Librarians and Media Collections Specialists

25-4022.00
Median wage $68,270/yr133,790 employed (US)Rank #199 of 923 scored · top 22% by substitution

Administer and maintain libraries or collections of information, for public or private access through reference or borrowing. Work in a variety of settings, such as educational institutions, museums, and corporations, and with various types of informational materials, such as books, periodicals, recordings, films, and databases. Tasks may include acquiring, cataloging, and circulating library materials, and user services such as locating and organizing information, providing instruction on how to access information, and setting up and operating a library's media equipment.

Sub-scores

0–100 · band = confidence interval from rater disagreement

Substitution39
Exposure35
Augmentation63

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

30 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

13%

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%36

panel mean rating 2.4/5 → substitution pressure 36/100

Technical feasibility todayw 20%33

panel mean rating 2.3/5 → substitution pressure 33/100

Cost vs. human wagew 15%38

panel mean rating 2.5/5 → substitution pressure 38/100

Adoption barriersw 20%inverted — strong barriers lower the score54

panel mean rating 2.9/5 (barrier strength) → substitution pressure 54/100

Sector adoption velocityw 10%30

panel mean rating 2.2/5 → substitution pressure 30/100

Task breakdown (30 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.

Check books in and out of the library.

92

CI 84–100 · exposure 92 · augmentation 50 · importance 4.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510015/5Automated and self-service circulation systems (including RFID kiosks and mobile check-out) are already widely deployed across public, academic, and special libraries; adoption is mature and ongoing rather than nascent.
Sector adoption velocityclaude-sonnet-54/5Self-service circulation technology has been widely adopted across public and academic libraries for over a decade, though smaller/rural libraries lag.
Augmentation potentialclaude-haiku-4-5-202510014/5Even in libraries with manual check-in/check-out, systems assist staff by automating holds queues, overdue notices, and collection analytics, significantly raising productivity of remaining human circulation workers.
Augmentation potentialclaude-sonnet-52/5The task is largely fully automated rather than augmented; any remaining human involvement is minimal oversight rather than productivity enhancement.
Task automatabilityclaude-haiku-4-5-202510014/5Checking books in and out is a straightforward, rule-based task involving barcode scanning, database lookup, and status updates. Current RFID and AI-powered library systems can perform this end-to-end with minimal human intervention, easily meeting the ≥50% time-saving threshold; only edge cases (damaged items, holds management) may require human judgment.
Task automatabilityclaude-sonnet-55/5Circulation check-in/check-out is a routine transactional task fully handled by barcode/RFID self-checkout kiosks and integrated library systems, requiring no human judgment.
Adoption barriersclaude-haiku-4-5-202510012/5While some libraries have organizational or staff-preference friction, there are no legal licensing requirements, regulatory mandates, or liability asymmetries that prevent full automation of book circulation; most barriers are organizational inertia rather than hard constraints.
Adoption barriersclaude-sonnet-51/5No licensing, legal, or professional requirement mandates a human librarian perform check-in/check-out; it's a purely administrative task.
Cost vs. human wageclaude-haiku-4-5-202510015/5The amortized cost of RFID infrastructure and automated circulation systems is dramatically lower than the loaded wage of a full-time library circulation clerk, achieving at least order-of-magnitude cost advantage once infrastructure is in place.
Cost vs. human wageclaude-sonnet-55/5Once installed, automated kiosks process transactions at negligible marginal cost compared to staff wages for repetitive circulation duties.
Technical feasibility todayclaude-haiku-4-5-202510015/5Deployed library management systems (ILS platforms like Alma, Evergreen, and RFID-based circulation systems) reliably perform check-in/check-out at scale in thousands of libraries worldwide, with mature APIs and integration proven in production.
Technical feasibility todayclaude-sonnet-55/5Self-checkout kiosks and automated sorting/return systems (e.g., 3M, Bibliotheca) are deployed at scale in public and academic libraries worldwide today.

Code, classify, and catalog books, publications, films, audio-visual aids, and other library materials, based on subject matter or standard library classification systems.

76

CI 67–85 · exposure 78 · augmentation 88 · importance 3.9/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510013/5Library systems are moderately digitized but often lag cutting-edge tech adoption; many libraries still rely on human catalogers and legacy systems, though forward-looking institutions and commercial vendors are actively deploying AI-assisted cataloging tools.
Sector adoption velocityclaude-sonnet-53/5Libraries have adopted shared cataloging databases and copy cataloging for decades, and are now piloting AI-assisted classification, but full-scale AI-driven original cataloging remains uneven across institutions with varying budgets.
Augmentation potentialclaude-haiku-4-5-202510015/5AI cataloging assistants (auto-generating draft metadata, subject suggestions, and hierarchical placements) substantially boost cataloger productivity by automating routine classification while preserving human judgment on complex or ambiguous materials, exemplifying high-value augmentation.
Augmentation potentialclaude-sonnet-54/5AI can suggest subject headings, classification numbers, and metadata drafts, dramatically speeding up librarians' cataloging workflow while they retain final review and correction responsibilities.
Task automatabilityclaude-haiku-4-5-202510015/5Classification and cataloging can be largely automated with subject-matter extraction (OCR/NLP), metadata assignment via machine learning models trained on existing library taxonomies, and systematic mapping to standard systems like Dewey Decimal or Library of Congress. Current LLMs and specialized NLP tools can extract subjects and assign codes with high consistency, meeting the ≥50% time-saving bar.
Task automatabilityclaude-sonnet-54/5Classification and cataloging (assigning subject headings, LC/Dewey codes, MARC records) from metadata or text is largely pattern-based work that AI can perform with high accuracy for standard materials, though edge cases and rare items still need human review.
Adoption barriersclaude-haiku-4-5-202510012/5While libraries may have internal workflows and preferences for human review, there are no legal or licensing barriers preventing AI automation of classification itself; organizational friction and quality-control preferences provide modest resistance but not hard constraints.
Adoption barriersclaude-sonnet-52/5No licensing requirement mandates a human cataloger; the main friction is quality control and institutional cataloging standards/policy rather than legal or professional barriers.
Cost vs. human wageclaude-haiku-4-5-202510015/5Automated classification via machine learning or cloud-based cataloging APIs costs a fraction of human librarian labor (≤$1–2 per item vs. $15–30+ in loaded wages), representing an order-of-magnitude cost advantage.
Cost vs. human wageclaude-sonnet-54/5Automated metadata extraction and classification is vastly cheaper per item than manual original cataloging, especially for high-volume standard materials, though oversight costs remain.
Technical feasibility todayclaude-haiku-4-5-202510014/5Mature products (library automation systems, vendor cataloging APIs, and emerging AI-powered metadata tools) perform classification and cataloging at scale in real library operations, though integration with legacy systems and occasional need for human review of edge cases keeps this from a perfect 5.
Technical feasibility todayclaude-sonnet-53/5Products like OCLC's tools, AI-assisted cataloging plugins, and vendor-supplied copy cataloging exist and are used, but fully automated original cataloging for unusual or non-English materials still has meaningful error rates requiring librarian verification.

Compile lists of books, periodicals, articles, and audio-visual materials on particular subjects.

73

CI 67–79 · exposure 70 · augmentation 88 · importance 3.6/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510014/5Libraries and information services sectors are actively digitizing and deploying AI-powered discovery tools, search engines, and recommendation systems; adoption is already visible in academic, public, and corporate library systems.
Sector adoption velocityclaude-sonnet-53/5Libraries and information services are moderately digitized with growing AI search adoption, but many institutions still rely on traditional cataloging workflows and pilot rather than production-scale AI use.
Augmentation potentialclaude-haiku-4-5-202510014/5AI dramatically accelerates the discovery and filtering process, allowing librarians to search vastly more materials in less time and focus on subjective curation, context, and patron guidance rather than manual compilation.
Augmentation potentialclaude-sonnet-55/5AI substantially speeds up discovery and compilation of relevant materials, letting librarians focus on curation, verification, and specialized subject expertise.
Task automatabilityclaude-haiku-4-5-202510014/5Current AI systems can effectively search databases, filter by subject metadata, and generate curated lists of books and materials with minimal human intervention, easily meeting a 50% time-saving threshold. The task requires no subjective judgment beyond applying pre-defined subject parameters, which AI excels at executing end-to-end.
Task automatabilityclaude-sonnet-54/5Compiling topical lists from catalogs, databases, and metadata is a well-structured retrieval task that LLM-powered search and citation tools can perform quickly, though final curation and relevance judgment still benefit from human review.
Adoption barriersclaude-haiku-4-5-202510012/5Librarians are not licensed professionals with mandatory sign-off requirements, and automation of list-generation does not trigger regulatory barriers; however, some organizational friction exists around job displacement and patron preference for human expertise in recommendations.
Adoption barriersclaude-sonnet-52/5No licensing requirement restricts this task, though institutional trust in accuracy and completeness of AI-generated bibliographies creates some friction before full reliance.
Cost vs. human wageclaude-haiku-4-5-202510015/5AI database querying and list compilation costs (API calls, integration overhead) are orders of magnitude cheaper than paying a professional librarian to manually search, retrieve, and organize materials across multiple sources.
Cost vs. human wageclaude-sonnet-54/5Automated bibliographic compilation via APIs and AI search tools costs a fraction of a librarian's hourly wage for equivalent list generation, even with some oversight required.
Technical feasibility todayclaude-haiku-4-5-202510014/5Multiple library management systems and content discovery platforms already deploy AI-assisted cataloging and recommendation engines in production; library systems routinely use automated search and list-generation features. While perfect accuracy in subject classification varies, the capability is demonstrably operational at scale in real institutions.
Technical feasibility todayclaude-sonnet-53/5Products like AI research assistants, citation managers, and library discovery tools with AI search exist and are used, but they still show gaps in comprehensiveness, accuracy, and coverage of niche or non-digitized materials.

Keep up-to-date records of circulation and materials, maintain inventory, and correct cataloging errors.

72

CI 72–72 · exposure 75 · augmentation 88 · importance 4.1/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510013/5Large urban and academic libraries are adopting AI-enhanced cataloging and inventory systems, but smaller public and rural libraries lag significantly due to budget and digitization gaps. Overall adoption is uneven: moderately fast in well-resourced information-sector institutions, slow in resource-constrained settings.
Sector adoption velocityclaude-sonnet-53/5Libraries have moderately adopted digital ILS and automated inventory systems, but many public and academic libraries still rely on manual verification and are slower tech adopters than corporate sectors.
Augmentation potentialclaude-haiku-4-5-202510014/5AI tools substantially assist librarians by automating error detection, flagging duplicate records, and suggesting cataloging corrections, allowing humans to focus on complex classification decisions and metadata enhancement. This creates a strong assistive layer that raises productivity without removing human judgment from the loop.
Augmentation potentialclaude-sonnet-55/5AI-assisted cataloging tools, automated error flagging, and inventory dashboards substantially boost librarian productivity while staff retain oversight for edge cases and judgment calls.
Task automatabilityclaude-haiku-4-5-202510014/5The task involves data management, record-keeping, and error correction—core capabilities of current AI systems. Library management systems can now integrate with AI for automated circulation tracking, inventory management, and cataloging error detection using OCR and database reconciliation, achieving substantial time savings, though some human judgment for complex cataloging decisions may remain.
Task automatabilityclaude-sonnet-54/5Circulation record-keeping, inventory tracking, and cataloging error correction are largely structured, rules-based data tasks well within current library management software and AI capabilities, though some physical inventory verification still requires human action.
Adoption barriersclaude-haiku-4-5-202510012/5Few regulatory or licensing barriers exist for automating these administrative and record-keeping functions; libraries are incentivized to adopt efficiency tools. Minor friction comes from staff retraining and preference to maintain human oversight of cataloging decisions, but nothing prevents substitution of the core tasks.
Adoption barriersclaude-sonnet-52/5No licensing requirement exists for this administrative task, though some institutional policy and quality-control oversight of catalog integrity remains common.
Cost vs. human wageclaude-haiku-4-5-202510014/5AI-powered library systems (circulation, inventory, cataloging engines) operate at far lower cost per transaction than hiring staff for these data-entry and verification tasks, especially at scale. The integration cost is modest relative to the loaded wage of library technicians performing these repetitive operations.
Cost vs. human wageclaude-sonnet-54/5Automated ILS software and batch cataloging tools cost far less per record processed than manual staff time for routine record-keeping and error correction.
Technical feasibility todayclaude-haiku-4-5-202510014/5Mature library management software with AI-assisted cataloging, barcode scanning, and automated inventory management systems are deployed in production at major libraries and institutions today. Systems reliably track circulation and flag cataloging inconsistencies, though oversight of complex corrections is still typically human-verified.
Technical feasibility todayclaude-sonnet-54/5Integrated Library Systems (ILS) with barcode/RFID scanning, automated cataloging validation, and MARC record error-checking are already deployed at scale in most libraries today.

Search standard reference materials, including online sources and the Internet, to answer patrons' reference questions.

70

CI 64–76 · exposure 67 · augmentation 88 · importance 4.1/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510013/5Libraries have begun piloting AI search assistants and chatbots, but deployment remains experimental in most institutions. Digital-forward sectors (universities, large systems) move faster, while smaller or less-resourced libraries lag.
Sector adoption velocityclaude-sonnet-53/5Libraries are a slower-moving public-sector/education-adjacent field, but AI search tools are being piloted and increasingly integrated into reference desks and library systems.
Augmentation potentialclaude-haiku-4-5-202510014/5AI-powered search tools significantly augment librarians' productivity by pre-screening sources, suggesting query refinements, and surfacing relevant materials faster, allowing librarians to focus on interpretation and patron guidance.
Augmentation potentialclaude-sonnet-55/5AI substantially speeds up reference search and answer drafting, letting librarians verify and contextualize results rather than manually searching multiple sources.
Task automatabilityclaude-haiku-4-5-202510013/5AI can search online sources and retrieve candidate answers reliably, but evaluating relevance, authority, and appropriateness for specific patron needs typically requires human judgment. The task could be partially automated with AI pre-filtering results, achieving modest time savings.
Task automatabilityclaude-sonnet-54/5AI search/chat systems can retrieve and synthesize reference information from online sources and answer factual questions with substantial time savings for many routine queries, though complex or specialized reference interviews still need human judgment.
Adoption barriersclaude-haiku-4-5-202510012/5Library patrons often prefer human assistance for complex questions, and some institutions have formal policies requiring human librarians for reference services, but no legal or strict licensing barriers prevent AI substitution of the search component itself.
Adoption barriersclaude-sonnet-52/5No licensing requirement mandates a human librarian for reference lookups, though some institutional preference for professional verification and accuracy concerns create mild friction.
Cost vs. human wageclaude-haiku-4-5-202510014/5API costs for search and LLM-based retrieval are now orders of magnitude cheaper than librarian wages per query handled. Integration and oversight overhead is low for routine reference tasks.
Cost vs. human wageclaude-sonnet-55/5Automated search/chat queries cost fractions of a cent per query versus a librarian's loaded wage for the same lookup task.
Technical feasibility todayclaude-haiku-4-5-202510014/5Deployed AI systems (search engines, retrieval-augmented generation tools, chatbots) already perform reference lookups at scale in production. However, human verification of answer quality and patron satisfaction assessment remains standard practice in most library settings.
Technical feasibility todayclaude-sonnet-54/5Production tools like AI-powered search engines, chatbots, and library-integrated AI assistants already answer reference questions reliably for common queries, though accuracy on niche or archival material is inconsistent.

Review and evaluate materials, using book reviews, catalogs, faculty recommendations, and current holdings to select and order print, audio-visual, and electronic resources.

52

CI 43–61 · exposure 53 · augmentation 75 · importance 4.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Library systems have historically lagged in digital transformation compared to other professional sectors. While some larger institutions pilot AI-assisted selection tools, most libraries still rely on manual review workflows, indicating slower adoption velocity in the sector overall.
Sector adoption velocityclaude-sonnet-52/5Libraries are historically slow adopters of AI tools compared to other professional services, with most current use limited to pilot programs or basic search/reporting tools.
Augmentation potentialclaude-haiku-4-5-202510014/5AI can substantially assist librarians by automating review aggregation, flagging relevant materials against current holdings, ranking recommendations by relevance, and generating summaries—allowing specialists to focus on judgment-heavy curation and institutional fit assessment.
Augmentation potentialclaude-sonnet-54/5AI can efficiently synthesize reviews, cross-reference catalogs, and surface trends in holdings, significantly speeding up the research phase of the librarian's evaluation process.
Task automatabilityclaude-haiku-4-5-202510014/5AI can process book reviews, catalogs, and holdings data to generate curated recommendations and flagged selections with high speed and consistency. However, final ordering decisions often require institutional judgment (budget constraints, curriculum alignment, patron needs) that still benefits from human oversight, limiting full 5-rating automation.
Task automatabilityclaude-sonnet-53/5AI can summarize reviews, compare catalogs, and flag gaps in holdings, but final selection requires judgment about institutional needs, budget, and curriculum fit that still requires human oversight.
Adoption barriersclaude-haiku-4-5-202510013/5Libraries face institutional friction around curation decisions (faculty input, collection development policies, community standards) and legal oversight of acquisitions (licensing, copyright compliance). Selection itself is not legally restricted to humans, but organizational practices and stakeholder preferences create meaningful adoption friction.
Adoption barriersclaude-sonnet-52/5No licensing requirement blocks AI assistance, but institutional purchasing policies, vendor relationships, and accountability for budget decisions create moderate friction.
Cost vs. human wageclaude-haiku-4-5-202510013/5AI-powered recommendation and evaluation systems cost roughly comparable to the overhead of a specialist's time spent on review and initial selection, though human review and final ordering approval still require staff involvement, preventing significant cost savings.
Cost vs. human wageclaude-sonnet-53/5AI can cheaply process reviews and holdings data, but integration with acquisitions systems and required human vetting keeps overall cost roughly comparable to a librarian's time for this task.
Technical feasibility todayclaude-haiku-4-5-202510013/5Products exist (recommendation engines, metadata aggregators, collection analytics tools) that assist in evaluation and selection, but they typically operate as decision-support systems rather than autonomous end-to-end ordering. Error rates and scope limitations remain material in production library systems.
Technical feasibility todayclaude-sonnet-52/5Some library systems offer collection-analysis tools and AI-assisted recommendation features, but no mature product autonomously performs full selection and ordering workflows in production at scale.

Maintain inventory of audio-visual equipment.

48

CI 39–57 · exposure 42 · augmentation 63 · importance 3.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Libraries and media institutions are traditionally slow adopters of automation technology; most still rely on manual or semi-automated systems. Digital transformation in this sector lags behind finance and information services.
Sector adoption velocityclaude-sonnet-52/5Libraries and educational institutions are generally slower adopters of advanced inventory automation compared to fast-moving sectors like finance or tech, though basic inventory software is common.
Augmentation potentialclaude-haiku-4-5-202510013/5AI-powered inventory tracking tools can assist librarians by automating barcode scanning, flagging discrepancies, and generating reports, meaningfully speeding up audit cycles while the human remains responsible for verification and decision-making on problem items.
Augmentation potentialclaude-sonnet-54/5AI-enabled inventory systems (barcode/RFID scanning, automated alerts, predictive maintenance flags) significantly streamline and reduce manual tracking effort for librarians managing AV equipment.
Task automatabilityclaude-haiku-4-5-202510012/5Inventory tracking of physical equipment involves scanning barcodes and updating databases, which AI can partially automate. However, verifying equipment condition, location accuracy in facilities, and handling exceptions (missing/damaged items) requires human judgment and physical inspection, preventing end-to-end automation at the 50% time-saving bar.
Task automatabilityclaude-sonnet-53/5Tracking AV equipment inventory (counts, locations, check-in/out, condition) can largely be automated using barcode/RFID systems and inventory software with AI-assisted tracking, though physical verification and handling still require human involvement.chunk
Adoption barriersclaude-haiku-4-5-202510013/5Institutional libraries typically require documented inventory procedures and human oversight of asset management for accountability and budgeting, though no legal mandate requires a licensed human to maintain the inventory itself. Organizational policies and audit requirements create moderate friction.
Adoption barriersclaude-sonnet-51/5There are no licensing or regulatory requirements for maintaining AV equipment inventories; it's an administrative task with minimal legal or liability barriers.
Cost vs. human wageclaude-haiku-4-5-202510013/5AI-assisted inventory systems (RFID scanning, database management) have modest setup and ongoing costs comparable to hiring staff for periodic inventory work, with significant organizational integration overhead offsetting labor savings.
Cost vs. human wageclaude-sonnet-53/5Inventory software subscriptions are relatively cheap compared to staff time, but initial setup, tagging equipment, and periodic physical checks still require paid labor, making savings moderate rather than order-of-magnitude.
Technical feasibility todayclaude-haiku-4-5-202510013/5Inventory management systems exist and can track equipment through barcodes and RFID, but deployed solutions often require manual audits, condition assessment, and exception handling. Real-world systems have material limitations in automated discovery and verification of physical assets.
Technical feasibility todayclaude-sonnet-53/5Asset management and inventory software products (e.g., barcode scanning apps, library asset trackers) are deployed and reliable for tracking, but full automation of physical audits and equipment condition checks still relies on humans.

Explain use of library facilities, resources, equipment, and services, and provide information about library policies.

44

CI 30–59 · exposure 38 · augmentation 63 · importance 3.9/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Libraries are traditionally slow-digitizing institutions with limited tech budgets and high emphasis on human customer service. While some large urban libraries pilot chatbots, widespread AI adoption for patron-facing information delivery remains nascent and unevenly distributed.
Sector adoption velocityclaude-sonnet-53/5Public and academic libraries have adopted chatbots and self-service kiosks at a moderate pace, but many still rely heavily on staffed reference desks.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist librarians by pre-drafting clear explanations of policies, generating FAQ content, and providing real-time lookup of resource details, allowing staff to focus on complex patron needs and personalized guidance. The assistance is meaningful but not transformative, as the human judgment and empathy remain central to the task.
Augmentation potentialclaude-sonnet-54/5AI tools can draft FAQ responses, power search interfaces, and triage common questions, letting librarians focus on complex reference needs.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can generate accurate information about library facilities, resources, and policies through retrieval systems, the task requires real-time responsiveness to diverse user contexts, clarification of nuanced policies, and dynamic troubleshooting of equipment issues. Current AI systems cannot handle the full end-to-end task with 50% time savings while maintaining equal quality across the heterogeneous patron interactions this task entails.
Task automatabilityclaude-sonnet-53/5Chatbots and AI assistants can answer routine FAQ-style questions about hours, policies, and resource locations, but nuanced in-person guidance and equipment troubleshooting still require human judgment.:
Adoption barriersclaude-haiku-4-5-202510013/5Public institutions face organizational friction and user preference for human staff contact, especially for vulnerable or non-digital-native patrons. However, there are no strict legal barriers preventing AI-assisted or AI-delivered explanations of publicly available policies and facility information.
Adoption barriersclaude-sonnet-52/5No licensing requirement exists for this informational task, though patrons often still prefer human interaction for complex or sensitive requests, creating mild friction.
Cost vs. human wageclaude-haiku-4-5-202510012/5Integration of AI systems requires significant library IT infrastructure, training data curation, and ongoing maintenance. The savings from automating a portion of routine questions is offset by these setup costs and the need for human oversight to catch errors in explaining complex policies or equipment use.
Cost vs. human wageclaude-sonnet-54/5An AI chatbot answering routine informational queries costs far less per interaction than staff time, though integration and oversight add some cost.
Technical feasibility todayclaude-haiku-4-5-202510012/5Chatbots and FAQ systems exist in some libraries, but they operate at narrow scope (policy lookup) and struggle with contextual questions, equipment-specific guidance, and patron-specific needs. Deployed products do not reliably perform the full task; humans remain gatekeepers for complex or unusual requests.
Technical feasibility todayclaude-sonnet-53/5Many libraries deploy chatbots and virtual reference services that handle basic policy/resource questions, but accuracy and scope are limited, especially for specialized collections or equipment help.

Arrange for interlibrary loans of materials not available in a particular library.

44

CI 25–64 · exposure 45 · augmentation 63 · importance 3.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Library automation adoption has been slow and fragmented; most interlibrary loan work still depends on legacy systems and manual review. Even digitization-forward libraries maintain human-centric workflows here, and public/academic library sectors have lagging overall digitization velocity compared to finance or information tech.
Sector adoption velocityclaude-sonnet-53/5Libraries have adopted ILL management software widely, but full AI-driven optimization is still emerging and libraries are generally slower technology adopters than commercial sectors.
Augmentation potentialclaude-haiku-4-5-202510013/5Current tools assist librarians by automating database searches, generating request forms, and tracking loan status, which raises throughput on routine requests. However, assistance is limited to well-structured portions; complex queries, holdings disambiguation, and patron education still require active human reasoning.
Augmentation potentialclaude-sonnet-54/5AI can substantially speed up locating materials, matching lending libraries, and automating request forms, letting librarians focus on exception handling and patron communication.
Task automatabilityclaude-haiku-4-5-202510012/5This task involves identifying availability across multiple library systems, filling out loan requests, and coordinating logistics—parts of which are automatable (database lookup, form generation). However, the human judgment required to interpret patron needs, select appropriate alternative materials, and handle edge cases means less than 50% time savings at equal quality with current systems.
Task automatabilityclaude-sonnet-54/5Interlibrary loan requests, tracking, and fulfillment are largely standardized workflows already handled by ILL software (ILLiad, WorldShare) that can be further automated with AI-driven matching and request routing, meeting the time-saving threshold for most of the process.
Adoption barriersclaude-haiku-4-5-202510014/5Significant barriers exist: interlibrary loan systems operate through formal consortial agreements requiring institutional authorization, liability for lost/damaged materials must be tracked, and many libraries require human judgment to navigate copyright/licensing restrictions and patron needs assessment. Regulatory frameworks around intellectual property and data sharing between institutions protect this workflow.
Adoption barriersclaude-sonnet-52/5No licensing requirement mandates a human librarian for ILL arrangement, though institutional agreements, courtesy protocols between libraries, and copyright compliance create some friction.
Cost vs. human wageclaude-haiku-4-5-202510012/5Interlibrary loan coordination involves specialized domain knowledge (library protocols, catalog systems, holdings databases) and integration with multiple institutional systems. The setup and maintenance costs are substantial relative to the relatively low per-transaction human labor cost for experienced library staff.
Cost vs. human wageclaude-sonnet-53/5Automated ILL systems reduce staff time meaningfully, but ongoing subscription costs, integration with library systems, and human oversight for edge cases keep the ratio only moderately favorable.
Technical feasibility todayclaude-haiku-4-5-202510012/5While library management systems exist and can query some union catalogs, end-to-end automation of interlibrary loan arrangement remains limited. Most institutions rely on semi-automated workflows where staff manually initiate requests, verify holdings, and handle exceptions; no mainstream product fully automates this task reliably today.
Technical feasibility todayclaude-sonnet-53/5Existing ILL systems automate much of the logistics, but they are rules-based rather than AI-driven, and exceptions (rare materials, licensing restrictions, negotiation with partner libraries) still require human intervention.

Analyze patrons' requests to determine needed information and assist in furnishing or locating that information.

41

CI 25–56 · exposure 38 · augmentation 75 · importance 4.1/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Library sector digitization is moderate; adoption of AI for patron-facing request analysis remains mostly in pilots (chatbots, discovery layer improvements) rather than production displacement. Legacy systems, limited IT budgets, and professional cautionary attitudes slow real-world adoption.
Sector adoption velocityclaude-sonnet-52/5Libraries are generally slower adopters of AI compared to finance or tech sectors, with pilots more common than widespread production deployment.
Augmentation potentialclaude-haiku-4-5-202510014/5AI tools meaningfully augment librarians' work by enabling faster searching, relevance ranking, and cataloging assistance, allowing them to focus on complex request interpretation and patron interaction. Current systems excel at reducing routine lookup time while humans guide the dialogue.
Augmentation potentialclaude-sonnet-54/5AI search and summarization tools significantly speed up librarians' ability to locate and synthesize information for patron requests while the librarian still handles interpretation and follow-up.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can help search databases and retrieve bibliographic information, the core task—understanding nuanced, often vague patron requests and determining actual information needs—requires contextual judgment and clarification dialogue that current systems struggle with. Automatability falls short of the 50% time-saving threshold because human librarians must typically validate and refine AI-retrieved results.
Task automatabilityclaude-sonnet-53/5AI chat/search assistants can interpret many reference queries and retrieve information quickly, but ambiguous or complex reference interviews requiring nuanced clarification and local resource knowledge still need human judgment.
Adoption barriersclaude-haiku-4-5-202510014/5Libraries operate under professional standards and patron expectations that a qualified human professional (often requiring an MLS) should understand and assist with information requests. Organizational culture, liability concerns around misinformation, and public trust norms create substantial friction against fully automated patron assistance.
Adoption barriersclaude-sonnet-52/5No licensing requirement for reference services, though patron trust, accuracy concerns, and institutional preference for human assistance create moderate friction.
Cost vs. human wageclaude-haiku-4-5-202510012/5Current AI systems (including chatbot overhead, integration, and required human review/correction) still cost more per successful request resolution than a librarian's loaded wage, especially when accounting for the need for human fallback and quality assurance.
Cost vs. human wageclaude-sonnet-54/5AI-based search/reference tools cost far less per query than a staffed reference desk, though integration with catalogs and oversight adds some cost.
Technical feasibility todayclaude-haiku-4-5-202510012/5No production system reliably performs full patron-request analysis and information location end-to-end. AI chatbots and search assistants exist but have high error rates in interpreting complex or ambiguous requests and often fail to match patron intent accurately enough for practical library deployment without human oversight.
Technical feasibility todayclaude-sonnet-53/5Library chatbots and AI search tools are deployed in some institutions for basic reference queries, but accuracy and coverage of specialized local collections remain limited in production use.

Develop, maintain, and troubleshoot information access aids, such as databases, annotated bibliographies, Web pages, electronic pathfinders, software programs, and online tutorials.

39

CI 25–52 · exposure 38 · augmentation 63 · importance 3.7/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Library sectors adopt digitization tools slowly relative to commercial sectors; many institutions remain moderately digitized with legacy systems. Actual AI-driven replacement in production library environments is minimal; most adoption is confined to pilots or narrow tool use for content creation assistance.
Sector adoption velocityclaude-sonnet-52/5Libraries and educational institutions are generally slower adopters of AI tooling compared to fast-moving tech/finance sectors, with pilots more common than full production deployment.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can meaningfully assist librarians by auto-generating metadata suggestions, drafting tutorial content, and identifying potential database inconsistencies, improving their productivity without removing human judgment from quality control and system design decisions.
Augmentation potentialclaude-sonnet-54/5AI substantially speeds up drafting of annotated bibliographies, tutorial content, and basic code/scripts, meaningfully boosting librarian productivity while they retain oversight of accuracy and system integration.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist with generating database schemas, bibliographic annotations, and tutorial content drafts, the full end-to-end task requires significant domain expertise, curatorial judgment, and understanding of user needs that current systems struggle with reliably. Troubleshooting and maintaining complex information systems remain largely manual and require human oversight.
Task automatabilityclaude-sonnet-53/5AI can draft bibliographies, generate web page content, and write scripts for pathfinders, but ongoing troubleshooting, database maintenance, and integration with library systems still require human oversight and domain expertise.
Adoption barriersclaude-haiku-4-5-202510014/5Libraries have strong organizational and professional norms favoring human curatorial judgment, and many institutions have licensing agreements with human information professionals. Additionally, the need for ongoing maintenance, user feedback integration, and accountability for information quality creates friction against full automation.
Adoption barriersclaude-sonnet-52/5No licensing requirement forces a human to do this, though institutional systems, legacy software, and quality control create moderate organizational friction.
Cost vs. human wageclaude-haiku-4-5-202510012/5Current AI tools require substantial human oversight, validation, and rework given the precision needed in bibliographic data and system reliability. The all-in cost (tool subscription, human review, maintenance) remains comparable to or exceeds the cost of a human specialist performing the work directly.
Cost vs. human wageclaude-sonnet-53/5AI tools can cheaply generate content and code snippets, but the specialized integration, testing, and maintenance work still requires paid human time, keeping costs roughly comparable for the full task.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed product reliably handles the full scope of developing and maintaining comprehensive information access aids independently. AI tools exist for narrow subtasks (content generation, basic metadata tagging) but lack the integrated capability to design, build, and troubleshoot multi-faceted information systems as librarians do.
Technical feasibility todayclaude-sonnet-53/5Products like AI coding assistants and content generators can produce drafts of tutorials, annotated bibliographies, and simple web pages today, but reliable end-to-end database/system troubleshooting in production library environments is limited.

Evaluate materials to determine outdated or unused items to be discarded.

36

CI 25–47 · exposure 33 · augmentation 75 · importance 3.7/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Library adoption of AI for operational tasks remains limited; most institutions rely on manual review and circulation reports rather than AI-driven deselection, reflecting conservative cultural and budgetary conditions in the sector.
Sector adoption velocityclaude-sonnet-52/5Libraries are generally slow technology adopters with limited budgets, and collection management software adoption for analytics-driven weeding remains uneven and mostly pilot-level.
Augmentation potentialclaude-haiku-4-5-202510014/5AI-powered usage analytics, recommendation systems, and metadata flagging can substantially assist librarians by surfacing candidates for review and providing decision support, while the librarian retains judgment over final retention decisions.
Augmentation potentialclaude-sonnet-54/5AI-driven usage analytics and reporting tools meaningfully speed up identification of weeding candidates, letting librarians focus review time on borderline or high-value items.
Task automatabilityclaude-haiku-4-5-202510012/5AI could assist in identifying usage statistics and flagging rarely-checked items, but determining what is truly 'outdated' or should be preserved requires nuanced judgment about institutional mission, patron needs, and collection value that exceeds current automation capability.
Task automatabilityclaude-sonnet-53/5AI can analyze circulation data, publication dates, and usage metrics to flag candidate items for deaccessioning, but final weeding decisions require contextual judgment about collection value, community needs, and rare/local materials that current systems can't fully replicate.
Adoption barriersclaude-haiku-4-5-202510014/5Librarians and collection specialists hold professional authority over collection stewardship; discarding decisions often require institutional review, stakeholder input, and accountability that creates organizational and professional friction against full automation.
Adoption barriersclaude-sonnet-52/5No licensing requirement dictates who performs weeding, though institutional policies often require professional librarian sign-off on deaccessioning decisions, creating mild organizational friction.
Cost vs. human wageclaude-haiku-4-5-202510012/5Building and maintaining an AI system for collection evaluation, including oversight for retention decisions, would likely exceed the cost of a librarian's time spent on selective batch review, especially given low-volume recurring cycles.
Cost vs. human wageclaude-sonnet-53/5Data-driven flagging tools are cheap to run compared to manual shelf-by-shelf review, but human verification still adds substantial cost, making overall savings moderate rather than dramatic.
Technical feasibility todayclaude-haiku-4-5-202510012/5While analytics tools can track circulation data and identify low-usage items, no deployed product reliably performs the full evaluative judgment of what should be discarded—this remains a human decision in library practice rather than an automated system.
Technical feasibility todayclaude-sonnet-52/5Some library systems offer analytics dashboards that surface low-circulation items, but no widely deployed product autonomously performs full collection weeding evaluations; librarians still manually review flagged lists.

Locate unusual or unique information in response to specific requests.

34

CI 30–39 · exposure 25 · augmentation 63 · importance 3.8/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Libraries and archives are traditionally slow to adopt automation due to their emphasis on specialized expertise, user relationships, and preservation of curatorial judgment. Although discovery tools and search systems have improved, displacement of human research specialists through AI remains limited and slow in the field.
Sector adoption velocityclaude-sonnet-52/5Libraries and information services are adopting AI search tools slowly and unevenly, with many still relying on manual expertise for specialized reference work.
Augmentation potentialclaude-haiku-4-5-202510013/5AI-powered search, citation tools, and information retrieval systems can meaningfully assist librarians by accelerating database searches and surfacing candidate materials, but the task of determining what is truly unusual or unique still depends on human judgment and domain expertise to evaluate results.
Augmentation potentialclaude-sonnet-54/5AI substantially assists librarians by quickly narrowing search spaces, suggesting sources, and drafting search strategies, even when final unusual-item location still requires human expertise.
Task automatabilityclaude-haiku-4-5-202510012/5Locating unusual or unique information requires deep domain knowledge, nuanced understanding of what constitutes 'unusual' in context, and judgment about relevance—capabilities where current AI systems struggle. While AI can search and retrieve information at scale, determining what is genuinely unique or unusual in response to specific, often vague requests requires human expertise and contextual reasoning that AI cannot reliably replicate end-to-end.
Task automatabilityclaude-sonnet-52/5AI can search and synthesize common information well but locating truly unusual, obscure, or archival-specific items often requires physical access, specialized databases, or judgment about source reliability that current AI cannot fully replace.5
Adoption barriersclaude-haiku-4-5-202510013/5Libraries and archives often have organizational policies favoring human expertise and face user expectations that a librarian will personally curate results. However, there are no hard legal or licensing barriers preventing AI-assisted research tools, though institutional inertia and professional standards create moderate friction.
Adoption barriersclaude-sonnet-52/5No licensing requirement blocks AI use, though institutional trust, accuracy expectations, and reliance on proprietary/archival materials create moderate friction.
Cost vs. human wageclaude-haiku-4-5-202510012/5The cost of AI search infrastructure, fine-tuning for specialized domains, human oversight of unusual findings, and verification of uniqueness claims is considerable, while a librarian's hourly cost for specialized research remains competitive for complex requests that require judgment.
Cost vs. human wageclaude-sonnet-53/5For straightforward lookups AI is much cheaper, but unusual requests often need extensive human verification and specialized access, narrowing the cost advantage.
Technical feasibility todayclaude-haiku-4-5-202510012/5No production systems reliably perform the full task of identifying genuinely unusual or unique information on demand. AI systems can search and retrieve information, but assessing uniqueness and relevance to a specific request's underlying intent remains a research problem; deployed products are limited to conventional search and retrieval.
Technical feasibility todayclaude-sonnet-52/5Deployed AI search/chat tools handle general reference queries but struggle with genuinely obscure or niche requests requiring access to specialized, non-indexed, or physical collections.

Train faculty and media staff on the use of software and audio-visual equipment.

33

CI 30–35 · exposure 25 · augmentation 75 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Libraries and educational institutions have begun adopting learning management systems and video tutorials, but live, in-person training by human librarians remains the standard; adoption of fully automated training is slow and limited to supplementary modules.
Sector adoption velocityclaude-sonnet-52/5Libraries and educational institutions are generally slower adopters of AI-driven training tools compared to fast-moving corporate or tech sectors.
Augmentation potentialclaude-haiku-4-5-202510014/5AI can meaningfully assist librarians by generating training scripts, creating video tutorials, designing interactive modules, and automating documentation, substantially raising instructor productivity while the librarian remains the authoritative trainer and troubleshooter.
Augmentation potentialclaude-sonnet-54/5AI can significantly help create training materials, FAQs, step-by-step guides, and even simulate practice scenarios, greatly boosting the librarian's efficiency in preparing and delivering training.
Task automatabilityclaude-haiku-4-5-202510012/5Training delivery has some automatable elements (automated video tutorials, interactive modules), but the task requires real-time adaptation to audience questions, technical troubleshooting, and contextual instruction that current AI cannot reliably handle end-to-end. Genuine human instruction with live feedback remains essential for >50% time savings at equal quality.
Task automatabilityclaude-sonnet-52/5Training involves live, adaptive interaction, hands-on troubleshooting, and relationship-building with faculty that current AI cannot fully replicate end-to-end, though AI can generate training materials and tutorials.'
Adoption barriersclaude-haiku-4-5-202510013/5Training effectiveness is partly dependent on institutional trust in instructors and human contact preference; however, there are no hard legal or regulatory barriers preventing AI-assisted or partially automated training delivery, only moderate organizational and pedagogical friction.
Adoption barriersclaude-sonnet-52/5No licensing requirement, but organizational preference for human-led training, equipment access issues, and need for real-time adaptive instruction create moderate friction against full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5While AI can lower the cost of content creation and self-paced learning modules, the replacement cost of actual live training (instructor labor, setup, interaction) remains substantial relative to AI inference; oversight and content quality assurance add overhead.
Cost vs. human wageclaude-sonnet-52/5Producing AI-generated training videos or documentation is cheap, but ongoing personalized instruction, live troubleshooting, and hands-on equipment training still require human labor, keeping blended costs comparable to human trainers.
Technical feasibility todayclaude-haiku-4-5-202510012/5AI systems can generate training materials and pre-recorded tutorials, but no deployed product reliably conducts live, interactive training sessions with hands-on equipment demonstration and personalized troubleshooting at the scale required for faculty and media staff.
Technical feasibility todayclaude-sonnet-52/5Some AI-driven chatbots and tutorial-generation tools exist for software help, but no deployed product reliably conducts full in-person or synchronous equipment/software training programs for staff.

Teach library patrons basic computer skills, such as searching computerized databases.

33

CI 30–35 · exposure 25 · augmentation 50 · importance 4.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Libraries have adopted digital tools and online tutorials slowly and unevenly. Most public and academic libraries remain heavily staffed with human librarians for instruction; no significant displacement or high-velocity adoption of AI teaching systems in the sector is evident.
Sector adoption velocityclaude-sonnet-52/5Public libraries are historically slow technology adopters with limited budgets, and this interpersonal teaching task sees minimal AI integration in practice.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist librarians by generating tutorial drafts, pre-screening patron questions, or offering interactive practice modules that patrons complete before a librarian session. However, the librarian's judgment, empathy, and adaptive teaching remain central to the task's success.
Augmentation potentialclaude-sonnet-53/5AI tutorials, chatbots, and guided walkthroughs can supplement a librarian's teaching materials and provide self-service options, moderately improving efficiency for common queries.
Task automatabilityclaude-haiku-4-5-202510012/5Teaching requires adaptive explanation, responsiveness to patron confusion, and personalized scaffolding—beyond current AI's reliable capabilities. While AI can draft tutorials or answer basic FAQs about databases, live teaching with real-time diagnosis of misunderstanding and adjustment of instructional approach remains fundamentally difficult for autonomous systems.
Task automatabilityclaude-sonnet-52/5Teaching involves live interaction, assessing patron skill level, and adapting to individual confusion, which current AI cannot fully replicate end-to-end in a real library setting.,, though chatbots could handle simple FAQ-style guidance.
Adoption barriersclaude-haiku-4-5-202510013/5Library institutions often emphasize human interaction and patron relationships as core value; teaching is seen as part of the librarian role. However, there are no hard legal or regulatory barriers preventing AI-assisted or AI-led instruction, and some institutions may adopt it readily.
Adoption barriersclaude-sonnet-52/5No licensing barrier exists, but many patrons (elderly, disabled, non-native speakers) prefer or require human assistance, and libraries value in-person community service as part of their mission.
Cost vs. human wageclaude-haiku-4-5-202510012/5Teaching labor remains cheap relative to the cost of building, integrating, and maintaining AI systems capable of reliable one-on-one or small-group instruction. Librarians' hourly cost is still competitive with the amortized cost of AI systems that perform this task adequately.
Cost vs. human wageclaude-sonnet-52/5While AI chat tools are cheap per interaction, effective in-person teaching for less tech-savvy patrons still requires human staff time, oversight, and troubleshooting, keeping costs comparable.
Technical feasibility todayclaude-haiku-4-5-202510012/5Chatbots and video tutorials exist, but deployed products do not reliably teach complex procedural skills to diverse learners with varying technical backgrounds. Current systems lack the contextual sensitivity and dynamic assessment needed to identify and remediate learning gaps in real-time during instruction.
Technical feasibility todayclaude-sonnet-52/5Some libraries deploy chatbots or online tutorials for basic database searching, but no product reliably replaces hands-on, adaptive one-on-one instruction for diverse patrons including those with low digital literacy.

Troubleshoot problems with audio-visual equipment.

33

CI 30–35 · exposure 25 · augmentation 50 · importance 3.8/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Libraries are low-digitization, laggard sectors for automation adoption. While some may use AI-assisted troubleshooting guides, production AI replacement of this task is rare and slow in cultural institutions.
Sector adoption velocityclaude-sonnet-52/5Libraries are generally slow adopters of AI-driven IT support tools, relying on traditional IT helpdesk or manual troubleshooting practices.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist by providing diagnostic flowcharts, suggesting common fixes, and matching symptoms to known issues, moderately boosting a technician's problem-solving speed. However, the heavy reliance on physical inspection limits the scope of augmentation.
Augmentation potentialclaude-sonnet-53/5AI chatbots and searchable manuals/knowledge bases can help staff quickly identify likely causes or solutions for common AV issues, improving speed of diagnosis even if physical fixes remain manual.
Task automatabilityclaude-haiku-4-5-202510012/5Troubleshooting AV equipment requires physical diagnosis, hands-on testing, and contextual problem-solving. While AI could assist with diagnosis flowcharts or symptom matching, current systems cannot physically inspect equipment, test connections, or manipulate hardware—core components needed for end-to-end troubleshooting at ≥50% time savings.
Task automatabilityclaude-sonnet-52/5Physical hands-on troubleshooting (checking cables, projectors, hardware faults) requires physical manipulation and diagnosis that current AI cannot perform end-to-end; AI can assist with diagnostic guidance but not execute repairs.
Adoption barriersclaude-haiku-4-5-202510013/5Libraries and institutions typically require staff to sign off on equipment status, and end-user safety and operational continuity create some institutional friction against full automation. However, no strict legal licensing bars AI-assisted diagnosis.
Adoption barriersclaude-sonnet-52/5No licensing requirement, but physical access to equipment and organizational reliance on on-site staff create moderate friction against full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI diagnostic support is cheap, but the labor cost for a librarian or technician to physically perform troubleshooting remains low relative to AI inference plus integration overhead. The task does not justify standalone AI replacement on cost grounds.
Cost vs. human wageclaude-sonnet-52/5AI diagnostic assistance is cheap, but since a human still must physically inspect and fix equipment, overall cost savings versus a librarian or IT staffer handling this are limited.
Technical feasibility todayclaude-haiku-4-5-202510012/5Deployed products can offer diagnostic guidance via chatbots or knowledge bases, but no end-to-end AI system reliably troubleshoots hardware problems without human intervention. The embodied nature of the task—examining cables, checking connections, replacing components—remains beyond production AI capabilities.
Technical feasibility todayclaude-sonnet-52/5Some chatbot/knowledge-base tools exist for IT troubleshooting guidance, but no deployed product reliably diagnoses and fixes AV equipment issues in libraries without human physical intervention.

Set up, adjust, and operate audio-visual equipment, such as cameras, film and slide projectors, and recording equipment, for meetings, events, classes, seminars, and video conferences.

33

CI 30–35 · exposure 25 · augmentation 50 · importance 3.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Adoption of AI-driven AV automation is slow in the library and educational sectors; most organizations still rely on human technicians or basic pre-programmed control, with full autonomous AV operation remaining rare even in well-funded institutions.
Sector adoption velocityclaude-sonnet-52/5Libraries and educational institutions are generally slower adopters of advanced AV automation compared to tech-forward corporate sectors, though basic conferencing tools are increasingly common.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist technicians through live monitoring, automated alerts, remote diagnostics, and control-panel assistance; these features improve efficiency and reduce setup time, but the human must remain in the loop to interpret context, make real-time decisions, and handle exceptions.
Augmentation potentialclaude-sonnet-53/5AI-enabled AV tools (auto-tracking cameras, voice-activated recording, smart conferencing platforms) can meaningfully assist a librarian in setup and operation, reducing manual adjustments during use.
Task automatabilityclaude-haiku-4-5-202510012/5While AI could theoretically control some AV equipment remotely or manage basic setup via scripts, the task requires real-time physical adjustment, troubleshooting variable hardware states, and responsive handling of live events—contexts where current AI lacks embodied capability and meaningful judgment. Partial automation (e.g., pre-programmed sequences) exists but cannot replace the full adaptive operation needed.
Task automatabilityclaude-sonnet-52/5This is a physical, hands-on task involving equipment setup and real-time operation that AI cannot perform without robotic embodiment; some smart AV systems can auto-configure but full setup/adjustment/operation remains largely manual.dı
Adoption barriersclaude-haiku-4-5-202510013/5Event venues often contract with or employ specialized AV technicians, and liability concerns around equipment failure during live events create organizational friction; however, no legal or licensing barrier strictly prevents automation, allowing incremental substitution where trust in reliability grows.
Adoption barriersclaude-sonnet-52/5No licensing barrier exists, but physical equipment handling, troubleshooting judgment, and on-the-spot adjustments during live events create practical friction against full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5Current AV automation and remote management systems require significant upfront investment in compatible hardware and integration, plus ongoing maintenance; the loaded cost per event often exceeds a trained technician, especially for small-to-medium institutions with variable event types.
Cost vs. human wageclaude-sonnet-52/5Automated AV systems (smart cameras, scheduling software) have upfront and integration costs comparable to or exceeding occasional human labor for setup tasks, especially in smaller institutions without dedicated infrastructure.
Technical feasibility todayclaude-haiku-4-5-202510012/5Some AV control systems offer automation and remote operation features, but deployed products are narrow (fixed setups, scripted workflows) and require human intervention for real-time problem-solving, equipment variability, and event-specific adjustments. No mature AI system reliably replaces a technician in a live or ad-hoc AV scenario.
Technical feasibility todayclaude-sonnet-52/5Smart conferencing systems (auto-framing cameras, automated recording) exist in production but typically still require a human to physically set up cables, projectors, and troubleshoot, so full task coverage isn't reliably automated.

Respond to customer complaints, taking action as necessary.

32

CI 30–34 · exposure 25 · augmentation 50 · importance 3.9/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Library systems have historically been slow to adopt AI automation, with limited digitization in many regions. Most complaint handling remains manual; automation adoption is in early pilot phases, not production.
Sector adoption velocityclaude-sonnet-52/5Libraries and public-sector institutions are generally slow adopters of AI for customer-facing complaint resolution, with most automation limited to basic FAQ bots.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist by summarizing complaint patterns, suggesting resolutions, and drafting responses that librarians then personalize and refine. This augmentation improves efficiency without removing the human from the loop.
Augmentation potentialclaude-sonnet-53/5AI can help draft responses, summarize complaint history, and suggest resolutions, giving moderate productivity benefit while a human remains in charge of decisions and actions.
Task automatabilityclaude-haiku-4-5-202510012/5AI can handle templated complaint acknowledgments and route simple issues, but cannot reliably assess complaint validity, negotiate resolution, or exercise the judgment required for complex patron disputes. The human element of trust and empathy is central to this task.
Task automatabilityclaude-sonnet-52/5Handling complaints often requires empathy, judgment about exceptions/policy, and authority to take corrective action (waive fines, resolve disputes) that current AI cannot reliably exercise end-to-end.
Adoption barriersclaude-haiku-4-5-202510013/5Libraries operate as public institutions with expectations of human responsiveness and accountability; patrons often require direct staff interaction for resolution. Organizational culture and customer preference for human contact create moderate friction.
Adoption barriersclaude-sonnet-53/5No licensing requirement, but institutional policy, patron preference for human interaction, and the need for discretionary authority to resolve disputes create moderate friction.
Cost vs. human wageclaude-haiku-4-5-202510012/5Integration costs for complaint handling systems, ongoing training, and human oversight for escalations make the all-in cost comparable to or higher than employing a specialist to handle complaints directly.
Cost vs. human wageclaude-sonnet-53/5AI-assisted triage/drafting is cheap, but complex complaints still need human review and decision-making, keeping blended costs roughly comparable to a human handling the full task.
Technical feasibility todayclaude-haiku-4-5-202510012/5Chatbots can partially handle routine inquiries, but deployed systems lack the contextual understanding and authority to take meaningful corrective action. Production deployments in libraries remain minimal; most systems are narrow pilots.
Technical feasibility todayclaude-sonnet-52/5Chatbots and ticketing systems can triage or draft responses to simple complaints, but resolving substantive disputes and taking action is not reliably handled by deployed products in library settings today.

Plan and teach classes on topics such as information literacy, library instruction, and technology use.

31

CI 30–32 · exposure 25 · augmentation 75 · importance 4.0/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Library instruction automation is slow. Sectors adopting AI-driven instruction (corporate training, higher ed) are moving gradually and remain primarily in pilot phases. Most public and school libraries continue traditional or hybrid models with limited AI integration in teaching.
Sector adoption velocityclaude-sonnet-53/5Academic and public libraries are experimenting with AI for content creation and chatbots, but actual classroom teaching automation remains rare and pilot-stage.
Augmentation potentialclaude-haiku-4-5-202510014/5AI meaningfully assists librarians in preparing lesson content, creating handouts, designing information literacy modules, and generating quizzes or assessment tools. These augmentations significantly reduce preparation time and improve instructional quality while keeping the librarian in the teaching role.
Augmentation potentialclaude-sonnet-54/5AI substantially helps librarians prepare lesson plans, quizzes, slides, and instructional materials, meaningfully boosting prep efficiency while the human still delivers instruction.
Task automatabilityclaude-haiku-4-5-202510012/5Teaching classes requires real-time classroom interaction, adaptive response to student questions, and dynamic pedagogical judgment. While AI can generate lesson plans and instructional materials, it cannot reliably conduct live instruction, manage classroom dynamics, or adjust teaching in response to student understanding at equal quality to a human instructor.
Task automatabilityclaude-sonnet-52/5AI can help draft lesson plans and materials, but live teaching requires interpersonal facilitation, adapting to student questions, and classroom management that current AI cannot fully perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510013/5Libraries and educational institutions have moderate adoption friction: stakeholders often prefer credentialed human instructors for pedagogical and accountability reasons, and there are cultural expectations around live teaching quality. However, no hard legal requirement mandates human-only instruction in all contexts.
Adoption barriersclaude-sonnet-53/5No formal licensing requires a human librarian to teach, but institutional norms, accreditation expectations, and student preference for live instructors create moderate friction.
Cost vs. human wageclaude-haiku-4-5-202510012/5The infrastructure cost for AI to generate personalized lesson content, manage learning platforms, and provide adaptive feedback is comparable to or exceeds the loaded cost of a human librarian delivering group instruction, especially when accounting for setup, customization, and oversight.
Cost vs. human wageclaude-sonnet-52/5AI can cheaply generate curriculum drafts, but actual classroom delivery still requires paid human instructor time, so overall cost savings are modest.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed product reliably teaches multi-student classes end-to-end today. AI can assist in content creation and tutoring individual students asynchronously, but classroom instruction at scale with heterogeneous learners remains beyond current reliable deployment. Chatbots show material limitations in real-time group teaching scenarios.
Technical feasibility todayclaude-sonnet-52/5No deployed product autonomously plans and delivers full instructional classes in libraries; AI tools are used as authoring aids rather than as instructors themselves.

Maintain hardware and software, including computers, media equipment, scanners, color copiers, and color laser printers.

30

CI 25–35 · exposure 25 · augmentation 50 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Library systems and media departments, particularly in smaller institutions, adopt IT automation slowly due to budget constraints, legacy infrastructure, and preference for established vendor service contracts. Adoption remains concentrated in large research libraries with dedicated IT staff.
Sector adoption velocityclaude-sonnet-52/5Library and facilities IT maintenance is a low-digitization, physical-labor-heavy niche with slow AI tool adoption compared to office/knowledge work sectors.
Augmentation potentialclaude-haiku-4-5-202510013/5AI-assisted monitoring systems (predictive alerts for failing hardware, automated patch recommendations) do usefully augment human technicians by prioritizing work and reducing unplanned downtime. However, the assistance remains largely informational rather than transformative to the core maintenance workflow.
Augmentation potentialclaude-sonnet-53/5AI chatbots and diagnostic tools can help identify software issues, suggest troubleshooting steps, or manage maintenance schedules, offering moderate assistance to the human performing physical upkeep.
Task automatabilityclaude-haiku-4-5-202510012/5Hardware and software maintenance requires physical intervention (replacements, repairs, cable management) and contextual troubleshooting that current AI cannot perform autonomously. While some diagnostic and software updates can be partially automated, the majority of tasks demand hands-on work that AI cannot execute.
Task automatabilityclaude-sonnet-52/5Physical maintenance and hands-on troubleshooting of hardware/equipment cannot be done by AI, though some software diagnostics could be assisted; overall time savings fall well short of 50%.
Adoption barriersclaude-haiku-4-5-202510014/5Significant barriers exist: equipment manufacturers often require authorized technicians for warranty compliance and liability purposes, institutional IT governance restricts who can modify systems, and safety/security concerns around equipment access are substantial. These licensing and organizational friction points limit substitution.
Adoption barriersclaude-sonnet-52/5No licensing barrier, but physical presence requirement for hardware maintenance creates a practical barrier to full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI-driven diagnostic tools and remote monitoring services have modest costs, but the physical maintenance and repair tasks still require human technicians. The loaded cost of human maintenance technicians in library settings is already relatively low per task, limiting the cost advantage of partial AI solutions.
Cost vs. human wageclaude-sonnet-52/5Physical repairs and hardware upkeep still require a human technician on-site, so AI cannot substitute the majority of the labor cost involved.
Technical feasibility todayclaude-haiku-4-5-202510012/5Some software patching and remote diagnostics are automated in IT environments, but comprehensive hardware maintenance across diverse equipment (scanners, printers, copiers) lacks reliable end-to-end automated solutions. Most deployed products handle narrow subsets (e.g., print queue monitoring) rather than full maintenance workflows.
Technical feasibility todayclaude-sonnet-52/5No deployed AI product autonomously maintains physical hardware like printers, scanners, or copiers; some IT helpdesk chatbots assist with software issues but reliability for full maintenance tasks is limited.

Plan and deliver client-centered programs and services, such as special services for corporate clients, storytelling for children, newsletters, or programs for special groups.

28

CI 25–30 · exposure 25 · augmentation 75 · importance 3.9/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Library systems are traditionally slow to automate (low digitization in some sectors, small-to-medium budgets, strong professional norms around human service); adoption of AI for administrative support exists but production-level automation of program planning and delivery remains rare.
Sector adoption velocityclaude-sonnet-52/5Libraries and cultural institutions are generally slow adopters of AI for public-facing programming, with pilots rare and mostly limited to back-office tasks.
Augmentation potentialclaude-haiku-4-5-202510014/5AI can meaningfully assist librarians by generating program ideas, drafting newsletters, managing scheduling, and suggesting content tailored to demographics—freeing time for human-centered relationship-building, creative customization, and live program delivery where the librarian remains central.
Augmentation potentialclaude-sonnet-54/5AI can meaningfully assist in generating newsletter content, program ideas, story scripts, and audience segmentation, boosting planning efficiency while humans still lead delivery.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist with content generation, scheduling, and newsletter drafting, the core task requires understanding diverse client needs, real-time audience engagement, and adaptive delivery that demands human judgment and personal connection—particularly for interactive programs like storytelling or corporate client relationship management.
Task automatabilityclaude-sonnet-52/5AI can help draft content like newsletters or story scripts, but planning and delivering client-centered programs requires in-person facilitation, needs assessment, and relationship management that current AI cannot execute end-to-end.'
Adoption barriersclaude-haiku-4-5-202510014/5Public libraries operate under community trust mandates and often require licensed librarians for program authority; corporate clients typically demand human relationship managers; regulatory and professional standards expect qualified human oversight of public services, creating strong adoption friction.
Adoption barriersclaude-sonnet-53/5No licensing requirement, but strong organizational and community expectations for human-led, personalized service create moderate friction against full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI tools for content generation and administrative tasks reduce some costs, but a trained librarian's deep domain knowledge, client relationship skills, and ability to deliver engaging programs in person remain difficult to replace cheaply; the all-in cost of AI assistance is comparable to or higher than partial human labor.
Cost vs. human wageclaude-sonnet-52/5While drafting assistance is cheap, the live delivery (storytelling sessions, corporate liaison work) still requires paid staff time, so overall cost savings are limited.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed products reliably execute full program planning and delivery end-to-end; AI tools exist for content drafting and scheduling, but human librarians remain essential for client consultation, program customization, and live delivery, especially for specialized or interactive services.
Technical feasibility todayclaude-sonnet-52/5No deployed product plans and delivers full library programming autonomously; existing tools only assist with content generation pieces, not the holistic service design and live delivery.

Develop library policies and procedures.

28

CI 25–30 · exposure 25 · augmentation 63 · importance 3.8/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Library adoption of AI for policy development is minimal and largely nascent; most libraries still rely on manual drafting, templates from associations, and human committees. No evidence of meaningful production-scale deployment in this function across the sector.
Sector adoption velocityclaude-sonnet-52/5Libraries are generally slower adopters of AI tools compared to fast-moving information/finance sectors, with AI use mostly limited to drafting assistance rather than policy-making itself.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can offer useful assistance by drafting boilerplate sections, suggesting regulatory language, or organizing policy structures for human review. However, the assistive value remains moderate because policy development is inherently judgmental and institution-specific, limiting how much AI can meaningfully elevate output.
Augmentation potentialclaude-sonnet-54/5AI can meaningfully assist by drafting policy language, summarizing best practices, comparing peer institution policies, and identifying regulatory considerations, significantly speeding up the human-led process.
Task automatabilityclaude-haiku-4-5-202510012/5Policy and procedure development requires nuanced organizational judgment, stakeholder input, and legal/ethical reasoning that current AI cannot reliably do end-to-end. While AI can draft text or suggest frameworks, the creative synthesis of institutional values, compliance needs, and community context remains fundamentally human-directed work.
Task automatabilityclaude-sonnet-52/5Drafting policy language can be assisted by AI, but developing actual policies requires institutional judgment, stakeholder negotiation, and contextual knowledge that AI cannot autonomously perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Significant organizational and governance barriers exist: library boards typically require human accountability for policy decisions, professional standards expect librarian authorship, and legal liability for compliance falls on the institution, not on automated systems. These friction points slow or prevent substitution.
Adoption barriersclaude-sonnet-53/5Policies often require approval by library boards, administrators, or governing bodies, and carry legal/liability implications (e.g., privacy, access), creating moderate organizational and procedural barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5Policy development requires institutional knowledge and accountability that currently demand human expertise; AI assistance does not yet reduce the labor cost below that of a librarian performing the task themselves, since substantial human oversight and revision is necessary.
Cost vs. human wageclaude-sonnet-52/5AI drafting assistance is cheap, but the overall task still requires substantial human deliberation, meetings, and institutional review, limiting overall cost savings.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed library management systems autonomously generate or substantially replace human-authored policies. AI can assist with templating or content suggestions in some tools, but production systems do not perform this task reliably without human authorship and organizational sign-off.
Technical feasibility todayclaude-sonnet-52/5No deployed product autonomously develops library policy; AI writing tools can produce draft text but library staff must define requirements, review legal/community context, and finalize decisions.

Evaluate vendor products and performance, negotiate contracts, and place orders.

28

CI 25–30 · exposure 25 · augmentation 63 · importance 3.6/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Libraries are information organizations but tend toward conservative, slower adoption of AI in procurement; most remain in pilot phases for AI-assisted vendor analysis rather than deployment of autonomous ordering systems.
Sector adoption velocityclaude-sonnet-52/5Library and cultural institution procurement processes are generally slow to digitize and adopt AI compared to fast-moving sectors like finance or tech.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can usefully assist by summarizing vendor specs, comparing pricing matrices, and drafting contract clauses, enabling librarians to focus on negotiation strategy and institutional fit—a productivity lift on analytical components while the human retains decision authority.
Augmentation potentialclaude-sonnet-54/5AI can meaningfully assist with vendor research, comparison summaries, price benchmarking, and drafting contract language, improving efficiency while humans retain final decision-making.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist with vendor comparison (specs, pricing) and draft contract language, the task requires judgment on performance trade-offs, relationship context, and custom negotiation—areas where AI cannot reliably replace the full decision-making loop without significant human oversight.
Task automatabilityclaude-sonnet-52/5Involves negotiation, relationship management, and judgment calls about institutional needs and budgets that current AI cannot fully replicate end-to-end, though drafting and analysis portions can be assisted.help.
Adoption barriersclaude-haiku-4-5-202510014/5Libraries often face organizational procurement policies, approval hierarchies, and vendor relationship continuity that require human sign-off; legal liability for contract terms and the need for institutional accountability create meaningful barriers to full automation.
Adoption barriersclaude-sonnet-53/5Contract negotiation and purchasing often require institutional authorization, procurement policies, and accountability for spending decisions, creating moderate organizational and legal friction.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI-based procurement support (LLMs for contract review, basic vendor comparison) is cost-competitive with outsourced research but does not yet undercut the librarian's loaded cost when legal and relationship judgment are factored in.
Cost vs. human wageclaude-sonnet-52/5Human negotiation and vendor relationship judgment still require significant oversight, so AI assistance reduces some labor but doesn't yet approach order-of-magnitude cost savings for the full task.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed product reliably handles the negotiation and contracting components end-to-end; AI tools can support document drafting and data analysis, but vendor evaluation and deal-making remain largely manual in library practice today.
Technical feasibility todayclaude-sonnet-52/5No deployed product autonomously evaluates library vendors, negotiates contracts, and places orders; AI is used only for narrow subtasks like summarizing product comparisons.

Direct and train library staff in duties, such as receiving, shelving, researching, cataloging, and equipment use.

26

CI 21–30 · exposure 25 · augmentation 50 · importance 3.7/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Libraries are traditionally low-digital-adoption sectors with limited AI infrastructure and strong professional norms around human mentorship. Even tech-forward libraries treat staff development as a core human management function rather than a candidate for automation.
Sector adoption velocityclaude-sonnet-52/5Libraries are a slower-adopting, often under-resourced public sector environment with limited AI integration into management and training workflows.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist librarians by generating training outlines, creating documentation, or providing learning resources, which could improve training consistency and reduce prep time. However, the interactive and relational core of staff direction limits transformative potential.
Augmentation potentialclaude-sonnet-53/5AI can help create training manuals, FAQs, or cataloging guides and answer staff questions, meaningfully supporting the training process even though it doesn't replace direct supervision.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist with cataloging guidance and provide training materials, the task requires real-time supervision, performance feedback, and interpersonal relationship-building that current AI systems cannot reliably deliver end-to-end. Training staff demands contextual judgment and adaptive response to individual learning needs.
Task automatabilityclaude-sonnet-52/5Directing and training staff involves interpersonal leadership, on-the-job coaching, and adaptive supervision that current AI cannot fully replicate end-to-end, though training materials could be AI-assisted.5;however the core management function remains human-led.
Adoption barriersclaude-haiku-4-5-202510014/5This task involves employment management, performance evaluation, and staff development, which carry legal liability and human-relations sensitivity. Organizations typically require human judgment and accountability in hiring, training, and personnel decisions due to employment law and professional standards in library services.
Adoption barriersclaude-sonnet-53/5No licensing barrier exists, but organizational norms expect a human supervisor to train and direct staff, and interpersonal trust/authority dynamics create moderate friction against full automation.
Cost vs. human wageclaude-haiku-4-5-202510011/5AI systems would still require significant human oversight, review, and intervention to ensure training quality and staff development. The cost of AI infrastructure plus necessary human supervisory layers would exceed the cost of direct librarian management.
Cost vs. human wageclaude-sonnet-52/5Since AI cannot substitute for the supervisory/training function itself, any cost comparison is limited to content-generation support, meaning the human labor cost for actual training delivery remains largely unavoidable.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed product reliably performs staff direction and training at scale. AI can generate training content or draft procedural documentation, but actual supervision, performance evaluation, and adaptive staff development remain beyond what current systems do reliably in production library environments.
Technical feasibility todayclaude-sonnet-52/5No deployed product autonomously manages and trains library staff; AI is at best used to generate training documents or answer procedural questions, not to perform supervisory training itself.

Confer with teachers to select course materials and to determine which training aids are best suited to particular grade levels.

23

CI 11–35 · exposure 13 · augmentation 75 · importance 3.6/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5K–12 and higher education sectors are slow to adopt autonomous AI systems for instructional material selection, with most adoption limited to assistive recommendation tools that still require librarian gatekeeping. Digital transformation in schools lags information and professional services sectors.
Sector adoption velocityclaude-sonnet-52/5K-12 and library settings are moderate-to-slow AI adopters, with usage concentrated in content search/recommendation rather than replacing consultative planning conversations.
Augmentation potentialclaude-haiku-4-5-202510014/5AI can substantially assist librarians by generating filtered material recommendations by grade level, subject, and learning standard, allowing them to focus on conferencing with teachers about fit and pedagogy. This augmentation significantly raises librarian productivity while preserving professional judgment.
Augmentation potentialclaude-sonnet-54/5AI can help librarians quickly identify age-appropriate materials, summarize resources, and prepare talking points, meaningfully speeding preparation for these conversations even though the human interaction remains central.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can recommend materials based on grade levels and curricula, the task requires nuanced judgment about pedagogical fit, teacher preferences, and institutional context that current systems handle only partially. The back-and-forth conferencing aspect with teachers involves interpersonal negotiation that AI cannot reliably perform end-to-end.
Task automatabilityclaude-sonnet-51/5This is an interpersonal conferring and negotiation task requiring live judgment about specific students, curricula, and relationships that AI cannot conduct end-to-end today.
Adoption barriersclaude-haiku-4-5-202510014/5Educational institutions typically require librarians or media specialists to make curated selections, and teachers often prefer direct human consultation on pedagogical fit. Schools maintain institutional controls and professional norms that reinforce human decision-making in this role.
Adoption barriersclaude-sonnet-52/5No licensing requirement blocks AI assistance, but organizational norms and the inherently relational nature of conferring with teachers create real friction against full substitution.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI tools for educational recommendations are available but typically require significant librarian review and curation to validate suggestions, limiting labor savings. The librarian's domain expertise remains necessary, making all-in costs comparable to or exceeding human labor for this task.
Cost vs. human wageclaude-sonnet-52/5AI tools could cheaply generate material lists, but the actual conferring and contextual judgment still requires a paid human, so overall cost savings are limited.
Technical feasibility todayclaude-haiku-4-5-202510012/5No production systems reliably perform librarian-teacher conferencing autonomously today. Educational material recommendation tools exist but require heavy human oversight and do not capture the conversational, context-specific nature of material selection with educators.
Technical feasibility todayclaude-sonnet-51/5No deployed product substitutes for a librarian's collaborative conversation with teachers about grade-appropriate materials; this remains a human-to-human consultative process.

Assemble and arrange display materials.

21

CI 19–24 · exposure 16 · augmentation 38 · importance 3.1/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Library systems are generally slower adopters of automation, with limited digitization pressure on physical display work; this task remains largely manual and human-centered in practice.
Sector adoption velocityclaude-sonnet-51/5Libraries are a low-digitization, physical-space-dependent sector with minimal AI adoption for tasks involving physical merchandising or display work.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist librarians by suggesting thematic arrangements, generating design mockups, or recommending materials for a display, meaningfully raising planning and curation efficiency while the librarian executes the physical assembly.
Augmentation potentialclaude-sonnet-52/5AI could help suggest themes, layouts, or generate design mockups for displays, but it offers limited direct assistance with the physical assembly itself.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can help design layouts and suggest arrangements digitally, physical assembly and spatial arrangement of display materials requires manual dexterity, real-time visual judgment, and adaptation to physical constraints that current AI systems cannot execute end-to-end without human intervention.
Task automatabilityclaude-sonnet-52/5Physically arranging displays, books, and materials in a library requires manual manipulation of physical objects and spatial-aesthetic judgment that current AI cannot perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510013/5Libraries often have institutional preferences for human curation and public-facing arrangement work; while no strict legal barrier exists, organizational culture and the creative judgment valued in displays create moderate friction against full automation.
Adoption barriersclaude-sonnet-52/5No licensing requirement exists, but the physical, hands-on nature of arranging materials in a real space creates practical barriers to automation beyond simple digital work.
Cost vs. human wageclaude-haiku-4-5-202510011/5The physical labor component means a human librarian remains necessary, and any AI assistance (design software, layout suggestions) does not reduce the core cost of human time spent on assembly and arrangement.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute for physical display assembly, so any AI-based approach would require robotics or human labor anyway, making AI more costly or infeasible.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product reliably performs physical assembly and arrangement of display materials autonomously; this task fundamentally requires embodied robotics or sustained human presence, which remains research-stage for library contexts.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product physically assembles or arranges library displays; this remains a manual, in-person task requiring human dexterity and taste.

Confer with colleagues, faculty, and community members and organizations to conduct informational programs, make collection decisions, and determine library services to offer.

19

CI 7–30 · exposure 13 · augmentation 50 · importance 4.0/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Library sectors remain relatively low-digitization, human-centered institutions with slower adoption of AI-driven workflow changes; pilot adoption of meeting support tools is emerging but production displacement is minimal.
Sector adoption velocityclaude-sonnet-52/5Libraries and educational/community institutions are generally slower adopters of AI for stakeholder engagement functions compared to fast-moving information sectors.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist by preparing meeting agendas, summarizing stakeholder feedback, and drafting documentation, meaningfully reducing prep work and increasing a librarian's capacity to conduct more consultations and manage larger datasets of community input.
Augmentation potentialclaude-sonnet-53/5AI can help summarize community feedback, draft agendas, analyze circulation data for collection decisions, or prepare briefing materials, meaningfully supporting but not replacing the conferring process.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can draft communications and synthesize information for librarians to review, the task fundamentally requires human judgment to weigh community needs, stakeholder preferences, and organizational constraints in real-time dialogue. AI cannot reliably conduct the interpersonal negotiation and consensus-building that 'conferring' demands.
Task automatabilityclaude-sonnet-51/5This task is fundamentally relational and consultative, requiring live human negotiation, trust-building, and contextual judgment about community needs that AI cannot conduct end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Librarians often hold public-facing roles and legal authority over collection decisions; many organizations require human accountability for resource allocation and community engagement, creating organizational and quasi-legal friction against full automation.
Adoption barriersclaude-sonnet-53/5No licensing mandates a human specifically for this, but strong organizational and community expectations for human relationship management and institutional trust create real friction against automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5Deploying AI to assist with meeting preparation and documentation adds cost (infrastructure, oversight) without displacing the librarian's core role of judgment and relationship-building; the human remains essential and the all-in cost remains comparable to hiring the librarian.
Cost vs. human wageclaude-sonnet-51/5Since AI cannot substitute for the human presence and relationship-building required, the human cost remains necessary and AI adds cost as a supplementary tool rather than replacing labor.
Technical feasibility todayclaude-haiku-4-5-202510012/5No production system reliably performs stakeholder conferencing and collection-decision facilitation end-to-end. AI chatbots can support note-taking and agenda-drafting, but they cannot autonomously represent library interests or synthesize complex organizational preferences in actual meetings.
Technical feasibility todayclaude-sonnet-51/5No deployed product independently confers with stakeholders, facilitates meetings, or negotiates collection/service decisions on behalf of a librarian in production settings.

Supervise daily library operations, budgeting, planning, and personnel activities, such as hiring, training, scheduling, and performance evaluations.

16

CI 7–25 · exposure 13 · augmentation 50 · importance 4.1/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Libraries are public-sector or nonprofit institutions with slower digital adoption and strong human-centered service cultures; pilot adoption of HR analytics tools exists, but supervisory automation is minimal and viewed with institutional skepticism.
Sector adoption velocityclaude-sonnet-52/5Libraries and public/educational institutions are typically slower adopters of AI for managerial functions compared to fast-moving private sectors.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can provide useful assistance on scheduling optimization, budget forecasting, and performance data visualization, enabling supervisors to spend less time on routine administrative analysis and more on mentorship and strategic planning.
Augmentation potentialclaude-sonnet-53/5AI can help with budgeting spreadsheets, scheduling optimization, drafting performance review language, and planning documents, meaningfully aiding but not replacing the supervisory role.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist with scheduling algorithms, budget analysis, and performance metric compilation, the core supervisory task—hiring decisions, personnel evaluation, training oversight, and judgment calls about library operations—requires human accountability and contextual judgment that current AI systems cannot reliably perform end-to-end at the 50% time-saving threshold.
Task automatabilityclaude-sonnet-51/5This is a managerial task involving human judgment, interpersonal decisions, and organizational leadership that cannot be executed end-to-end by current AI systems.
Adoption barriersclaude-haiku-4-5-202510014/5Hiring, training, and performance evaluations carry legal liability (employment law, discrimination risk), and organizational policy typically requires a human manager to sign off on personnel decisions; additionally, library board oversight and institutional governance create friction against full automation.
Adoption barriersclaude-sonnet-54/5Personnel decisions like hiring, evaluations, and terminations carry legal, HR, and liability requirements that generally require human accountability and signoff.
Cost vs. human wageclaude-haiku-4-5-202510012/5The human salary for a library supervisor is substantial, and the cost of integrating AI tools plus oversight (to ensure hiring and personnel decisions are sound) approaches or exceeds the savings from partial automation of scheduling and budgeting tasks.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing this managerial bundle, so cost comparison favors the human entirely.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed product reliably executes full library operations supervision; existing HR software handles narrow subtasks (scheduling, payroll), but hiring, performance evaluation, and strategic operational decisions remain dependent on human judgment and are not automated in production systems.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously supervises staff, conducts hiring, or manages personnel evaluations; these remain human-led activities with AI only as a peripheral tool.

Engage in professional development activities, such as taking continuing education classes and attending or participating in conferences, workshops, professional meetings, and associations.

5

CI 0–10 · exposure 0 · augmentation 50 · importance 3.6/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5This task is not subject to automation adoption because it inherently requires human participation. No sector is automating professional development activities away.
Sector adoption velocityclaude-sonnet-51/5This is an inherently human, identity- and career-based activity not subject to AI adoption trends in the way substitutable work tasks are.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist by recommending relevant conferences, summarizing proceedings, organizing professional materials, or suggesting continuing education courses aligned with career goals, but the actual engagement remains human-driven.
Augmentation potentialclaude-sonnet-53/5AI can help identify relevant conferences, summarize learning materials, generate study notes, or curate continuing education content, moderately aiding preparation and follow-up.
Task automatabilityclaude-haiku-4-5-202510011/5Professional development is fundamentally a human activity requiring personal judgment about career growth, networking, and skill acquisition. AI cannot attend conferences, participate in professional discussions, or meaningfully engage in the learning process itself.
Task automatabilityclaude-sonnet-51/5This task requires the human librarian to personally engage in learning, networking, and professional growth activities; AI cannot attend conferences or complete continuing education on someone's behalf.
Adoption barriersclaude-haiku-4-5-202510015/5Professional development is a volitional human activity tied to career advancement and personal agency. There are no pathways to automate human judgment about which development activities to pursue or how to benefit from them.
Adoption barriersclaude-sonnet-53/5While no legal requirement mandates a human perform this, professional norms, certification/licensing renewal requirements, and organizational expectations tie this to the individual professional's identity and credentials.
Cost vs. human wageclaude-haiku-4-5-202510011/5Professional development requires human time investment and associated costs (registration, travel, materials). AI cannot substitute for this; if anything, an AI system would add cost rather than replace human participation.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute for the human activity itself, so cost comparison is not meaningful; the human must incur the cost of participation regardless.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product can autonomously attend conferences or participate in professional meetings. While AI can summarize conference materials or schedule tasks, the core task of engaging in professional development is inherently human-driven.
Technical feasibility todayclaude-sonnet-51/5No deployed product substitutes for a person's professional development participation, as the value lies in the individual's personal learning and networking.

Represent library or institution on internal and external committees.

0

CI 0–0 · exposure 0 · augmentation 38 · importance 3.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5No adoption pathway exists because institutional and legal structures require humans to represent organizations. Even highly digitized sectors maintain this as a human-only function.
Sector adoption velocityclaude-sonnet-51/5Committee representation is a governance/relational function with essentially no AI adoption trend in libraries or elsewhere.
Augmentation potentialclaude-haiku-4-5-202510012/5AI might assist by preparing briefing materials or summarizing committee agendas beforehand, but the core task of representation itself offers limited augmentation opportunity.
Augmentation potentialclaude-sonnet-53/5AI can help prepare briefing materials, summarize prior meeting notes, or draft talking points, aiding preparation even though it cannot perform the representation itself.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires live committee participation, negotiation, relationship-building, and real-time decision-making. AI cannot meaningfully represent an institution's interests or engage in the interactive, authority-dependent aspects of committee work.
Task automatabilityclaude-sonnet-51/5Representing an institution on committees requires relationship-building, judgment, and real-time human presence/advocacy that current AI cannot perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Committees require authorized institutional representatives with accountability, decision-making authority, and legal standing. Organizational governance structures mandate human participation in these roles.
Adoption barriersclaude-sonnet-55/5This is an inherently interpersonal, representational, and often politically/organizationally sensitive role requiring a human with institutional authority and accountability.
Cost vs. human wageclaude-haiku-4-5-202510011/5A librarian's salary for committee work cannot be meaningfully compared to AI cost because the task fundamentally requires human representation—cost comparison is moot when substitution is infeasible.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute delivering this output, so cost comparison favors the human by default.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI system can serve as a legal or institutional representative on committees. Human presence and accountability are non-negotiable prerequisites that no current technology addresses.
Technical feasibility todayclaude-sonnet-51/5No deployed product substitutes for a human representative in committee settings; this remains firmly outside AI product scope.

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.