Archivists
25-4011.00Median wage $64,550/yr7,970 employed (US)Rank #430 of 923 scored · top 47% by substitution
Appraise, edit, and direct safekeeping of permanent records and historically valuable documents. Participate in research activities based on archival materials.
Sub-scores
0–100 · band = confidence interval from rater disagreement
Substitution29
Exposure26
Augmentation65
Substitution — the headline: capability discounted by cost, barriers and adoption.
Exposure — technical capability alone, regardless of whether anyone deploys it.
Augmentation — how much AI assists without replacing. High here + moderate substitution = a changing job, not a disappearing one.
Tasks on the substitution scale
13 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%27
panel mean rating 2.1/5 → substitution pressure 27/100
Technical feasibility todayw 20%22
panel mean rating 1.9/5 → substitution pressure 22/100
Cost vs. human wagew 15%28
panel mean rating 2.1/5 → substitution pressure 28/100
Adoption barriersw 20%inverted — strong barriers lower the score41
panel mean rating 3.4/5 (barrier strength) → substitution pressure 41/100
Sector adoption velocityw 10%26
panel mean rating 2.0/5 → substitution pressure 26/100
Task breakdown (13 tasks)
Substitution pressure per task, weighted by O*NET importance in the composite. Expand a task for the full rater audit trail — every rating, every model, every rationale.
Prepare archival records, such as document descriptions, to allow easy access to information.
47CI 43–52 · exposure 45 · augmentation 75 · importance 4.7/5 · click for rater detail
Prepare archival records, such as document descriptions, to allow easy access to information.
47| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Archival organizations are primarily smaller institutions, universities, and cultural heritage bodies with slower digital transformation and tight budgets; pilots are growing but production adoption remains limited compared to information-rich sectors like finance or tech. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Archives and libraries are traditionally slow to adopt AI due to underfunding, legacy systems, and small-scale operations, though some large institutions are piloting AI tools. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI excels at drafting descriptions, flagging duplicates, suggesting metadata tags, and accelerating indexing, meaningfully raising archivist productivity while they remain responsible for accuracy and judgment on significance and access restrictions. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI significantly speeds up drafting descriptions, transcription, and metadata suggestions, letting archivists focus on verification and complex judgment calls. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI can generate document descriptions and metadata at scale via OCR and NLP, creating 50%+ time savings on routine cataloging. However, complex contextual judgment about archival significance, relationships, and authority control still typically requires human review, preventing full end-to-end automation. |
| Task automatability | claude-sonnet-5 | 3/5 | AI can draft finding aids, metadata, and descriptions from digitized text, but archivists must verify accuracy, provenance, and context, limiting full end-to-end automation. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Most archival work lacks hard legal barriers, but professional standards (DACS, EAD), institutional preferences for human expertise, and the high cost of bad metadata create material friction against full automation without expert sign-off. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement, but institutional standards (e.g., DACS, EAD) and provenance/authenticity concerns create moderate friction requiring expert oversight. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 3/5 | AI-assisted description generation is becoming cost-competitive with manual archivist time, especially for high-volume routine cataloging, but integration overhead and human review still keep all-in costs roughly equivalent to hiring archivists for complex collections. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI can cut drafting time for descriptions significantly, but human review, quality control, and handling of physical/rare materials keep costs closer to comparable rather than order-of-magnitude cheaper. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Products like DocumentAI and archival management systems with ML-assisted description exist and perform reliably on straightforward tasks, but error rates on nuanced description quality and edge cases mean most organizations require human oversight, limiting production reliability to partial deployment. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some archival institutions pilot AI-assisted metadata generation and OCR-based description tools, but reliable production-scale deployment for nuanced archival description is still limited and error-prone. |
Create and maintain accessible, retrievable computer archives and databases, incorporating current advances in electronic information storage technology.
36CI 25–46 · exposure 38 · augmentation 75 · importance 4.6/5 · click for rater detail
Create and maintain accessible, retrievable computer archives and databases, incorporating current advances in electronic information storage technology.
36| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Archives and heritage institutions are typically slower-adopting sectors with legacy systems, smaller budgets, and strong professional norms favoring human curatorial judgment. Pilot projects exist but production-scale displacement remains limited and concentrated in well-funded institutions. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Libraries, museums, and archival institutions are adopting AI-assisted metadata and digitization tools at a moderate pace, but many remain under-resourced and slow to modernize infrastructure fully. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI excels at assisting archivists with bulk metadata generation, duplicate detection, format conversion, and search-optimization while the archivist maintains interpretive control and final curation decisions. This substantially raises archival productivity without removing human oversight. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools significantly boost productivity in areas like automated metadata extraction, OCR, classification, and search indexing, meaningfully augmenting archivists' ability to build and maintain these systems. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI can automate significant portions—metadata generation, database schema design, information extraction from documents, and basic indexing—but human judgment remains essential for evaluating archival significance, determining access restrictions, and making preservation decisions. This covers roughly half the workflow with setup. |
| Task automatability | claude-sonnet-5 | 2/5 | AI can assist with database schema design, metadata generation, and some data entry, but designing, implementing, and maintaining an institution-specific archival system requires ongoing human judgment about preservation standards, security, and organizational needs that current AI cannot fully replace end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Archives operate under strict legal and regulatory frameworks (FERPA, HIPAA, institutional access policies) that require human accountability; many institutions legally mandate human archivists to certify preservation standards and make access decisions. Professional certification and institutional trust in human stewardship create high adoption friction. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing requirement to build a database, but organizational standards, compliance, and long-term preservation liability create meaningful friction against full automation without human oversight. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | While OCR and basic metadata extraction are cheap, archival work demands specialized expertise in preservation technology and institutional knowledge that AI cannot yet replace. The full end-to-end cost (AI + human review + integration) remains close to or exceeds skilled archivists' loaded wages. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | While cloud storage and some automated tools are cheap, the specialized design, migration planning, and long-term stewardship work still requires skilled human labor, keeping all-in AI substitution costs comparable to or above human cost for full task scope. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Deployed products exist for document digitization, OCR, and database management, but they work reliably only within narrow scopes and require substantial human oversight for archival-specific tasks like provenance verification and access control classification. Material error rates persist in automated categorization. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Products exist for automated metadata tagging, digital asset management, and cloud storage integration, but full end-to-end system creation and maintenance in production archival settings still relies heavily on human architects and archivists. |
Provide reference services and assistance for users needing archival materials.
30CI 25–35 · exposure 25 · augmentation 75 · importance 4.8/5 · click for rater detail
Provide reference services and assistance for users needing archival materials.
30| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Archival institutions are historically laggard in automation; most are under-resourced and operate in the public/cultural sector with limited investment in AI infrastructure. Pilots exist but production adoption of AI reference systems remains rare. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Archives and libraries are traditionally slower-adopting sectors relative to finance or tech, with AI reference tools still largely in pilot phases rather than widespread production use. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can powerfully assist archivists by rapidly searching large collections, suggesting related materials, or transcribing handwritten documents, thereby freeing curators to focus on patron interaction and complex research guidance. Current tools meaningfully boost productivity when the human remains in control. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can meaningfully speed up preliminary research, catalog searching, and drafting responses, significantly aiding archivists while they retain responsibility for verification and specialized judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can help search and retrieve basic archival metadata, the task requires understanding patron needs, contextual knowledge of collections, and often handling rare/fragile materials or complex research guidance—activities that resist full automation. Current systems cannot reliably replace the interpretive and advisory components that define reference service. |
| Task automatability | claude-sonnet-5 | 2/5 | AI can help search catalogs and answer basic questions, but interpreting user needs, locating obscure physical materials, and providing nuanced archival guidance requires human judgment that current systems cannot fully replicate end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Archival institutions often have institutional cultures and legal/professional standards requiring human intermediation for sensitive materials, access restrictions, and complex genealogical or scholarly inquiries. Libraries and archives have slow digital transformation and strong norms around professional expertise. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No strict licensing requirement exists, but institutional trust, accuracy expectations for scholarly/legal research, and physical access to unique materials create meaningful friction against full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI-powered search and retrieval systems have modest infrastructure costs, but the oversight burden (fact-checking results, handling exceptions, training curators) remains labor-intensive. The all-in cost is not yet substantially lower than paying an archivist, especially for high-accuracy reference work. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | While AI chat assistance is cheap per query, the oversight and specialized knowledge integration needed for accurate archival reference still requires costly human expert involvement, keeping overall cost comparable to human staff. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Some AI chatbots and search tools exist to help users find archival materials, but they typically have narrow scope, limited domain knowledge, and high error rates when handling specialized requests. Deployed systems rarely match the judgment and contextual depth a trained archivist brings to reference work. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some libraries deploy chatbots for basic reference triage, but reliable handling of complex archival reference queries involving specialized collections is not yet a mature deployed capability. |
Specialize in an area of history or technology, researching topics or items relevant to collections to determine what should be retained or acquired.
29CI 25–34 · exposure 25 · augmentation 63 · importance 3.6/5 · click for rater detail
Specialize in an area of history or technology, researching topics or items relevant to collections to determine what should be retained or acquired.
29| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Archival institutions tend to be smaller, slower-digitizing, and conservative; while some are piloting AI-assisted cataloging, autonomous acquisition decisions remain rare. Adoption remains limited to metadata assistance rather than core curatorial tasks. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Archival and library science institutions are generally slower adopters of AI compared to finance or tech sectors, with pilots for research assistance but little deployment for collection decision-making. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can usefully assist archivists by automating literature searches, suggesting relevant items from databases, and generating initial metadata, which raises productivity on research and assessment phases. The human archivist remains the decision-maker but benefits from faster information retrieval. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can significantly speed up literature review, historical research, and identification of relevant items, greatly aiding an archivist's research phase even though final judgment remains human. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Only limited parts of this task are automatable—AI can assist with keyword extraction, metadata tagging, and initial relevance assessment of items. However, the core judgment about historical or technological significance and acquisition decisions requires domain expertise and curatorial judgment that AI cannot reliably replicate end-to-end. |
| Task automatability | claude-sonnet-5 | 2/5 | This requires deep domain expertise, contextual judgment about institutional collecting priorities, and evaluation of provenance/significance that current AI cannot reliably perform end-to-end; AI can support research but not make acquisition determinations autonomously.</br> |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Archival acquisition decisions carry liability and reputational risk; institutions typically require human archivists with subject expertise to make and sign off on retention and acquisition decisions. Professional credentials and institutional accountability create strong friction against full automation. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No formal licensing requirement, but institutional trust, curatorial authority, and long-term stewardship responsibilities create organizational friction against fully automating acquisition judgment. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 3/5 | AI-assisted research tools reduce time on literature review and metadata tasks, but the cost of AI infrastructure plus human oversight and final decision-making remains roughly comparable to direct archivist labor for the same acquisition quality. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | While AI research assistance is cheap per query, the human oversight, verification, and specialized judgment needed to validate acquisition decisions keeps costs comparable to or only modestly below human-only workflows. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While AI tools exist for text classification and metadata generation, no mature product reliably performs the full task of specialized historical research and acquisition recommendation independently. Archivists still need to validate and override automated suggestions significantly. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | No deployed product independently researches and determines archival acquisition/retention decisions; existing tools are limited to search/summarization support for human archivists. |
Preserve records, documents, and objects, copying records to film, videotape, audiotape, disk, or computer formats as necessary.
29CI 25–32 · exposure 25 · augmentation 63 · importance 4.3/5 · click for rater detail
Preserve records, documents, and objects, copying records to film, videotape, audiotape, disk, or computer formats as necessary.
29| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Large cultural institutions and libraries have adopted digitization equipment and workflows, but adoption is uneven and often project-based rather than fully replacing archivist roles. Many smaller archives lack resources for automation; the sector remains labor-intensive and slow to fully digitize backlogs. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Archives and cultural heritage institutions are typically under-resourced and slow adopters of new technology, with digitization projects often manual, grant-funded, and not automated at scale. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI-assisted OCR, metadata extraction, format conversion, and quality-control flagging significantly boost archivist productivity by automating routine digitization steps and indexing, allowing archivists to focus on preservation strategy, complex materials, and validation. This is a strong case of human-in-the-loop augmentation. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI tools assist with metadata tagging, OCR, and organizing digitized content, improving efficiency, but the physical preservation and format transfer work still requires substantial human execution. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While digitization of physical records into common formats (scanning to PDF, converting audio to WAV) can be partially automated with specialized equipment and software, the task also requires judgment about preservation method selection, handling fragile materials, metadata creation, and quality control that current AI cannot reliably handle end-to-end. Automation covers the mechanical copying but not the preservation decision-making. |
| Task automatability | claude-sonnet-5 | 2/5 | Digitization workflows can be partially automated (scanning, OCR, format conversion), but preservation decisions, handling of fragile physical media, and quality control require human judgment and physical manipulation that current AI cannot replace end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Archival preservation is governed by professional standards (Library of Congress, ISO guidelines) and institutional policies; many organizations require certified archivists to validate preservation decisions and sign off on digital authenticity. Legal liability for loss or degradation of irreplaceable cultural records creates a strong incentive to retain human expert oversight. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing requirement, but institutional standards, chain-of-custody requirements, and risk of damaging irreplaceable objects create meaningful procedural friction favoring trained human oversight. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Automated digitization equipment and software are available but require significant upfront investment, trained operators, and quality-control staff. The labor cost of skilled archivists remains competitive with or cheaper than the all-in cost of high-reliability digitization systems plus oversight. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Physical media conversion still requires specialized hardware, handling, and human oversight, so while software tools reduce some costs, the overall process remains labor- and equipment-intensive relative to a human archivist doing it directly. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Commercial digitization services and OCR/document-conversion software exist, but they require human oversight for quality assurance, format selection, and metadata accuracy. No mature product reliably handles the full preservation workflow—material assessment, format choice, copying, and validation—without substantial human intervention. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Digitization equipment and software exist and are used in production, but AI-driven autonomous preservation of diverse physical formats (film, audiotape, objects) is not a mature deployed product handling the full task. |
Organize archival records and develop classification systems to facilitate access to archival materials.
28CI 25–30 · exposure 25 · augmentation 75 · importance 4.8/5 · click for rater detail
Organize archival records and develop classification systems to facilitate access to archival materials.
28| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Archives operate in lower-digitization, traditional sectors with limited tech infrastructure and funding; pilot projects exist but production deployment of AI-driven classification remains rare and largely confined to larger research institutions. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Archives and libraries are historically slow adopters of AI relative to finance or tech sectors, with most AI use still in pilot or experimental stages for classification tasks. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI significantly enhances archivist productivity by automating routine tagging, suggesting subject headings, and bulk metadata extraction, allowing human curators to focus on complex classification decisions and policy development rather than data entry. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can meaningfully assist archivists by suggesting metadata tags, extracting entities from documents, and drafting finding aids, significantly speeding up parts of the classification workflow while the archivist retains final judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can assist with metadata extraction and basic document sorting, the task requires substantial human judgment to develop meaningful classification systems that reflect institutional context, historical nuance, and future research needs. Significant manual review and curatorial decision-making remain essential. |
| Task automatability | claude-sonnet-5 | 2/5 | Developing classification systems requires domain judgment about provenance, historical context, and institutional research needs that current AI cannot reliably determine autonomously; AI can assist with metadata tagging but not the full organizational design task.“}, |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Strong barriers exist: many archives are subject to legal restrictions on access and handling, professional standards (e.g., ISAD(G), DACS) require human expertise to implement properly, and institutional accountability for misclassification creates liability concerns that demand licensed or accredited archivists sign off on systems. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing barrier exists, but archival standards (e.g., ISAD(G), institutional policies) and the need for professional judgment on provenance and access create moderate organizational friction against full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Integration costs, model fine-tuning for domain-specific vocabularies, and mandatory human oversight (particularly for rare or sensitive materials) make the all-in cost comparable to or higher than trained archivists performing the work directly. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Given the need for extensive human oversight, correction, and domain expertise to validate classification decisions, AI tools reduce some labor but do not yet approach order-of-magnitude cost savings for this task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | AI tools exist for document OCR and automated tagging, but no deployed product reliably performs end-to-end archival classification without material human intervention. Systems struggle with ambiguous provenance, mixed-media materials, and the need for contextual understanding that exceeds current capabilities. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some digital archive tools use AI-assisted metadata extraction and tagging, but no deployed product independently organizes archival records or designs classification schemas at production scale in real archives. |
Research and record the origins and historical significance of archival materials.
28CI 25–30 · exposure 25 · augmentation 63 · importance 4.2/5 · click for rater detail
Research and record the origins and historical significance of archival materials.
28| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Archive institutions tend to be risk-averse, under-resourced, and geographically dispersed; adoption of AI in archival work remains limited to digitization assistance and basic indexing, not core research tasks. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Archives and libraries are historically slow adopters of AI due to funding constraints, specialized workflows, and preservation-focused institutional culture. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can meaningfully assist archivists by automating metadata extraction, suggesting date ranges, and flagging related materials, which speeds up the research phase; however, the human must validate findings and make final significance judgments. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can meaningfully speed up literature search, transcription, translation, and preliminary contextual research, significantly aiding archivists while they retain final judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | AI can assist with basic metadata extraction and date/source identification from documents, but determining historical significance requires deep contextual understanding, expert judgment, and nuanced interpretation of primary sources that current systems cannot reliably perform end-to-end. |
| Task automatability | claude-sonnet-5 | 2/5 | AI can assist with drafting summaries and searching digitized records, but authoritative provenance research requires cross-referencing physical materials, contextual judgment, and verification that current systems cannot reliably perform end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Archives are typically governed by institutional standards, professional archival ethics, and accreditation requirements; institutions are cautious about delegating provenance and significance determination to machines, and human expertise is valued and professionally certified. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No strict licensing requirement, but institutional standards, scholarly credibility, and risk of misattribution create meaningful friction against full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Current AI assistance (OCR, metadata tagging) reduces some clerical overhead, but the core research and judgment work requires expert archivist time; integration costs and oversight needs mean AI cost-per-task remains high relative to human expertise. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI tools can reduce time on preliminary searches but still require substantial expert human verification and correction, so all-in cost savings versus a trained archivist are modest. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While OCR and basic document classification tools exist, no deployed system reliably performs the full research task of establishing origins and assessing historical significance; most archive work still relies on human experts consulting multiple sources and applying domain knowledge. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some archival institutions use AI-assisted metadata generation and OCR-based research tools, but no mature product independently performs full provenance and historical significance research reliably. |
Select and edit documents for publication and display, applying knowledge of subject, literary expression, and presentation techniques.
28CI 25–30 · exposure 25 · augmentation 63 · importance 3.1/5 · click for rater detail
Select and edit documents for publication and display, applying knowledge of subject, literary expression, and presentation techniques.
28| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Cultural institutions and archives adopt AI slowly; most remain conservative on core curatorial tasks, preferring human experts, and digitization/publication budgets are often constrained, limiting experimental AI deployment. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Archives and cultural heritage institutions are typically under-resourced and slow to adopt AI tools compared to fast-moving sectors like finance or tech. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can usefully assist archivists by suggesting candidate documents, proposing editorial corrections, and flagging formatting inconsistencies, thereby speeding parts of the workflow while the human retains curatorial control and final judgment. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can meaningfully assist with drafting captions, summarizing content, suggesting edits, and checking style/consistency, significantly speeding up parts of the workflow while archivists retain final judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can assist with document selection via metadata analysis and basic editing suggestions, the task requires nuanced judgment about subject expertise, literary merit, and appropriate presentation—areas where AI alone cannot reliably meet the 50% time-saving threshold without substantial human oversight and revision. |
| Task automatability | claude-sonnet-5 | 2/5 | AI can draft summaries or suggest edits, but selecting archival documents for publication requires deep subject expertise, contextual judgment, and curatorial taste that current systems cannot reliably replicate end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Archives operate under institutional, legal, and professional standards (provenance, authenticity, ethical curatorial practice) that typically require a credentialed archivist to make and sign off on publication decisions; organizational tradition and professional liability also protect the human role. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No formal licensing requirement, but institutional standards, provenance verification, and reputational risk around public displays create meaningful friction against full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI text processing and suggestion tools are inexpensive, but the specialized knowledge, judgment, and quality assurance required for archival work means total cost (including expert oversight) remains comparable to or higher than direct human curation. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI editing assistance is cheap, but the human curatorial judgment and verification of authenticity/context still dominate cost, keeping overall savings modest. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Some AI editing and organization tools exist, but no deployed product reliably performs end-to-end curation and literary-quality editing for archival publication without significant expert human intervention and domain-specific knowledge application. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some AI writing/editing tools assist with text polishing, but no deployed product performs the full curation-and-editing workflow for archival publication reliably in production. |
Coordinate educational and public outreach programs, such as tours, workshops, lectures, and classes.
26CI 16–35 · exposure 13 · augmentation 63 · importance 3.9/5 · click for rater detail
Coordinate educational and public outreach programs, such as tours, workshops, lectures, and classes.
26| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Archives and cultural institutions have historically been slower to digitize operations and adopt AI-driven coordination tools. Pilot adoption of scheduling and communication aids exists, but production-scale autonomous program coordination in archives remains rare. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Archives and cultural institutions are typically under-resourced and slow to adopt AI tools compared to fast-moving sectors like finance or tech, though some libraries use AI for content generation and marketing. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can usefully assist with drafting promotional materials, managing participant databases, suggesting scheduling options, and automating email reminders. These augmentations modestly raise a human coordinator's productivity without removing them from decision-making or relationship management. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can meaningfully assist with drafting workshop materials, marketing copy, scheduling logistics, and generating lecture outlines, significantly aiding archivists while they retain control of program design and delivery. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can help draft outreach materials, schedules, and promotional content, coordinating live educational programs requires real-time logistical management, interpersonal communication, and responsiveness to participant needs that resist full automation. Substantial human oversight remains essential for scheduling, venue coordination, and on-site facilitation. |
| Task automatability | claude-sonnet-5 | 1/5 | Coordinating in-person events, scheduling presenters, engaging with community members and adapting programs to audiences requires human judgment, relationship management, and physical presence that AI cannot replicate end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Organizations often prefer human coordinators for community trust, relationship-building, and liability in educational settings. There is no legal mandate requiring a human, but reputational and customer-preference friction is moderate, particularly for workshops and lectures serving diverse public audiences. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement bars AI involvement, but the task's reliance on in-person engagement, institutional relationships, and public-facing judgment creates practical friction against full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Current AI tools (chatbots, scheduling assistants) cost roughly as much to implement and maintain as they save in staff time on routine administrative tasks. The human coordinator still drives critical decisions, so the labor burden remains substantial relative to AI deployment cost. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Some administrative sub-tasks (emails, scheduling, drafting outreach copy) can be done cheaply by AI, but the core coordination and interpersonal work still requires paid staff time, keeping overall cost savings limited. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed product reliably handles end-to-end coordination of educational programs including scheduling, participant communication, and logistics. AI can assist with email templates and calendar management, but current systems lack the contextual reasoning and error recovery needed for reliable production-scale program coordination. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product manages the full coordination of tours, workshops, lectures and classes; at best AI tools assist with scheduling or drafting materials, not the coordination task itself. |
Direct activities of workers who assist in arranging, cataloguing, exhibiting, and maintaining collections of valuable materials.
25CI 20–30 · exposure 20 · augmentation 63 · importance 4.5/5 · click for rater detail
Direct activities of workers who assist in arranging, cataloguing, exhibiting, and maintaining collections of valuable materials.
25| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Archival institutions are generally slower to digitize and adopt automation due to small budgets, specialized workforces, and conservative stewardship practices. While some are piloting AI cataloguing aids, production-level workforce displacement is minimal and concentrated in a few well-resourced institutions. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Archival and library institutions are generally slow adopters of AI for managerial functions, though they may use AI tools for cataloguing support. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can meaningfully augment archivist productivity through automated metadata extraction, cataloguing suggestions, and collection analysis. These tools enhance researcher discovery and reduce manual data entry, allowing archivists to focus on curatorial and preservation decisions while staying in control. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help managers by generating task schedules, tracking collection status, and drafting instructions, improving efficiency while the human retains direction and oversight. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI could assist with cataloguing and metadata generation, the task requires significant human judgment in arrangement decisions, preservation assessment, and curatorial choices. End-to-end automation would struggle with the domain expertise needed for valuable collections and the requirement to direct other workers. |
| Task automatability | claude-sonnet-5 | 2/5 | Directing and supervising other workers requires interpersonal judgment, delegation, and adaptive management that current AI cannot perform end-to-end; only sub-components like scheduling or checklist generation could be automated. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Cultural institutions, museums, and archives operate under stewardship and legal duties regarding collections; many require credentialed professionals (archivists) to be responsible for preservation and access decisions. Liability and custodial responsibility create strong legal barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing requirement specifically bars AI from directing staff, but organizational structures and accountability for personnel management create real friction against automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI tools for cataloguing assistance exist but represent a small fraction of the archivist's total labor cost. The oversight, quality assurance, and specialized judgment required mean the effective cost ratio remains unfavorable for full substitution. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI could cheaply assist with task lists or workflow tracking, but the human supervisory role and accountability still require a paid manager, so cost savings are limited to auxiliary functions. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed product reliably performs the full supervisory and curatorial direction this task entails. AI can support cataloguing workflows, but managing collections of valuable materials and directing workers requires contextual expertise and accountability that current systems cannot reliably deliver in production archival settings. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product manages or directs human archival staff; this is a supervisory/management function outside current AI product scope. |
Locate new materials and direct their acquisition and display.
24CI 18–30 · exposure 20 · augmentation 50 · importance 3.9/5 · click for rater detail
Locate new materials and direct their acquisition and display.
24| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Archives and museums are typically lower-digitization, small-team organizations with limited tech adoption. Pilot AI projects exist but are rare; production deployment of AI-driven acquisition and display remains negligible in cultural heritage sectors. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Archives and museums are generally slow adopters of AI tools relative to fast-moving sectors like finance or tech, with pilots for cataloging but little production-level automation of acquisition decisions. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist archivists by surfacing candidate materials via search, flagging duplicates, suggesting metadata, and organizing digital assets for review, meaningfully raising archivists' ability to survey and organize. However, the human curatorial and acquisition decisions remain central to the task. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help by searching databases, tracking metadata, identifying potential acquisition targets, and suggesting display layouts, meaningfully aiding but not replacing archivist judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Locating materials can leverage search algorithms and AI-driven cataloging, but the judgment of what is 'new,' valuable, or acquisition-worthy requires curatorial expertise, provenance verification, and strategic collection decisions that remain deeply human tasks. AI cannot reliably evaluate acquisition priority or appropriateness without substantial human oversight. |
| Task automatability | claude-sonnet-5 | 2/5 | Identifying and physically/digitally acquiring new archival materials and directing their display involves negotiation, judgment about historical value, and physical logistics that AI cannot perform end-to-end today.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Archivists are credentialed professionals, and acquisition decisions often require institutional authority, legal responsibility for provenance authentication, and donor relationships that carry liability. Display and curation decisions rest on professional judgment with institutional accountability, creating regulatory and organizational barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No strict licensing requirement, but institutional trust, donor relationships, provenance verification, and curatorial authority create significant organizational friction against full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Current AI tools for search and cataloging are relatively inexpensive, but the domain expertise required (archivists with subject and provenance knowledge) commands higher wages. The full task including acquisition negotiation and curation strategy costs more to perform via human archivists than any near-term AI augmentation can replace. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI could cheaply assist with searching databases or cataloging, but the core acquisition negotiation and display curation still requires paid human expertise, keeping overall cost comparable to human labor. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Search and database tools exist, but no deployed product reliably automates the curatorial acquisition workflow end-to-end. Products for digital asset management and cataloging support portions of this work, but evaluating, negotiating acquisition, and directing display remain largely manual and require human judgment. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product autonomously locates, negotiates acquisition of, and directs display of archival materials; this remains a human curatorial and relationship-driven function. |
Establish and administer policy guidelines concerning public access and use of materials.
23CI 20–25 · exposure 20 · augmentation 50 · importance 4.6/5 · click for rater detail
Establish and administer policy guidelines concerning public access and use of materials.
23| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Archives and cultural institutions are typically slow to adopt automation, particularly for governance and policy functions. Adoption of AI assistance in archival work remains limited and concentrated in larger institutions, with most policy work still conducted through traditional human-led processes. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Archives and cultural heritage institutions are generally slow adopters of AI for governance-level decisions, with pilots more focused on cataloging/digitization than policy-setting. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can meaningfully assist archivists by analyzing usage patterns, suggesting policy language, identifying regulatory gaps, and drafting framework documents. These tools help accelerate policy development while the archivist retains final authority over institutional guidelines and compliance decisions. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help draft policy language, summarize precedents, or benchmark peer institution policies, providing moderate assistance while humans retain decision authority. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Establishing and administering public access policy requires nuanced judgment about organizational goals, legal constraints, and stakeholder needs. While AI could draft policy text or analyze existing frameworks, the core task of setting institutional direction and ensuring appropriate governance cannot be meaningfully automated to meet the 50% time-savings threshold. |
| Task automatability | claude-sonnet-5 | 2/5 | This requires organizational judgment, legal/ethical reasoning, and stakeholder negotiation to set policy, which current AI cannot autonomously establish or administer end-to-end.assistance is limited to drafting support. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Establishing access policy for archival materials often involves legal compliance, institutional authority, and fiduciary responsibility. Many archives operate under regulatory frameworks or institutional mandates that require a qualified human archivist to formally establish and sign off on access policies. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Archival policy often ties to institutional governance, legal compliance (e.g., privacy, FOIA-like access laws), and accountability structures that require human authorization and oversight. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI tools for policy drafting and analysis are relatively inexpensive, but the task still requires significant human expertise and oversight to validate output. The labor cost of a qualified archivist overseeing the process remains comparable to or lower than a full automation setup when accounting for risk and quality assurance. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Because the task requires human accountability and institutional decision-making, AI cannot substitute the core labor, so cost comparison favors the human role despite AI's cheap drafting assistance. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed products reliably perform policy establishment and administration end-to-end. AI can assist with policy analysis and drafting, but the authoritative decision-making and oversight required for archival access governance remains a human responsibility in current systems. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product independently establishes or administers institutional access policy; this remains a human governance function with AI at most drafting reference documents. |
Authenticate and appraise historical documents and archival materials.
23CI 20–25 · exposure 20 · augmentation 63 · importance 3.9/5 · click for rater detail
Authenticate and appraise historical documents and archival materials.
23| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Archives and cultural heritage institutions operate on constrained budgets and traditionally low digitization rates. Adoption of AI in archival work remains pilot-stage; most institutions are still digitizing and cataloging manually rather than deploying autonomous systems for authentication or appraisal. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Archival and library science is a small, specialized, historically low-digitization sector where AI pilots exist (e.g., handwriting recognition, metadata tagging) but full production adoption for appraisal/authentication is rare. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can meaningfully assist archivists by automating document digitization, suggesting metadata, flagging potential anomalies in images, and cross-referencing databases. These tools raise productivity on preliminary and analytical work, though human expertise remains central to final authentication and appraisal decisions. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools significantly aid archivists via OCR, handwriting transcription, image forensics, and database cross-referencing, meaningfully speeding up research that feeds into human-led authentication and appraisal decisions. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Authentication and appraisal require domain expertise, contextual judgment, and often physical inspection to detect forgeries or assess condition. While AI can assist with OCR, metadata extraction, and pattern matching on document images, fully autonomous authentication of historical materials—particularly handwriting verification, paper/ink analysis, and provenance assessment—remains beyond reliable automation and would require human validation, preventing the 50% time-saving threshold at equal quality. |
| Task automatability | claude-sonnet-5 | 2/5 | AI can assist with provenance research, dating, and cross-referencing but final authentication and appraisal require expert judgment involving physical inspection, historical context, and legal/financial stakes that current systems cannot reliably replicate end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Legal and professional standards govern archival authentication; institutions typically require credentialed archivists to certify appraisals for legal, tax, and historical-record purposes. Liability for authentication errors (e.g., false attribution) and regulatory compliance in heritage sectors create strong adoption friction against full automation. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Authentication often carries legal, financial, and provenance implications (e.g., forgery detection, chain of custody) where professional expertise and institutional accountability are effectively required, creating strong barriers to full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI tools for document processing (OCR, scanning automation) are inexpensive, but they handle only preliminary tasks. The core authentication and appraisal work still requires trained archivists or conservators whose labor cost far exceeds the marginal cost of AI assistance, making the total solution costlier than human-only approaches for the full task. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | While AI-assisted text analysis and image comparison are cheap, the human expertise, physical handling, and liability involved in authentication keep overall costs dominated by skilled labor, making AI cost savings marginal for the full task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Deployed products exist for limited tasks (document digitization, OCR, some metadata tagging), but no production systems reliably authenticate or appraise historical documents end-to-end. Forensic analysis, handwriting comparison, and expert judgment on historical significance remain primarily manual. Current AI tools lack the domain knowledge and error tolerance required for authentication in archival contexts. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs autonomous authentication or appraisal of archival materials in production; this remains a specialist human function supported at most by research-stage tools for metadata extraction or forgery detection. |
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.