupations — Educational Instruction & Library

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.