Library Assistants, Clerical
43-4121.00Median wage $36,910/yr85,520 employed (US)Rank #102 of 923 scored · top 11% by substitution
Compile records, and sort, shelve, issue, and receive library materials such as books, electronic media, pictures, cards, slides and microfilm. Locate library materials for loan and replace material in shelving area, stacks, or files according to identification number and title. Register patrons to permit them to borrow books, periodicals, and other library materials.
Sub-scores
0–100 · band = confidence interval from rater disagreement
Substitution46
Exposure43
Augmentation53
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
32 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
25%
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%44
panel mean rating 2.7/5 → substitution pressure 44/100
Technical feasibility todayw 20%41
panel mean rating 2.6/5 → substitution pressure 41/100
Cost vs. human wagew 15%46
panel mean rating 2.8/5 → substitution pressure 46/100
Adoption barriersw 20%inverted — strong barriers lower the score65
panel mean rating 2.4/5 (barrier strength) → substitution pressure 65/100
Sector adoption velocityw 10%29
panel mean rating 2.2/5 → substitution pressure 29/100
Task breakdown (32 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.
Send out notices and accept fine payments for lost or overdue books.
89CI 79–100 · exposure 87 · augmentation 50 · importance 3.8/5 · click for rater detail
Send out notices and accept fine payments for lost or overdue books.
89| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 5/5 | Library systems have been automating notice and payment processing for 15+ years; most medium-to-large libraries use integrated systems, and cloud-based solutions are rapidly spreading automation into smaller institutions. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Public and academic libraries have widely adopted integrated library systems with automated notifications and online payments over the past decade, though smaller libraries lag. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI/automation assists library staff by handling routine notices and payments, freeing them to focus on dispute resolution and patron communication, though the human role is diminishing rather than being substantially elevated. |
| Augmentation potential | claude-sonnet-5 | 3/5 | Where not fully automated, AI/software still assists clerical staff by flagging overdue accounts and pre-drafting notices, though the task is largely already automated rather than merely augmented. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Sending automated notices and accepting payments for overdue/lost books are largely automatable tasks. Email/SMS notifications and payment processing via web forms or APIs can handle the majority of this workflow, though handling edge cases (disputes, special circumstances) requires some human judgment. This likely exceeds the 50% time-saving threshold. |
| Task automatability | claude-sonnet-5 | 5/5 | Automated notice generation and payment processing are routine, rule-based tasks fully handled by existing library management software and online payment systems with no quality loss. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few hard barriers exist; libraries are already automating this task at scale. Some friction comes from patron communication preferences and the need for human oversight of payment disputes, but no legal or licensing requirement mandates human performance of this task. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing, liability, or regulatory requirement mandates human involvement in sending notices or collecting small fines. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Automated email/SMS notice systems cost pennies per message, and payment processing fees are typically 2-3% of transaction value. This is an order of magnitude cheaper than paying library staff (loaded hourly wage ~$20-30+) to manually process notices and payments. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated batch notice systems and online payment portals cost a fraction of a cent per transaction versus staff time for manual notice mailing and cash handling. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Mature library management systems and automated notification platforms already perform these functions reliably in production across thousands of libraries worldwide. Payment processing is well-established through standard online systems, though integration complexity varies by institution. |
| Technical feasibility today | claude-sonnet-5 | 5/5 | Integrated library systems (e.g., Koha, Sierra, Polaris) already send automated overdue notices and process fine payments online in production at scale across libraries. |
Review records, such as microfilm and issue cards, to identify titles of overdue materials and delinquent borrowers.
86CI 72–100 · exposure 87 · augmentation 63 · importance 3.8/5 · click for rater detail
Review records, such as microfilm and issue cards, to identify titles of overdue materials and delinquent borrowers.
86| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Many libraries have adopted automated overdue systems, but conversion of legacy microfilm collections and implementation in smaller or underfunded systems remain incomplete; adoption is steady but not universal. |
| Sector adoption velocity | claude-sonnet-5 | 5/5 | Library systems have already deeply adopted automated circulation and overdue-tracking software as a baseline standard practice, not just a pilot. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI-assisted systems help clerks prioritize delinquent accounts, flag suspicious patterns (e.g., systematic non-returns), and generate mailing lists, substantially raising the productivity of staff review and follow-up actions. |
| Augmentation potential | claude-sonnet-5 | 3/5 | Where legacy microfilm or non-digitized records still exist, AI/OCR tools can assist in extracting and cross-referencing information, but this applies to a shrinking subset of libraries still using non-digital records. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Current OCR and record-matching systems can reliably extract titles and borrower IDs from microfilm and digital records, then cross-reference due dates against current holdings with minimal human intervention, achieving >50% time savings on typical workflows. |
| Task automatability | claude-sonnet-5 | 5/5 | This is a structured data-lookup and matching task (comparing due dates against current date to flag overdue items and borrowers) that is trivially handled by database queries and automated systems already integrated into library management software. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Library policies may prefer human review for dispute resolution and patron relations, but there are no legal or licensing barriers preventing automated overdue detection; libraries control their own adoption pace. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing, liability, or human-judgment requirement exists for identifying overdue materials; this is purely administrative record-keeping with no legal barrier to automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | OCR and database query costs are negligible compared to the hourly wage of a clerical library assistant; a fully automated pipeline costs orders of magnitude less per identification than paying staff to manually review records. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated database queries cost fractions of a cent compared to manual clerical review of physical or microfilm records, an order-of-magnitude or greater savings. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Library management systems with integrated OCR, barcode scanning, and automated overdue detection are widely deployed in production; systems like Alma and Koha handle this task at scale, though legacy microfilm conversion still requires some manual oversight. |
| Technical feasibility today | claude-sonnet-5 | 5/5 | Integrated Library Systems (ILS) like Koha, Sierra, and Alma have automated overdue tracking and delinquent borrower reports in production use across virtually all libraries for decades. |
Enter and update patrons' records on computers.
81CI 72–90 · exposure 87 · augmentation 75 · importance 4.3/5 · click for rater detail
Enter and update patrons' records on computers.
81| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Libraries are information-sector organizations, but adoption of AI agents for clerical tasks remains in pilot/early phases; many still rely on manual entry or basic ILS workflows rather than full agent automation. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Public libraries are often under-resourced and slower to modernize systems compared to fast-adopting private sectors, though self-service kiosks are increasingly common. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can assist staff by auto-populating fields, suggesting corrections, and flagging data anomalies before submission, significantly reducing manual review time while a human remains in quality-control loop. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-assisted forms, autofill, and validation tools can meaningfully speed up clerical staff's data entry and reduce errors while they remain in the loop. |
| Task automatability | claude-haiku-4-5-20251001 | 5/5 | Entering and updating patron records is a highly structured, rule-based task with clear data entry patterns. Current AI systems with access to patron management databases can automate this end-to-end with 50%+ time savings using form-filling agents and API integrations. |
| Task automatability | claude-sonnet-5 | 4/5 | Data entry and record updates in structured library management systems are highly routinizable and can largely be automated via self-service kiosks, online forms, and API integrations with minimal human intervention.- |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | No licensing requirement for AI to modify patron records; main friction is organizational (staff retraining, system integration effort) and minor privacy/audit oversight, but nothing legally prevents automation. |
| Adoption barriers | claude-sonnet-5 | 2/5 | Minor privacy/data protection considerations exist, but no licensing requirement mandates a human clerk perform basic record entry. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Inference cost for data entry is negligible (cents per record), while human data entry assistants cost $25–35k/year loaded; automation is 1–2 orders of magnitude cheaper per record processed. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated data entry via self-service portals or batch imports costs a fraction of a clerical worker's time per record once systems are configured. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Library management systems (ILS platforms like Evergreen, Koha, Symphony) have mature automation capabilities, and RPA/agent tools routinely handle database record updates in production at scale across institutions. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Most integrated library systems (ILS) already support automated patron self-registration, online renewals, and record syncing, and many libraries deploy these in production today. |
Maintain records of items received, stored, issued, and returned and file catalog cards according to system used.
78CI 72–84 · exposure 75 · augmentation 63 · importance 4.0/5 · click for rater detail
Maintain records of items received, stored, issued, and returned and file catalog cards according to system used.
78| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Library systems are heavily digitized and have been adopting automated cataloging and tracking systems for decades. Modern libraries increasingly deploy self-checkout, RFID inventory, and integrated management systems; adoption in well-resourced institutions is already mature and widespread. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Public libraries and educational institutions have adopted ILS/circulation automation broadly, but adoption pace varies with funding and digitization level, and many smaller libraries retain manual elements. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI-assisted cataloging, intelligent search suggestions, and automated data validation can significantly boost the productivity of library assistants who remain in the loop, allowing them to focus on complex items or patron interactions while systems handle routine ingestion and filing. |
| Augmentation potential | claude-sonnet-5 | 3/5 | For remaining manual aspects (e.g., transitioning old card catalogs or exception handling), AI-assisted data entry and search tools can meaningfully speed up the clerical work while humans oversee accuracy. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Record-keeping and filing of catalog cards are largely data-entry and organizational tasks that can be substantially automated using document processing, database systems, and categorization algorithms. Current systems can handle scanning, data extraction, and automated filing with minimal human intervention, easily achieving >50% time savings. |
| Task automatability | claude-sonnet-5 | 4/5 | This is largely structured data entry and record-keeping that integrated library systems (ILS) and barcode/RFID scanning already automate; catalog filing is essentially obsolete in most digital-first systems. Remaining manual work is mostly exception handling and legacy card catalogs. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | While no legal requirement mandates human involvement in record-keeping or filing, some organizational conservatism, staff retention concerns, and the need for human oversight of system configuration create modest friction. However, these are not hard licensing or liability barriers, and substitution is technically unblocked. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing, liability, or regulatory requirement mandates a human perform routine record maintenance or filing; this is a purely administrative task with minimal friction to automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Automated systems for tracking and cataloging items have very low per-transaction costs once deployed, and inference/integration costs are negligible compared to the loaded wage of a clerical library assistant ($35–45k annually). The cost advantage is substantial and multi-order-of-magnitude. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated circulation/inventory systems cost far less per transaction than manual clerical record-keeping once deployed, though initial system integration and maintenance carry some cost. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Library management systems and automated cataloging software are mature, production-deployed technologies used in thousands of libraries worldwide. Barcode scanning, RFID systems, and integrated library management platforms (e.g., Koha, Evergreen) reliably track items and maintain records at scale, though some edge cases may require human review. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Modern ILS platforms (e.g., Koha, Ex Libris, SirsiDynix) reliably track item circulation, holdings, and status at scale in production libraries today, though legacy card catalog filing is now rare and not directly served by AI products. |
Inspect returned books for condition and due-date status and compute any applicable fines.
77CI 65–89 · exposure 75 · augmentation 50 · importance 4.1/5 · click for rater detail
Inspect returned books for condition and due-date status and compute any applicable fines.
77| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Libraries are typically under-resourced, conservative institutions with slow IT adoption. While some larger systems use automated sorters and barcode readers, widespread deployment of end-to-end inspection and fine automation in library networks remains limited; most adoption is in specialized sorter hardware, not AI agents. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Public and academic libraries have broadly adopted automated circulation and fine systems for years, representing mature, deep adoption in this specific workflow. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist library staff by pre-screening books, flagging condition issues, and auto-calculating fines, allowing assistants to focus on exceptions and customer service. This is useful productivity lift on routine parts of the task, though the human remains in the loop for judgment calls on damage severity. |
| Augmentation potential | claude-sonnet-5 | 3/5 | For the physical inspection component, AI offers limited assistance today, though software augments staff by auto-calculating fines and flagging overdue status. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Current AI can automate most of this task: optical character recognition (OCR) can read due dates and barcodes, computer vision can assess book condition (damage, wear), and rule-based systems can compute fines. End-to-end automation with modest setup achieves >50% time savings, though edge cases (ambiguous damage severity) may require human review. |
| Task automatability | claude-sonnet-5 | 4/5 | Fine computation from due dates is trivial logic already automated by library management systems, though physical book condition inspection still requires a human or vision-based check.the combined workflow is largely automatable today. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few regulatory or legal barriers exist; libraries operate the systems and set policies. Minor friction may come from patron preference for human interaction and organizational inertia, but nothing legally or structurally prevents automation of the core inspection and fine computation. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing, liability, or regulatory barrier prevents automated fine computation; libraries already widely use self-service kiosks and automated systems. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | The task is routine and rule-driven. OCR, barcode scanning, and fine calculation are cheap; even with integration and occasional human oversight, the all-in cost per book processed is substantially below the loaded wage of a library assistant performing the same work manually. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated fine calculation via existing ILS software costs a tiny fraction of clerical labor time per transaction, already deployed at negligible marginal cost. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Products exist for barcode/date reading and rule-based fine calculation in library management systems, and some libraries deploy computer vision for damage detection in pilots. However, reliable production-grade systems that handle full condition assessment variation across all book types remain limited; most deployments still require human verification of borderline cases. |
| Technical feasibility today | claude-sonnet-5 | 5/5 | Integrated library systems (e.g., Koha, SirsiDynix) already automate due-date tracking and fine calculation reliably at scale in production; self-checkout kiosks handle this routinely. |
Prepare, store, and retrieve classification and catalog information, lecture notes, or other information related to stored documents, using computers.
74CI 65–84 · exposure 75 · augmentation 75 · importance 3.6/5 · click for rater detail
Prepare, store, and retrieve classification and catalog information, lecture notes, or other information related to stored documents, using computers.
74| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Libraries and educational institutions have rapidly adopted digital catalog systems, metadata automation, and AI-assisted retrieval over the past decade. Modern library systems are highly digitized and already embed significant automation; further adoption velocity is high. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Libraries, especially public and academic ones, are historically slow technology adopters with limited budgets, so AI-driven cataloging tools are in pilot stages rather than widespread production use. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI powerfully augments clerical work by auto-suggesting classifications, accelerating metadata entry, improving search relevance, and flagging inconsistencies. Librarians and assistants using AI-assisted systems see substantial productivity gains while retaining final judgment over catalog quality. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools can significantly speed up search, indexing, and metadata tagging tasks for library assistants, serving as strong productivity aids even where full automation isn't yet trusted. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Current AI systems can reliably prepare, store, and retrieve structured classification, catalog metadata, and document information with strong time savings. Database management, metadata tagging, full-text search indexing, and retrieval automation are mature capabilities that exceed the 50% productivity threshold for most library catalog operations. |
| Task automatability | claude-sonnet-5 | 4/5 | Cataloging, metadata tagging, and retrieval of documents are largely digital, structured tasks that current AI (OCR, classification models, database queries, LLM-assisted metadata generation) can handle with substantial time savings, though some edge cases still need human verification. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few hard regulatory barriers exist for automating catalog and retrieval operations, though some organizations prefer human oversight of metadata standards and legacy system integration adds organizational friction. No licensing requirement prevents AI substitution. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement blocks automation, but some libraries maintain preferences for human-verified cataloging accuracy and standardized taxonomies (e.g., MARC, Dewey) that require oversight. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | AI-driven library systems, cloud storage, and automated indexing cost a small fraction of the equivalent full-time clerical labor. Once deployed, per-transaction costs of retrieval and storage are orders of magnitude cheaper than human labor. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated metadata extraction and database management software cost far less per record than clerical labor once implemented, though initial integration with legacy library systems adds some cost. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Enterprise library management systems (ILS platforms like Alma, Koha) and AI-enhanced document retrieval tools are deployed at scale in libraries and knowledge organizations today. Systems reliably handle classification, cataloging, storage metadata, and retrieval at production quality across thousands of institutions. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Integrated library systems already offer automated cataloging suggestions and search/retrieval tools, but full end-to-end AI-driven cataloging without human review is not yet standard in most library deployments. |
Process new materials including books, audio-visual materials, and computer software.
72CI 52–92 · exposure 70 · augmentation 75 · importance 3.9/5 · click for rater detail
Process new materials including books, audio-visual materials, and computer software.
72| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Many public and academic libraries have already deployed barcode scanning, RFID, and automated sorting systems. Adoption is rapid in well-funded institutions and has become industry standard practice over the past 10–15 years. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Libraries are typically under-resourced, slower-adopting public/nonprofit institutions with modest IT budgets, so AI-driven automation of physical materials processing remains limited in practice. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI systems augment library staff by automating the tedious aspects (scanning, cataloging lookup, physical sorting) while humans focus on exception handling, condition assessment, and curation decisions. This significantly raises productivity for the human-in-the-loop. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-assisted cataloging tools, automated metadata retrieval, and barcode/RFID systems meaningfully speed up the administrative portion of processing while staff still handle physical materials. |
| Task automatability | claude-haiku-4-5-20251001 | 5/5 | Processing new library materials involves routine, repetitive steps: scanning barcodes, data entry, physical organization, and labeling. Current AI systems can handle end-to-end workflows from image recognition of items, barcode scanning, metadata lookup, and automated cataloging/shelving assignments, saving well over 50% of the human time required. |
| Task automatability | claude-sonnet-5 | 3/5 | Cataloging metadata lookup, barcode/tagging, and record entry can be substantially automated, but physical handling (unpacking, stamping, shelving prep) and quality checks still require human labor. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Library automation is already widespread and faces minimal regulatory or licensing barriers; the task is clerical rather than requiring human authorization. Some organizational inertia and staff preference for human handling of rare/valuable materials exist, but they are not hard barriers. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing or legal requirement mandates a human perform this clerical processing task; it's routine administrative work. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | AI-based scanning, RFID, and automated cataloging infrastructure have low per-item marginal costs after setup, typically a fraction of the wage cost of a clerical assistant performing these repetitive, manual tasks. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Software for metadata/cataloging is cheap relative to labor, but physical processing (unboxing, labeling, covering) still requires paid staff time, keeping overall cost roughly comparable. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Deployed products exist for library automation, including barcode scanning systems, RFID tagging, and integrated library management software with automated cataloging. While mature in core functions, some edge cases (unusual formats, damaged materials) still require human judgment, keeping it slightly below perfect reliability at scale. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Integrated library systems (ILS) with barcode scanning, MARC record import, and vendor-supplied cataloging data are widely deployed, but full processing still involves manual physical steps not handled by AI products. |
Prepare library statistics reports.
71CI 65–76 · exposure 70 · augmentation 75 · importance 3.4/5 · click for rater detail
Prepare library statistics reports.
71| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Libraries are moderately digitized public/academic institutions with slower IT adoption than private sector, but many have already implemented automated reporting via their ILS (Integrated Library Systems) and BI tools. Adoption is steady but not rapid, typical of mid-tier institutional sectors. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Libraries, especially public/academic ones, are generally slow technology adopters with limited IT budgets, so uptake of automated reporting tools remains uneven. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI augmentation is strong here: AI can draft reports, flag anomalies, and auto-generate visualizations, allowing a librarian or assistant to focus on interpreting results and making decisions. This significantly multiplies human productivity without removing the person from the workflow. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI and analytics tools significantly speed up data aggregation, visualization, and drafting of narrative summaries, letting staff focus on interpretation and decision-making. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Preparing library statistics reports is largely data aggregation, transformation, and formatting—tasks where modern AI and automation excel. Current systems can extract data from library management databases, calculate metrics, and generate structured reports with minimal human oversight, easily meeting the 50% time-saving threshold. |
| Task automatability | claude-sonnet-5 | 4/5 | Compiling counts, formatting, and generating summary reports from library data (circulation, usage logs) is a structured data task that current AI/BI tools can largely automate given database access.atlantic |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Library statistics reporting is not heavily regulated or legally mandated to be performed by humans, and most libraries have weak organizational resistance to automation of internal reporting. The main friction is institutional inertia and preference for human review, not hard legal or liability barriers. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing or legal requirement mandates human preparation of statistics; main friction is organizational habit and data system compatibility. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Automated statistical reporting via cloud-based BI tools or scripts costs pennies per report, while a human clerical worker requires loaded wages of $25–40/hour for report preparation. AI-driven solutions are orders of magnitude cheaper once initial setup is amortized. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated report generation via existing software or scripts is far cheaper than manual compilation once set up, though some initial integration cost exists. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Mature products and off-the-shelf tools (data pipeline automation, BI platforms, LLM-based report generators) demonstrably perform statistical aggregation and report generation in production environments. Libraries increasingly use automated reporting systems, though some customization and validation still require human review. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Library management systems already have built-in reporting/analytics modules, and general spreadsheet/AI tools can generate statistics, but full end-to-end automation without staff configuration and validation is not yet standard practice in most libraries. |
Perform accounting and bookkeeping activities, such as invoicing, maintaining financial records, budgeting, and handling cash.
69CI 52–86 · exposure 67 · augmentation 75 · importance 3.7/5 · click for rater detail
Perform accounting and bookkeeping activities, such as invoicing, maintaining financial records, budgeting, and handling cash.
69| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 5/5 | Accounting and bookkeeping automation is widespread and deeply adopted across sectors, including libraries and cultural institutions. Cloud-based accounting platforms and integrated library management systems with accounting modules are standard in many library systems. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Libraries and educational institutions are typically slow adopters of AI tools relative to finance or tech sectors, with limited budget for specialized automation. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI significantly assists humans in this domain by automating routine invoice and ledger entry, flagging discrepancies, and auto-reconciling accounts, while a human reviews and approves. This transforms productivity on bookkeeping tasks while maintaining human oversight. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-powered accounting tools can meaningfully speed up invoicing, categorization, and record maintenance, letting the clerical worker focus on cash handling and exceptions. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Invoicing, financial record maintenance, budgeting, and cash handling are highly structured, rule-based processes well-suited to automation. Current AI systems and accounting software can perform end-to-end invoice generation, ledger entry, budget reconciliation, and cash reconciliation with minimal human intervention, easily meeting the 50% time-saving threshold. |
| Task automatability | claude-sonnet-5 | 3/5 | Bookkeeping tasks like invoicing and record-keeping are highly structured and partially automatable with existing accounting software, but cash handling and reconciliation still require human execution and physical presence.4 |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few legal barriers exist for automating accounting in libraries; no licensed accountant signature is universally required for library assistant-level bookkeeping. Minor friction comes from internal controls, audit trails, and the need for human oversight of reconciliation, but these do not prevent substitution. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement for library bookkeeping, but institutional financial controls, audit requirements, and trust/verification norms around cash handling add some friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Accounting automation via SaaS platforms costs fractions of a clerical wage per transaction and per month. AI-driven bookkeeping is an order of magnitude cheaper than hiring library assistants for these tasks once the system is operational. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Software licensing costs are low relative to a clerical wage, but integration, oversight, and the physical cash-handling component keep overall savings moderate rather than dramatic. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Mature accounting software (QuickBooks, NetSuite, Xero) and AI-augmented bookkeeping tools have long been deployed in production at scale. These systems reliably handle invoicing, record-keeping, and basic budgeting across millions of organizations with high accuracy. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Accounting software (QuickBooks, etc.) with AI-assisted categorization and invoicing is widely deployed, but cash handling and error correction still rely on human oversight in library settings. |
Answer routine inquiries and refer patrons in need of professional assistance to librarians.
66CI 56–76 · exposure 62 · augmentation 75 · importance 4.3/5 · click for rater detail
Answer routine inquiries and refer patrons in need of professional assistance to librarians.
66| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Library systems have been slower to adopt AI compared to commercial customer service, but chatbot and virtual reference desk pilots are now common in public and academic libraries; production rollout is increasing but not yet ubiquitous. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Libraries are typically under-resourced, public-sector institutions with slower technology adoption compared to fast-moving private sectors like finance or tech. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI assistants can substantially boost library clerical staff productivity by pre-screening questions, drafting responses to common inquiries, and flagging patterns, allowing staff to focus on complex patron needs and librarian coordination. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI chat tools and knowledge bases can meaningfully speed up answering routine questions and help staff quickly identify when escalation to a librarian is needed. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Most of this task involves pattern matching (recognizing routine inquiries) and applying simple rules (referring to librarians when needed). Current AI chatbots and rule-based systems can handle the majority of routine patron questions autonomously, with referral logic easily automated, achieving >50% time savings while maintaining quality on standard queries. |
| Task automatability | claude-sonnet-5 | 3/5 | AI chatbots can handle routine FAQ-style patron inquiries (hours, policies, location of materials) and triage to human librarians, but real-world library service involves varied in-person interactions and multitasking that current systems can't fully replace. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Libraries have some organizational friction around customer preference for human interaction and desire to maintain reference services with human judgment, but there are no legal or licensing barriers to automating routine inquiry response and referral routing. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement for this clerical task, though patron preference for human interaction and institutional norms in public-facing service roles create mild friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | An AI chatbot inference cost per interaction is typically pennies, while a clerical library assistant's loaded wage is $25–35/hour. AI is at least two orders of magnitude cheaper per routine interaction. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | A chatbot or simple AI assistant handling routine questions costs far less per interaction than staffing a clerical position, though integration and maintenance add some cost. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Deployed library chatbot systems (e.g., LibraryBot, Ask a Librarian AI) already perform routine inquiry answering in production at many institutions. These systems reliably handle FAQs, account issues, and basic reference questions, though they still require human oversight for escalation decisions. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Library chatbots and virtual reference services exist and are deployed in some systems, but coverage is uneven and many libraries still rely on staff for even basic desk inquiries. |
Register new patrons and issue borrower identification cards that permit patrons to borrow books and other materials.
66CI 60–72 · exposure 70 · augmentation 63 · importance 4.0/5 · click for rater detail
Register new patrons and issue borrower identification cards that permit patrons to borrow books and other materials.
66| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Adoption is uneven; large urban and well-resourced library systems increasingly deploy self-service registration, while smaller and underfunded branches lag. Pilots and partial implementations are common, but library sectors lag behind information-sector norms in digitization velocity. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Public libraries are typically underfunded, slow-moving public institutions with uneven digitization, so despite the technical feasibility, actual adoption of automated registration remains patchy. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI-assisted tools (auto-fill from document OCR, fraud-risk flagging, real-time address validation) significantly boost clerk productivity and accuracy, allowing faster processing and fewer errors while the librarian remains in control of exceptions and approvals. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI/software tools help pre-fill forms, verify ID data, and manage patron databases, easing clerical burden even where humans still handle exceptions or in-person interactions. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | The core components—data entry, form completion, identity verification via documents, and card generation—are highly automatable. A system could capture patron information, validate against databases, and produce physical or digital ID cards with minimal human intervention, easily meeting the 50% time-saving threshold. |
| Task automatability | claude-sonnet-5 | 4/5 | Self-service kiosks and online registration portals already handle most of patron intake and card issuance with minimal human intervention, meeting the time-saving bar for the bulk of the workflow. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Some friction exists: libraries may prefer human interaction for patron service, many patron-facing systems still require staff sign-off for fraud or policy exceptions, and organizational inertia slows digital adoption in some systems. However, no hard legal or licensing barrier mandates human registration. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement for this clerical task, but some libraries impose policies requiring photo ID verification or in-person visits for security/privacy, creating mild friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | Automated registration systems (kiosks, integrated library software, identity APIs) cost far less per transaction than staff time, especially at volume. Once deployed, marginal cost per patron is negligible compared to a clerk's loaded hourly wage. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated registration via kiosks or web forms costs far less per transaction than staff time, though initial software/ILS integration and occasional ID verification still require some investment. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Self-service kiosks and library management systems partially automate registration in some modern libraries, but full end-to-end automation faces challenges: identity verification still often requires human judgment, fraud detection is imperfect, and many systems require staff oversight. Products exist but are not universally reliable or deployed at scale. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Many library systems already deploy self-checkout/self-registration kiosks and online account creation tied to ILS systems, functioning reliably in production, though some libraries still require in-person verification. |
Perform clerical activities, such as answering phones, sorting mail, filing, typing, word processing, and photocopying and mailing out material.
63CI 61–65 · exposure 66 · augmentation 63 · importance 4.0/5 · click for rater detail
Perform clerical activities, such as answering phones, sorting mail, filing, typing, word processing, and photocopying and mailing out material.
63| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Libraries and educational institutions are typically lower-digitization, slower-adoption sectors. While some large systems experiment with automation, clerical automation in libraries is not yet widespread in production. Adoption remains at pilot stage in most public and academic library settings. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Libraries and clerical support functions are generally slow adopters of AI compared to sectors like finance or tech, with budget constraints and legacy systems slowing deployment. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist library assistants by auto-sorting and routing mail, suggesting file locations, auto-generating formatted documents, and drafting phone responses. These augmentations improve efficiency on parts of the task, though they do not fundamentally transform overall productivity the way specialized tools might. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools (auto-transcription, smart filing, templated correspondence, scheduling assistants) meaningfully speed up clerical work even where full automation of the physical tasks isn't yet achieved. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Most of these clerical subtasks—sorting mail, filing, photocopying, word processing, and mailing—can be substantially automated with current systems (document processing, OCR, workflow automation). Answering phones is the main exception, though basic call routing and FAQ-answering chatbots handle parts of it. End-to-end automation of the composite task would likely achieve >50% time savings with AI-assisted workflow tools. |
| Task automatability | claude-sonnet-5 | 4/5 | Most listed sub-tasks (typing, word processing, filing digital records, drafting mail responses) are highly automatable with current AI and office automation tools, though physical actions like photocopying and sorting physical mail still need a human or dedicated hardware. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Library clerical work has minimal regulatory barriers; no license is legally required. The main friction is organizational (legacy systems, staff preferences, modest capital to invest in RPA) and customer comfort with reduced human contact, but these are soft barriers that do not prevent substitution. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing or legal barriers to automating these tasks, though some institutional inertia and preference for human interaction (e.g., patrons calling in) create minor friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | RPA, OCR, and document-processing cloud services (AWS Textract, UiPath, etc.) operate at pennies to low dollars per task, far below the loaded wage of a library assistant ($30–40k annually, ~$15–20/hour). Automation cost per unit of work is substantially cheaper, especially for repetitive subtasks like filing and sorting. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Software automation (word processing, digital filing, auto-responses) is cheap, but physical tasks (photocopying, sorting mail, mailing) still require human labor or specialized equipment, keeping overall cost roughly comparable to hiring clerical staff for the full bundle. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Document management, OCR, and email/mail sorting systems exist in production (e.g., RPA solutions, document automation platforms), but they often require significant setup and handle edge cases with material error rates. No single deployed product reliably covers the entire task end-to-end; integration across heterogeneous inputs (phone calls, physical mail, digital documents) remains patchy. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Products like email/phone triage bots, document management systems, and OCR-based filing exist and are deployed, but a fully integrated clerical assistant handling all these physical and digital tasks reliably is not yet standard in most libraries. |
Acquire books, pamphlets, periodicals, audio-visual materials, and other library supplies by checking prices, figuring costs, and preparing appropriate order forms and facilitating the ordering process by providing such information to others.
59CI 43–76 · exposure 58 · augmentation 75 · importance 3.5/5 · click for rater detail
Acquire books, pamphlets, periodicals, audio-visual materials, and other library supplies by checking prices, figuring costs, and preparing appropriate order forms and facilitating the ordering process by providing such information to others.
59| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Libraries and educational institutions are adopting e-procurement and automation tools, but adoption remains fragmented. Many smaller libraries still use legacy systems or manual processes; large systems are faster to adopt. Overall adoption is in the pilot-to-early-production phase rather than deep, widespread deployment. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Libraries are a slow-adopting, often under-resourced sector with limited AI deployment in back-office procurement functions. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can assist clerical staff by automatically drafting orders, flagging price anomalies, and suggesting suppliers, significantly raising the productivity of a human librarian who retains control over vendor relationships and institutional policy compliance. The human remains in the loop while AI handles routine data processing. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools can meaningfully speed up price comparisons, cost calculations, and form drafting, letting staff focus on vendor negotiation and collection decisions. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Checking prices, calculating costs, and preparing order forms are highly structured, data-driven activities that current AI systems can perform reliably. The only moderately complex part is interfacing with diverse supplier systems, but most institutional ordering today uses APIs or standard forms that AI agents can navigate, allowing >50% time savings at equal quality. |
| Task automatability | claude-sonnet-5 | 3/5 | Price checking, cost calculation, and order form preparation are structured data tasks AI can largely handle, but vendor selection, budget judgment, and system integration still require human oversight. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Most libraries have institutional procurement policies and may require human sign-off on orders or supplier relationships, but there are no legal licensing barriers to automating the clerical work itself. Organizational friction and preference for human oversight over budgets create moderate friction without hard legal blocks. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement, but institutional purchasing rules, vendor relationships, and budget approval chains create moderate organizational friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Once integrated, the marginal cost of AI-driven procurement (API calls, inference, minimal human oversight) is orders of magnitude cheaper than paying a clerical worker's loaded wage ($35k–50k annually) to manually check prices and fill forms repeatedly. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Automating price lookups and form generation is cheap, but integrating with existing library ILS/acquisition systems and human verification adds cost comparable to a clerical wage. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Procurement and supply-chain automation tools are deployed in production across libraries and institutions. Systems can parse supplier catalogs, compare prices, auto-populate order forms, and flag cost discrepancies. While some edge cases and non-standard suppliers may require human review, the core task is demonstrably handled by mature products in real organizations. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | While procurement software and AI-assisted purchasing tools exist broadly, library-specific acquisition workflows are typically still manual or semi-automated with library staff verifying details. |
Classify and catalog items according to content and purpose.
54CI 52–55 · exposure 50 · augmentation 75 · importance 4.3/5 · click for rater detail
Classify and catalog items according to content and purpose.
54| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Libraries are gradually adopting automated cataloging and AI-assisted classification tools, but adoption remains uneven across public, academic, and special libraries; widespread production deployment exists but has not yet dominated the sector. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Libraries and archival institutions are generally slower adopters of AI compared to fast-moving information/finance sectors, with budget constraints and legacy systems limiting deployment. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI dramatically improves librarian and assistant productivity by pre-classifying items, suggesting metadata, and flagging potential cross-references, allowing humans to focus on complex judgments and quality assurance rather than routine data entry and classification lookup. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can meaningfully speed up cataloging by suggesting subject headings, classifications, and metadata, letting clerical staff review and finalize rather than starting from scratch. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI can automate significant portions of cataloging through optical character recognition, metadata extraction, and classification against existing taxonomies (Dewey Decimal, Library of Congress), but requires human judgment for ambiguous items, cross-references, and institutional-specific rules, yielding roughly 50% time savings with setup. |
| Task automatability | claude-sonnet-5 | 3/5 | AI can suggest classifications and metadata using content analysis, but final cataloging often needs verification against specific classification schemes (Dewey, LC) and institutional standards, requiring human oversight for accuracy. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Libraries have no legal requirement for human catalogers to perform this task, and many use automated systems already; however, institutional preference for human curation and concern over classification errors create modest friction to full automation. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement for this clerical task, though many libraries have institutional standards and quality control processes that create some friction against pure automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 3/5 | AI-assisted cataloging tools are comparable in cost to human labor when accounting for software licensing, integration, and necessary human oversight; neither dramatically undercuts the other on a per-item basis. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI tools can speed up initial classification suggestions cheaply, but the need for human verification and correction of errors keeps overall costs closer to parity with traditional clerical cataloging workflows. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Library management systems with automated classification modules exist and are deployed in production, but they still require human review and correction; error rates on complex or rare items remain material, limiting full autonomous end-to-end performance. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Library systems increasingly use AI-assisted cataloging tools and metadata generation, but these are typically semi-automated with librarian review rather than fully autonomous production cataloging. |
Design or maintain library web site and online catalogues.
51CI 49–52 · exposure 50 · augmentation 75 · importance 3.6/5 · click for rater detail
Design or maintain library web site and online catalogues.
51| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Library systems are typically part of lower-digitization, budget-constrained public/nonprofit sectors with slow adoption of cutting-edge automation; pilots exist but widespread production deployment of AI-driven catalog and web maintenance remains limited. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Libraries, especially public and small institutional ones, are generally slower technology adopters compared to fast-moving sectors like finance or tech, with limited AI-driven catalog system deployment reported. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI tools (search optimization, metadata suggestion, content recommendation engines, automated quality checks) meaningfully assist library assistants in catalog curation, website updates, and user experience improvement while keeping humans in control of strategic and design decisions. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI coding assistants and design tools substantially speed up website creation and updates, and can help with metadata tagging and content generation for catalogs, meaningfully boosting productivity. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI can automate significant portions of catalog maintenance (metadata standardization, duplicate detection) and basic website updates (content management, link validation), but design decisions, information architecture, and user experience refinement still require human judgment; likely 40–50% time savings achievable with AI assistance tools. |
| Task automatability | claude-sonnet-5 | 3/5 | AI can generate website code, templates, and assist with catalog metadata, but ongoing maintenance, integration with library systems (ILS), and design decisions still require human oversight and customization.web design and catalog work is partially but not fully automatable end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Library web and catalog systems have some institutional inertia and require IT governance approval, but there are no hard legal barriers or licensing requirements preventing AI automation; user preference for human library staff and organizational friction exist but are moderate. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement for this task, though library systems may have institutional IT policies or vendor lock-in for catalog software that create some friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Current AI/SaaS tooling for library systems and web maintenance is not dramatically cheaper than hiring a library assistant; integration overhead, customization, and ongoing supervision reduce cost advantage to modest at best. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI tools can cut initial design/coding time significantly, but ongoing catalog maintenance requires specialized integration work and human review, keeping costs roughly comparable to a clerical assistant's time for the full task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Deployed CMS platforms and AI-assisted tools exist for website maintenance and catalog management, but they typically require human oversight for quality assurance and design changes; mature production systems handle routine tasks reliably but struggle with novel design or complex catalog restructuring. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | AI website builders and coding assistants are deployed in production for general web design, but library-specific catalog systems (ILS/OPAC integration) have narrower, less mature AI tooling. |
Manage reserve materials by placing items on reserve for library patrons, checking items in and out of library, and removing out-of-date items.
45CI 35–55 · exposure 42 · augmentation 50 · importance 4.2/5 · click for rater detail
Manage reserve materials by placing items on reserve for library patrons, checking items in and out of library, and removing out-of-date items.
45| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Libraries are digitization laggards relative to finance or tech sectors. Many still rely on legacy systems; adoption of advanced AI-driven reserve and collection management is in pilot stages at best, with most libraries maintaining human-centric workflows. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Libraries have moderately adopted self-checkout and automated systems over the past decade, but library sector overall is a slower adopter of newer AI/agentic tools compared to finance or tech. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist by suggesting items for removal (via usage analytics), flagging checkout anomalies, and automating routine barcode scanning, but the human library assistant remains central to curation, patron service, and judgment calls. Productivity gains are moderate rather than transformative. |
| Augmentation potential | claude-sonnet-5 | 3/5 | Automated systems and software assist staff in tracking reserves and due dates, improving efficiency, though physical tasks remain manual. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While checking items in/out could be partially automated via barcode scanning and inventory systems, the judgment-required aspects (determining what is out-of-date, deciding placement on reserves, handling exceptions) require human discretion. Current AI cannot reliably perform the full end-to-end task with 50% time savings at equal quality. |
| Task automatability | claude-sonnet-5 | 3/5 | Check-in/check-out and reserve tracking are already largely automated via ILS/RFID self-checkout systems, but physically shelving, retrieving, and removing outdated items still requires human handling. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Library policies, patron access requirements, and institutional workflows create moderate friction. No hard legal barrier exists, but organizational inertia, staff training, and patron preference for human interaction provide meaningful resistance to full automation. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing or legal requirement mandates a human for these clerical tasks, though some patron preference for in-person help and physical handling needs create mild friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Automated checkout systems exist but require significant upfront infrastructure and ongoing integration costs. The full task bundle (reserve selection, condition assessment, removal decisions) still requires substantial human oversight, keeping total AI cost-per-output similar to or exceeding loaded human wage. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Self-checkout and automated circulation systems reduce labor costs substantially for the transactional part, but physical material handling still requires paid staff, keeping overall cost roughly comparable. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Library management systems and RFID/barcode checkout automation exist in production, but these are specialized library-domain tools rather than general AI. Fully autonomous reserve management with exception handling and curation decisions is not deployed reliably at scale. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Integrated library systems and self-service kiosks reliably handle circulation transactions today, but the physical placement/removal of materials and exception handling still depend on staff. |
Lend, reserve, and collect books, periodicals, videotapes, and other materials at circulation desks and process materials for inter-library loans.
44CI 35–52 · exposure 42 · augmentation 63 · importance 4.1/5 · click for rater detail
Lend, reserve, and collect books, periodicals, videotapes, and other materials at circulation desks and process materials for inter-library loans.
44| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Library adoption of circulation automation is moderate; many branches still rely heavily on staff-mediated lending and inter-library loan coordination, and budget constraints limit investment in advanced systems. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Public/academic libraries are typically slow-adopting, resource-constrained institutions where self-checkout technology has spread unevenly and clerical staff remain common for exceptions and physical processing. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI-assisted catalog search, automated due-date management, and smart routing of inter-library loans can significantly boost staff productivity while they remain in the loop for exceptions, patron interaction, and material condition assessment. |
| Augmentation potential | claude-sonnet-5 | 3/5 | Automated systems and kiosks meaningfully speed up transactional lending/reserving, letting clerical staff focus on exceptions and inter-library loan coordination, though physical tasks remain manual. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Physical handling of materials (picking, lending, collecting) cannot be automated without robotic systems not broadly deployed in libraries. Inventory tracking and reservation systems can be partially automated, but the complete task of physically processing materials at circulation desks falls well short of 50% time savings today. |
| Task automatability | claude-sonnet-5 | 3/5 | Check-in/check-out and reservation processes are largely automated via self-service kiosks and ILS systems, but physical handling of materials and inter-library loan processing require human presence and physical action. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Libraries have institutional and community preferences for human staff presence at circulation desks, and inter-library loan processing often requires authorization and verification steps; however, no strict legal barrier prevents automation of the clerical portion. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement, but physical material handling, shelving, and interlibrary loan logistics create practical friction that pure software cannot eliminate. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Existing library systems reduce labor but don't eliminate it; the integrated cost of circulation software, self-checkout hardware, maintenance, and required human oversight remains comparable to or exceeds the loaded wage of a library assistant for this task. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Self-service kiosks and automated systems reduce staffing needs but require capital investment in RFID/scanning hardware and ongoing maintenance, making costs comparable rather than dramatically cheaper than clerical wages. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Library management systems and self-checkout kiosks handle portions of lending and returns, and barcode scanning is standard; however, systems cannot handle damaged materials assessment, patron problem-solving, or inter-library loan exceptions reliably without human oversight. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Self-checkout kiosks, RFID systems, and automated hold-shelf notifications are widely deployed in libraries today, but physical circulation desks still require staff for exceptions, ILL packaging/shipping, and patron assistance. |
Instruct patrons on how to use reference sources, card catalogs, and automated information systems.
41CI 34–47 · exposure 33 · augmentation 63 · importance 4.1/5 · click for rater detail
Instruct patrons on how to use reference sources, card catalogs, and automated information systems.
41| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Library systems are traditionally conservative on automation and budget-constrained; while some libraries have deployed chatbots, production adoption remains sparse and pilots are common rather than deep rollouts. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Libraries and public-sector information services are generally slow adopters of AI tools compared to finance or tech sectors. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI-powered search suggestions, auto-generated FAQs, and decision trees can assist librarians in fielding routine inquiries and routing patrons, improving throughput on straightforward reference tasks while humans handle complex cases. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI chat tools and search assistants can help staff quickly find answers or draft instructions, improving efficiency in explaining systems to patrons. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI could handle scripted reference instructions via chatbots, the task requires responsive teaching adapted to diverse patron needs, questions, and technical proficiency levels. Current systems struggle with the dynamic pedagogical adjustment and troubleshooting that constitutes most of this work. |
| Task automatability | claude-sonnet-5 | 3/5 | Chatbots and AI search assistants can explain how to use catalogs and databases, but in-person, adaptive instruction with follow-up demonstration is harder to fully replicate end-to-end.dmin |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Library services often carry institutional and user-experience expectations favoring human interaction; regulatory barriers are low, but organizational inertia and patron preference for human help create moderate friction to full substitution. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement, but patron preference for human help and library service norms create mild friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 3/5 | Chatbot hosting and maintenance are modest, but staff oversight and failure recovery add cost. The all-in AI cost approaches human wage for the same instructional output due to integration, monitoring, and human fallback needs. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | A chatbot answering common questions is cheap, but staffing costs for library assistants are already low, so savings are moderate rather than dramatic. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Deployed chatbots exist for library FAQs and basic system instructions, but they perform narrowly and fail on out-of-scope questions or complex patron confusion. Production systems are thin and supplementary, not reliable replacements for live instruction. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some libraries deploy chatbots or AI FAQ systems for basic guidance, but reliable in-person instructional support at scale is not yet standard practice. |
Locate library materials for patrons, including books, periodicals, tape cassettes, Braille volumes, and pictures.
36CI 28–44 · exposure 33 · augmentation 63 · importance 4.3/5 · click for rater detail
Locate library materials for patrons, including books, periodicals, tape cassettes, Braille volumes, and pictures.
36| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Library systems are typically under-resourced and slow to digitize or automate; most libraries rely on manual shelving and retrieval. Adoption of retrieval automation is minimal outside a small number of well-funded research institutions. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Libraries are a low-digitization, resource-constrained sector with slow technology adoption relative to finance or professional services. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI-powered catalog search and recommendation systems can assist staff in locating materials faster and suggesting related items to patrons, meaningfully improving the assistant's ability to serve multiple patrons and reduce search time. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-enhanced catalog search, recommendation systems, and chatbots can meaningfully speed up locating relevant materials and answering patron queries, while staff still perform physical retrieval. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Locating materials by catalog number in a well-organized physical space requires navigation and retrieval of objects in the real world. While AI could search a digital catalog end-to-end, the physical retrieval step—finding and retrieving specific items from shelves—remains difficult for current systems without specialized robotics, which are not yet deployed at scale in libraries. |
| Task automatability | claude-sonnet-5 | 3/5 | Digital catalog lookups and location guidance can be partly automated via search interfaces, but physical retrieval of items from shelves (especially specialized formats like Braille or cassettes) still requires human action. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Public libraries and institutional settings often prefer human staff presence for patron interaction and wayfinding support. Physical library layouts and mixed-format collections create organizational friction around full automation, though no hard legal barrier exists. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement, but physical handling of materials, accessibility needs (e.g., Braille), and patron interaction create practical friction against pure automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Implementing fully automated retrieval (robotics, shelf-scanning systems, integration) is capital-intensive and ongoing operational cost, often exceeding the wage of a library assistant who performs the task manually. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI catalog search is cheap, but physical retrieval still requires paid staff time or robotics infrastructure that is costly relative to a low-wage clerical worker. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Library management systems can search digital catalogs reliably, but no mainstream product performs full end-to-end retrieval including physical shelf navigation and item retrieval at production scale. Some libraries experiment with RFID and robotic retrieval, but these remain niche deployments. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Library catalog search systems and some robotic retrieval pilots exist, but no widely deployed product physically locates and hands over diverse physical materials to patrons reliably. |
Select substitute titles when requested materials are unavailable, following criteria such as age, education, and interests.
34CI 30–39 · exposure 25 · augmentation 63 · importance 3.4/5 · click for rater detail
Select substitute titles when requested materials are unavailable, following criteria such as age, education, and interests.
34| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Public and academic libraries are relatively low-digitization sectors with slower technology adoption; most library automation has focused on cataloging and checkout, not recommendation. Few libraries have deployed AI-driven substitute selection in production. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Libraries are a low-digitization, resource-constrained sector with slow AI adoption for reference-type tasks, mostly still using catalog software over AI-driven recommendations. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can usefully assist librarians by surfacing candidate substitutes based on metadata and circulation patterns, allowing staff to review and refine recommendations rather than browsing manually. This augmentation is meaningful but incremental, not transformative. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI recommendation systems and chatbots can meaningfully assist clerical staff in brainstorming substitute titles or checking metadata quickly, improving speed while the human finalizes the choice. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can parse requests and match metadata, recommending substitutes requires understanding patron context (age, education, interests) and nuanced judgment about literary equivalence—factors that often demand human familiarity with collections and individual patrons. Current systems lack reliable access to patron-specific information and the reasoning depth needed for consistent quality. |
| Task automatability | claude-sonnet-5 | 2/5 | Recommending substitute titles requires contextual judgment about patron preferences and nuanced matching that current AI can partially support but not reliably execute end-to-end without human oversight in most library settings.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Library staff often work in unionized or publicly regulated environments where human mediation of patron requests is valued; some patrons may prefer human recommendation. No hard licensing barrier exists, but organizational norms and customer preference for human interaction create moderate friction. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement exists, but organizational preference for personalized human interaction and the need for local collection knowledge create some friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Integration costs (API setup, library system connection, oversight workflows) plus inference overhead likely exceed or match the wage cost of a library assistant performing this task, especially given the need for accuracy and patron-specific customization. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI-based recommendation tools are cheap to run, but integration and the need for human verification to match patron-specific criteria keeps costs roughly comparable to a clerical worker's time for this narrow task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed library system reliably automates substitute selection end-to-end; most recommendations engines are narrow (e.g., content-based filtering for retail). Library environments require institutional integration, patron verification, and human sign-off, making fully autonomous deployment rare in production. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some recommendation engines exist (e.g., library catalog suggestion tools), but few production systems handle nuanced substitute-title selection based on age, education, and interest criteria reliably. |
Assist in the preparation of book displays.
34CI 19–49 · exposure 28 · augmentation 63 · importance 3.1/5 · click for rater detail
Assist in the preparation of book displays.
34| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Libraries remain relatively low-digitization, physically-grounded environments with limited track records of robotic or autonomous automation for display tasks. Adoption of AI for this specific work is minimal even among early-adopter institutions. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Libraries are a slow-adopting, under-resourced sector with limited AI integration into physical curatorial workflows. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist library staff by recommending books by theme, generating display descriptions, or suggesting color/layout combinations—useful enhancements that improve productivity without replacing human curatorial judgment and physical execution. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools can meaningfully help brainstorm themes, write promotional blurbs, and create visual mockups, boosting efficiency while humans still execute the physical display. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Preparing book displays requires physical arrangement, aesthetic judgment, and contextual curation that current AI cannot perform end-to-end. While AI could suggest themes or generate display text, the core manual and creative components—physically positioning books, assessing visual balance, responding to space constraints—remain beyond autonomous AI capability. |
| Task automatability | claude-sonnet-5 | 3/5 | AI can suggest themes, generate descriptive copy, and even design visual layouts, but physical arrangement of books and space planning still require human execution.tomatability is partial. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Library institutions typically prefer human-curated displays for their community-facing value and aesthetic judgment. While not legally restricted, organizational preferences and the desire for human creativity and cultural relevance create moderate friction against full automation. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing, legal, or safety barriers restrict AI assistance in creative/curatorial planning tasks like this. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The task involves primarily physical work and local decision-making that requires human presence on-site. AI assistance would require expensive robotics and computer vision integration, making the all-in cost substantially higher than employing a library assistant. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI can cheaply generate ideas or captions, but the bulk of the task—physical setup, book selection, and arrangement—still requires paid human labor, keeping overall cost comparable to human-only work. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed products reliably perform this task autonomously. While AI can assist with recommendation algorithms or text generation for signage, the embodied, spatial, and creative elements of display preparation are not addressed by production systems. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some libraries use AI tools (e.g., design software, generative image tools) for planning displays, but no mature deployed product handles the full physical/curatorial task in production at scale. |
Provide assistance to librarians in the maintenance of collections of books, periodicals, magazines, newspapers, and audio-visual and other materials.
30CI 25–35 · exposure 25 · augmentation 50 · importance 3.8/5 · click for rater detail
Provide assistance to librarians in the maintenance of collections of books, periodicals, magazines, newspapers, and audio-visual and other materials.
30| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Library adoption of AI automation lags significantly behind information-sector leaders; most libraries remain in pilot phases for digital tools, with minimal production deployment of collection-maintenance automation. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Libraries are a low-digitization, often underfunded sector with slow AI adoption, mostly limited to catalog software rather than task automation. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI-assisted cataloging systems and searchable metadata can help assistants locate and organize materials more efficiently, offering useful productivity gains on the clerical and research components of collection maintenance. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-driven cataloging, metadata tagging, and inventory software can meaningfully assist clerical staff in organizing and tracking collections. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI could potentially assist with cataloging and database maintenance, the physical tasks of shelving, handling materials, and condition assessment require human judgment and dexterity. Current AI systems cannot perform the full suite of collection maintenance work end-to-end with 50% time savings. |
| Task automatability | claude-sonnet-5 | 2/5 | Much of collection maintenance involves physical handling, shelving, and organizing materials which AI cannot perform; only sub-tasks like cataloging metadata or tracking could be assisted. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Libraries operate under professional standards and governance that typically require human librarians or trained assistants to maintain collections; institutional inertia, staff union considerations, and the hands-on nature of the work create meaningful adoption friction. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement, but physical presence and handling of materials creates practical barriers to full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Library assistant wages are modest, and the cost of AI infrastructure plus human oversight for quality assurance makes full automation economically marginal for routine clerical collection maintenance tasks. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Physical tasks still require human labor; software tools reduce some administrative overhead but do not replace the bulk of the labor cost. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Deployed products exist for metadata management and cataloging assistance, but no integrated system reliably handles the complete scope of collection maintenance—physical inventory, damage assessment, and organizational decisions—in production library environments at scale. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Library systems use software for inventory and cataloging, but no deployed product autonomously performs physical collection maintenance tasks reliably. |
Operate and maintain audio-visual equipment.
27CI 19–35 · exposure 20 · augmentation 38 · importance 3.7/5 · click for rater detail
Operate and maintain audio-visual equipment.
27| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Library systems are generally slow to adopt advanced automation, remain budget-constrained, and lack the digital infrastructure and ROI incentives seen in high-velocity sectors; A/V automation adoption is minimal and pilot-stage. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Libraries and clerical support roles are low-digitization, low AI-adoption environments, and physical equipment tasks see essentially no automation deployment. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI could assist staff by monitoring equipment health, logging issues, suggesting troubleshooting steps, or automating routine startup sequences, materially reducing manual oversight while humans retain control and decision-making for complex faults. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI chatbots or manuals could help troubleshoot equipment issues or provide instructions, offering minor assistance, but cannot meaningfully transform the physical maintenance work. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Basic operation of audio-visual equipment (playing, pausing, adjusting volume) could be partially automated, but maintenance tasks require physical intervention, troubleshooting expertise, and contextual judgment that current AI cannot reliably perform without human oversight. |
| Task automatability | claude-sonnet-5 | 2/5 | This is a hands-on physical task involving setup, troubleshooting, and maintenance of equipment, which current AI systems cannot physically perform; at most AI can offer troubleshooting guidance via text. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Libraries may have institutional preferences for human staff managing valuable equipment and patron interactions, and some physical maintenance tasks legally or practically require in-person credentialed technicians, creating moderate friction to full substitution. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement, but the physical nature of handling and maintaining equipment creates a practical barrier to any non-physical AI substitution. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | The cost of building, deploying, and maintaining robotic systems or specialized AI for equipment maintenance remains high relative to the modest wage of a library assistant, especially for the mixed operation and maintenance duties involved. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI substitute performing the physical operation/maintenance, so the human remains the only cost-effective option for this task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While some monitoring and simple control tasks exist in deployed systems, real-world library A/V equipment maintenance involves diverse hardware, fault diagnosis, and physical repairs that no production AI system reliably handles today at the scope of this task. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product physically operates or maintains AV equipment; this remains a manual, in-person task requiring physical dexterity and hardware interaction. |
Sort books, publications, and other items according to established procedure and return them to shelves, files, or other designated storage areas.
24CI 19–29 · exposure 20 · augmentation 25 · importance 4.3/5 · click for rater detail
Sort books, publications, and other items according to established procedure and return them to shelves, files, or other designated storage areas.
24| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Library systems are typically budget-constrained, risk-averse institutions with slow technology adoption. Shelving remains almost entirely manual across most public and academic libraries, with minimal real-world displacement by automation. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Libraries are a low-digitization, low-margin sector with minimal investment in robotic automation; adoption of AI for physical shelving tasks is essentially nonexistent. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could marginally assist by sorting call numbers optically or managing inventory databases, but the core manual shelving task benefits little from AI assistance today. Any augmentation is confined to pre-sorting or tracking, not the physical placement itself. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can help with cataloging, sorting logic, or generating shelving lists, but it offers little assistance for the physical act of retrieving and placing items. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While sorting and shelving require significant physical manipulation and spatial reasoning, current robotic systems struggle with the variability of book sizes, fragility considerations, and the need to read call numbers accurately in dim lighting. End-to-end automation with 50% time savings is not yet achievable with deployed systems. |
| Task automatability | claude-sonnet-5 | 2/5 | Sorting and physical shelving requires physical manipulation of items in real-world space, which current AI (software-based) cannot perform; only robotics could address this and that is not off-the-shelf standard practice in libraries.the |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Libraries face moderate adoption friction: they require physical infrastructure redesign, staff retraining, and patron acceptance of automated systems in spaces traditionally staffed by humans. However, there are no strict legal or licensing barriers to automation itself. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing or legal barrier prevents automation, but physical infrastructure, item handling variability, and library layouts create practical organizational friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Current robotic systems capable of this task are extremely expensive to purchase, install, and maintain, while library assistants earn modest wages. The all-in cost of automation significantly exceeds the loaded human wage for this repetitive but low-skill task. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | Physical sorting/shelving requires robotic hardware and navigation systems that are far more expensive per task than a clerical worker's wage, making AI costlier all-in. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Some experimental library automation systems exist (e.g., robotic arms in specialized settings), but reliable, production-scale deployment across general library shelving remains rare. Most libraries still rely on human staff, and existing robots have narrow operating constraints. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product reliably performs physical book sorting and shelving in production libraries today; this remains a manual or occasionally robotic-assisted task in rare pilot settings. |
Maintain library equipment, such as photocopiers, scanners, and computers, and instruct patrons in proper use of such equipment.
23CI 10–35 · exposure 13 · augmentation 38 · importance 3.7/5 · click for rater detail
Maintain library equipment, such as photocopiers, scanners, and computers, and instruct patrons in proper use of such equipment.
23| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Library systems are slow adopters of automation, operating primarily in physical spaces with budget constraints; AI adoption is minimal beyond basic self-service kiosks and digital catalogs. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Libraries are a low-digitization, low-AI-adoption sector, and physical equipment maintenance tasks are not part of any current AI adoption wave. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist by providing automated troubleshooting guides, instructional videos, and patron FAQ systems that reduce the load on staff, though the human assistant remains essential for hands-on help and equipment repair coordination. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI chatbots or documentation tools could provide some troubleshooting guidance or FAQ support, but this offers only marginal assistance to the core physical and instructional task. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI could potentially automate parts of patron instruction (e.g., generating instructional content or basic troubleshooting guides), the physical maintenance of equipment and real-time adaptive instruction to diverse patrons requires human presence and judgment that current systems cannot replicate end-to-end at 50% time savings. |
| Task automatability | claude-sonnet-5 | 1/5 | This task requires physical maintenance of equipment and in-person, hands-on instruction of patrons, neither of which current AI systems can perform end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Library systems typically prefer human staff for patron interaction and equipment handling due to liability concerns and patron preference for direct assistance, though no formal licensing requirement exists for this support role. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement exists, but the physical nature of equipment repair and face-to-face instruction creates practical barriers to any automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | The task requires physical presence on-site and equipment expertise; AI cannot perform physical maintenance, so the cost of human labor plus AI tools for instruction (if deployed) likely still exceeds pure human performance in a low-wage library assistant role. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | AI cannot substitute for the physical labor and troubleshooting involved, so there is no meaningful AI cost comparison—human labor remains the only viable option. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed product reliably performs both equipment maintenance and patron instruction autonomously; instruction systems exist (chatbots, videos) but lack the embodied problem-solving and interpersonal adaptation needed for hands-on equipment support in a library setting. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product performs physical equipment maintenance or in-person patron instruction; this remains a manual, physical-presence task. |
Schedule, supervise, and train clerical workers, volunteers, student assistants, and other library employees.
21CI 11–30 · exposure 13 · augmentation 50 · importance 3.7/5 · click for rater detail
Schedule, supervise, and train clerical workers, volunteers, student assistants, and other library employees.
21| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Libraries are typically small, public-sector or nonprofit institutions with limited digitization and budget; adoption of advanced workforce management AI is slow. Pilots of scheduling software exist but full AI-driven supervision and training adoption is rare. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Libraries are typically slow adopters of AI, especially for HR-related functions like supervision and training, which remain largely manual. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI-powered scheduling assistants and staff performance dashboards can help a human manager allocate time and identify training gaps, but the core work of supervision and training still relies heavily on direct human judgment and interpersonal skill. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can assist with scheduling optimization, training material creation, and onboarding documentation, but the core supervisory and interpersonal training tasks remain human-driven. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While scheduling tasks can be partially automated using staff management software, the supervision and training components require human judgment, relationship-building, and adaptive feedback that current AI cannot reliably perform end-to-end. AI cannot achieve 50% time savings on the full scope of this task. |
| Task automatability | claude-sonnet-5 | 1/5 | Supervising, scheduling, and training people involves interpersonal leadership, real-time judgment, and physical presence that current AI cannot perform end-to-end.'},'feasibility':{'rating':1,'rationale':'No deployed AI product manages, trains, or supervises human staff autonomously in library settings today."},'cost_ratio'}, this needs fixing |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Library staffing decisions, particularly training and supervision of employees, carry organizational and employment law liability; many organizations require human managers to own these decisions. Regulatory frameworks around employment and volunteer management create friction against full automation. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing requirement, but organizational norms and accountability for personnel management create meaningful friction against full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI scheduling tools are cheap but cover only a fraction of the task. The oversight of AI outputs and the still-human supervision and training components mean the all-in cost per full task execution remains comparable to or exceeds a human manager's labor. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI scheduling software can lower some costs, but human supervision and training still require paid staff time, keeping costs comparable rather than dramatically lower. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Scheduling software exists and performs well on calendar logistics, but no deployed AI product reliably handles live supervision, performance feedback, or training delivery at the quality a manager expects. These elements remain largely manual. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product independently supervises or trains staff; scheduling tools exist but the full task remains human-led. |
Deliver and retrieve items to and from departments by hand or using push carts.
19CI 15–24 · exposure 5 · augmentation 13 · importance 3.3/5 · click for rater detail
Deliver and retrieve items to and from departments by hand or using push carts.
19| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Libraries are typically lower-digitization, budget-constrained institutions; adoption of specialized robotics for this task remains rare and concentrated in very large university or research libraries. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Libraries and clerical support roles are a low-digitization, low-capital sector with minimal evidence of physical automation adoption for internal item transport. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI/robotic assistance could organize delivery queues or optimize routes, but the core physical task of moving items leaves limited scope for meaningful augmentation while a human remains in the loop. |
| Augmentation potential | claude-sonnet-5 | 1/5 | AI offers essentially no meaningful assistance for the physical act of carrying or wheeling items between departments. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires physical movement of items through buildings and navigation to multiple departments—capabilities that current mobile robotics in general deployment cannot reliably perform in typical library environments without extensive retrofitting and mapping. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a physical manipulation and transport task requiring mobility through a building, which current AI systems (software-based) cannot perform; only specialized robotics could attempt it, and those are not standard deployed solutions in libraries.'} , |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Library delivery involves navigation through public spaces with safety considerations and occasional human interaction (handing off items), creating some friction; however, there are no hard legal or licensing requirements that mandate human performance. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement exists, but physical infrastructure, building layout, and liability for damaged materials or navigation errors create moderate practical friction against automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Autonomous delivery robots for library settings remain expensive to procure, install, and maintain relative to the loaded wage of a library assistant, particularly for variable routes and irregular item volumes. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | Robotic delivery systems capable of navigating library stacks and handling books would require expensive hardware and infrastructure investment far exceeding the cost of a human clerical worker. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While autonomous material handling robots exist in warehouses, they are not reliably deployed in general-purpose library spaces; library-specific solutions remain experimental or require highly structured environments. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No mature commercial product performs autonomous item delivery/retrieval within library environments at scale; warehouse robots exist in narrow logistics contexts but not deployed for this clerical task. |
Repair books using mending tape, paste, and brushes or prepare books to be sent to a bindery for repair.
15CI 15–15 · exposure 0 · augmentation 13 · importance 3.7/5 · click for rater detail
Repair books using mending tape, paste, and brushes or prepare books to be sent to a bindery for repair.
15| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Libraries and archive sectors are slower to digitize and adopt automation; this task is manual and physical, characteristic of laggard adoption environments. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Libraries are a low-digitization sector for physical tasks like this, and there is no evidence of any robotic or AI adoption for book repair work. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could assist by identifying which books need repair or by assisting with damage assessment and repair method recommendations, but only marginally and alongside human execution of the actual physical work. |
| Augmentation potential | claude-sonnet-5 | 1/5 | AI could conceivably help identify damage or suggest repair vs. bindery decisions via image analysis, but this is not a demonstrated or common augmentation for this specific manual task today. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Book repair requires fine motor control, visual assessment of damage extent, and judgment about which repair method suits each book—capabilities well beyond current AI systems. Physical manipulation of materials and tools cannot be performed by existing AI. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a physical manual craft task involving hand-eye coordination with adhesives and tools; no current AI system can physically mend books or perform hands-on binding prep. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | While not legally restricted, this task involves physical work in organizational settings where human presence is expected and where library patrons may prefer human handling of materials. Organizational friction is low but physical constraints are high. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing barrier exists, but the physical nature of the task itself is the primary barrier—it requires manual dexterity and physical presence rather than regulatory protection. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | AI cannot yet perform this task at all, so comparative cost analysis is moot; the human remains the only viable option. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI substitute for physical labor here, so AI cost is not applicable/comparable—human labor remains the only option, making AI relatively 'more expensive' by default since it cannot perform the task at all. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI system can physically repair books or assess damage and prepare books for bindery services. The task involves embodied manipulation that lies outside current robotic or AI capabilities in production. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product performs physical book repair; this remains purely a manual task performed by humans and specialized bindery staff. |
Operate small branch libraries, under the direction of off-site librarian supervisors.
7CI 5–10 · exposure 0 · augmentation 38 · importance 4.0/5 · click for rater detail
Operate small branch libraries, under the direction of off-site librarian supervisors.
7| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Library systems are traditionally conservative institutions with limited IT budgets and high reliance on in-person service delivery. Automation adoption in this sector remains slow and primarily limited to administrative back-office functions, not branch operations. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Public libraries are a low-digitization, low-AI-investment sector with minimal production deployment of autonomous operational agents. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could assist with catalogue searches, administrative record-keeping, or answering basic reference questions, but the bulk of branch operation—patron relations, physical collection management, decision-making—remains human-centric with limited augmentation potential. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help with scheduling, cataloging queries, communications with the off-site supervisor, and administrative tasks, but the core on-site operational task remains largely unassisted by AI directly. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Operating a small branch library involves complex interpersonal interactions (assisting patrons, handling complaints), physical management of materials, and judgment calls about policy application that require human presence and discretion. Current AI cannot reliably replace the full scope of these operational duties. |
| Task automatability | claude-sonnet-5 | 1/5 | Operating a physical branch library involves in-person patron service, physical material handling, facility oversight, and on-site presence that current AI cannot perform end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Libraries are public or institutional entities with governance structures, liability concerns around unattended facilities, and statutory or policy requirements for human staff presence. Customer and organizational expectations strongly favor human librarians and assistants for service delivery. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing requirement exists, but the role requires physical presence, custodial responsibility, and supervisory trust that create organizational friction against remote automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The task demands a full-time human presence on-site to manage patron services, circulation, and facility operations. Any AI assistance would only supplement, not replace, the core wage cost of the library assistant position. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute for on-site branch operation, so cost comparison favors the human by default. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI system today can autonomously operate a physical library branch, which requires on-site human judgment, patron interaction, and responsibility for resources. This remains a human-dependent task in all production settings. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product independently runs a physical library branch; this remains entirely a human staffing role. |
Plan or participate in library events and programs, such as story time with children.
7CI 0–14 · exposure 0 · augmentation 38 · importance 3.8/5 · click for rater detail
Plan or participate in library events and programs, such as story time with children.
7| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Library systems operate in laggard sectors—public institutions with constrained budgets, traditional missions, and low digitization pressure. Adoption of automation for core patron-facing programming is minimal. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Public libraries are a slow-adopting, low-digitization, physically-oriented sector with minimal AI deployment for community programming. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could assist with planning (event agendas, activity ideas) or administrative support (scheduling, promotion drafts), but the core task—live, interactive engagement with patrons—receives minimal productivity boost from current AI tools while humans remain in the loop. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help brainstorm themes, draft scripts, create flyers, or suggest book lists for story time, meaningfully aiding the planning portion of the task. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Library events and programs require direct human interaction, emotional engagement, and real-time responsiveness to audience needs. Current AI systems cannot substitute for the live, relational presence that makes story time and children's programs valuable. |
| Task automatability | claude-sonnet-5 | 1/5 | Live in-person program delivery like storytelling and event facilitation requires physical presence, real-time audience engagement, and social improvisation that current AI cannot perform. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Strong regulatory and organizational barriers protect this task: libraries operate under child safety requirements, parental expectations of human staff, and professional standards that strongly favor human-led programming. Substitute automation would face high institutional friction. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing requirement, but strong organizational and community expectations for human interaction with children and patrons create real friction against automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | A library assistant conducting story time or programs costs roughly $15–25/hour (loaded wage); AI systems currently offer no viable substitute that would be cheaper when factoring in integration and oversight, and cannot deliver the human value of these services. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Since AI cannot substitute for the in-person event delivery, there is no viable AI cost comparison for the core task, only marginal savings on planning materials. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product reliably conducts children's story time or library programs end-to-end. While AI can generate text or assist with planning, the interactive, performative nature of these events requires human facilitation that production systems do not replace. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product runs or delivers children's story time or library events autonomously; this remains firmly a human-performed activity. |
Open and close library during specified hours and secure library equipment, such as computers and audio-visual equipment.
5CI 5–5 · exposure 0 · augmentation 0 · importance 4.3/5 · click for rater detail
Open and close library during specified hours and secure library equipment, such as computers and audio-visual equipment.
5| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Library automation in opening/closing remains minimal; few libraries have deployed autonomous systems for these tasks, reflecting both low digitization of physical facility management and organizational conservatism around unattended building operations. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Libraries and physical facility security are low-digitization, slow-adopting contexts, and this specific physical task is not part of any AI adoption trend. |
| Augmentation potential | claude-haiku-4-5-20251001 | 1/5 | AI provides minimal assistance to a human performing physical opening/closing and equipment security tasks; the core actions require human judgment on-site and cannot be substantially augmented by current AI tools. |
| Augmentation potential | claude-sonnet-5 | 1/5 | AI offers essentially no assistance for physically opening/closing a building or securing equipment, though non-AI tools like smart locks or alarm systems could help this is unrelated to AI augmentation. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires physical presence to unlock/lock doors, activate/deactivate security systems, and physically check equipment—actions that current AI cannot perform without dedicated robotics, which is not yet deployable at scale in libraries. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a physical task requiring presence to unlock/lock doors and physically secure equipment; no AI system can perform physical building access and security tasks. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Libraries have liability and security requirements that often legally mandate a human employee on-site to verify equipment security, handle emergency protocols, and sign off on facility status; human presence is effectively required. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Physical security responsibility, liability for building safety, and equipment custody typically require an accountable human employee, creating strong organizational and liability barriers. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The loaded cost of a human library assistant (hourly wage plus benefits) is substantially lower than deploying a mobile robot or security system to perform equivalent opening/closing and equipment-checking tasks. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | AI has no capability to substitute for this physical task, so there is no viable AI cost comparison—human labor (or traditional automated locks/alarms, not AI) is required. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No AI system today can reliably open/close physical buildings or secure tangible equipment; these require embodied presence and actuators that exist only in research or specialized contexts, not in production library deployments. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product opens/closes physical facilities or secures physical equipment; this remains entirely a human/security-system function. |
Take action to deal with disruptive or problem patrons.
3CI 0–5 · exposure 0 · augmentation 25 · importance 3.7/5 · click for rater detail
Take action to deal with disruptive or problem patrons.
3| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Libraries are generally low-tech adopters and this particular task involves human safety and legal accountability, making adoption of autonomous AI solutions extremely slow even in digitally advanced institutions. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Libraries and clerical patron-service roles are low-digitization, physically-situated environments with minimal AI agent deployment for interpersonal conflict management. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI might assist marginally by flagging repeat problem patrons via data analysis or alerting staff to concerning patterns, but the core task of real-time intervention requires human presence and judgment that AI cannot substantially enhance. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI could offer minor support such as suggesting de-escalation scripts or logging incidents, but it provides little real-time assistance during an active disruptive encounter. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires real-time judgment about patron behavior, safety assessment, de-escalation, and contextual decision-making in unpredictable social situations. Current AI systems cannot autonomously handle confrontational human interactions or make liability-bearing decisions about patron removal or intervention. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires physical presence, real-time social judgment, de-escalation skills, and often enforcement of behavioral policies with patrons—no AI system can perform this end-to-end today. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Strong barriers exist: library staff must have institutional authority to enforce rules, legal liability for patron safety and rights, potential discrimination concerns, and the requirement for human judgment in sensitive interpersonal contexts. Many jurisdictions legally require human staff for patron management. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Handling disruptive patrons often involves safety, security, potential legal liability, and sometimes policy enforcement requiring authorized staff judgment, creating strong organizational and liability-based barriers to automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The cost of deploying AI monitoring systems plus required human oversight and liability insurance would exceed the cost of a library assistant handling this task, which is typically infrequent enough not to justify full automation. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute performing this task, so any cost comparison favors the human by default since AI cannot deliver the output at all. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product reliably performs autonomous patron management or behavioral intervention in libraries. This requires human presence, emotional intelligence, and institutional authority that AI systems do not possess in production. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product handles in-person conflict resolution or physical/social intervention with disruptive individuals; this remains entirely outside current AI product capability. |
Related occupations — Office & Administrative Support
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.