Library Technicians
25-4031.00Median wage $44,580/yr68,690 employed (US)Rank #77 of 923 scored · top 8% by substitution
Assist librarians by helping readers in the use of library catalogs, databases, and indexes to locate books and other materials; and by answering questions that require only brief consultation of standard reference. Compile records; sort and shelve books or other media; remove or repair damaged books or other media; register patrons; and check materials in and out of the circulation process. Replace materials in shelving area (stacks) or files. Includes bookmobile drivers who assist with providing services in mobile libraries.
Sub-scores
0–100 · band = confidence interval from rater disagreement
Substitution51
Exposure48
Augmentation59
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
31 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
26%
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%49
panel mean rating 3.0/5 → substitution pressure 49/100
Technical feasibility todayw 20%47
panel mean rating 2.9/5 → substitution pressure 47/100
Cost vs. human wagew 15%52
panel mean rating 3.1/5 → substitution pressure 52/100
Adoption barriersw 20%inverted — strong barriers lower the score68
panel mean rating 2.3/5 (barrier strength) → substitution pressure 68/100
Sector adoption velocityw 10%31
panel mean rating 2.2/5 → substitution pressure 31/100
Task breakdown (31 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 about lost or overdue books.
96CI 92–100 · exposure 100 · augmentation 38 · importance 3.7/5 · click for rater detail
Send out notices about lost or overdue books.
96| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Library systems are highly digitized and standardized; automated notices have been mainstream in academic and large public libraries for 10+ years. Adoption is deep and mature in the sector, though small rural libraries may lag. |
| Sector adoption velocity | claude-sonnet-5 | 5/5 | Nearly universal adoption across public, academic, and school libraries; this automation has been standard practice for over a decade via ILS platforms. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist by flagging edge cases (multiple holds, lost items requiring investigation) or suggesting hold-fulfillment alternatives before a technician sends a notice, but the core task itself is so routine that augmentation is modest compared to full automation. |
| Augmentation potential | claude-sonnet-5 | 2/5 | Since the task is already fully automated by existing systems, there is little incremental augmentation value for a human performing it manually. |
| Task automatability | claude-haiku-4-5-20251001 | 5/5 | Sending overdue/lost book notices is a highly structured, rule-based task with clear triggers (due dates, item status) and templated content. Current AI systems can fully automate this end-to-end—reading database records, generating personalized notices, and dispatching via email or mail—with minimal human oversight, easily meeting the 50% time-saving threshold. |
| Task automatability | claude-sonnet-5 | 5/5 | Sending overdue/lost book notices is a templated, rule-based communication task that library management systems already automate via scheduled emails/SMS with no AI needed beyond simple scripting. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | No legal licensing or human sign-off requirement exists for sending routine notices; most barriers are organizational inertia and choice to retain human oversight for customer service. Some libraries may prefer a technician review tone or personalization, but this is discretionary, not mandated. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing, liability, or human-judgment requirement exists for sending routine overdue notices; libraries already delegate this fully to software. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Fully automated notice generation (database query, template merge, SMTP/print dispatch) costs pennies per notice versus a technician's loaded wage (~$20–30/hour) for manual composition and sending. The cost ratio is at least an order of magnitude in AI's favor. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated notice generation costs pennies in software/server time versus staff time manually tracking and sending notices, an order-of-magnitude or greater saving. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Library management systems already integrate automated notice-sending; many vendors (Innovative, Evergreen, Koha) deploy this as standard production functionality. The task is narrowly scoped and deterministic, making it reliably executable at scale in real library operations today. |
| Technical feasibility today | claude-sonnet-5 | 5/5 | Integrated library systems (e.g., Koha, SirsiDynix, Follett) have deployed automated overdue notice features in production for years, reliably triggering notifications based on due dates. |
Reserve, circulate, renew, and discharge books and other materials.
92CI 84–100 · exposure 92 · augmentation 63 · importance 4.3/5 · click for rater detail
Reserve, circulate, renew, and discharge books and other materials.
92| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 5/5 | Library automation has been mainstream and deeply adopted across public, academic, and special libraries for 20+ years; circulation automation is nearly universal in digitized library systems. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Self-service circulation technology has been widely adopted across public and academic libraries for over a decade, representing mature, deep adoption rather than early piloting. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | ILS systems augment technician productivity by handling routine transactions, tracking inventory, and flagging overdue/reserved items, allowing staff to focus on patron service, problem resolution, and specialized collection management. |
| Augmentation potential | claude-sonnet-5 | 3/5 | Where human staff still perform circulation duties, ILS software and scanning tools significantly speed up the process, though the task itself is largely already automated rather than merely augmented. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Most of this task—reserve/checkout/renewal/discharge of physical and digital materials—is already partially or fully automated in library management systems (ILS). Barcode scanning, patron lookup, due-date calculation, and renewal processing are routine; however, shelving, physical condition assessment, and patron interaction still require human oversight, so full end-to-end automation with 50% time savings is near but not quite complete today. |
| Task automatability | claude-sonnet-5 | 5/5 | Circulation tasks like reserving, checking out, renewing, and discharging materials are highly structured, rule-based transactions already handled end-to-end by integrated library systems (ILS) with barcode/RFID scanning and self-service kiosks. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | There are minimal legal barriers (no licensing requirement for automation itself), though some libraries retain human staff for patron service and dispute resolution; organizational inertia and preference for human interaction at circulation desks provide some friction but not hard barriers. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing, legal, or human-judgment requirement exists for checking out or renewing books; this is a purely administrative/logistics task with no regulatory protection. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | ILS automation costs pennies per transaction in operation; the labor cost of a technician performing these tasks manually is orders of magnitude higher, making AI/automation vastly cheaper per task-equivalent. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated kiosks and backend ILS software cost a fraction of staff wages per transaction once installed, with near-zero marginal cost per circulation event compared to a human technician's hourly wage. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Integrated Library Systems (ILS) like Polaris, Sierra, Evergreen, and Koha have been in production at scale in thousands of libraries for decades, reliably automating checkout, renewal, reservations, and discharge workflows with minimal error rates. |
| Technical feasibility today | claude-sonnet-5 | 5/5 | Self-checkout kiosks, automated holds systems, and ILS software (e.g., Koha, Sierra) are deployed at scale in public and academic libraries worldwide, reliably handling these transactions daily. |
Compile and maintain records relating to circulation, materials, and equipment.
91CI 84–97 · exposure 92 · augmentation 63 · importance 3.8/5 · click for rater detail
Compile and maintain records relating to circulation, materials, and equipment.
91| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Library systems have been digitizing and automating circulation and inventory management for decades; modern cloud-based library platforms with API-driven workflows are increasingly standard, especially in larger and well-funded institutions. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Libraries have broadly adopted automated ILS and self-checkout/RFID systems over the past two decades, making this a well-established, near-universal automation pattern in the sector. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can meaningfully assist technicians by auto-flagging anomalies, suggesting deduplication matches, generating summary reports, and cross-referencing equipment maintenance schedules, keeping humans in oversight while dramatically raising throughput. |
| Augmentation potential | claude-sonnet-5 | 3/5 | Where manual oversight or exception handling remains (e.g., reconciling discrepancies, damaged items), AI/software tools assist technicians but the core recording is already largely automated rather than merely augmented. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Record compilation and maintenance for circulation, materials, and equipment involves structured data entry, categorization, and updating—tasks highly amenable to automation. Most library management systems already automate significant portions; AI agents could handle data ingestion, deduplication, reconciliation, and report generation at >50% time savings. |
| Task automatability | claude-sonnet-5 | 5/5 | Recordkeeping for circulation, materials, and equipment is a structured data-entry and database-maintenance task that integrated library systems (ILS) already automate almost entirely via barcode/RFID scanning and automated logging. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few regulatory or legal barriers exist; library records are internal administrative data. Main friction stems from legacy system integration and institutional resistance to change, but nothing legally prevents automation of record maintenance. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing, liability, or regulatory requirement mandates human record-keeping for circulation or inventory; libraries have already widely automated this without friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Library management software operates at marginal cost per record once deployed; inference and data processing are computationally cheap compared to the loaded wage of a full-time technician maintaining these records manually. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated scanning and database systems handle thousands of transactions at a fraction of the cost of manual record-keeping by a technician. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Library management systems with automated circulation tracking, inventory management, and equipment logging are mature, deployed products in production across thousands of institutions. Database systems and APIs reliably perform these core functions at scale. |
| Technical feasibility today | claude-sonnet-5 | 5/5 | Mature ILS products (e.g., Koha, SirsiDynix, Ex Libris) reliably automate circulation and inventory record-keeping in production libraries worldwide today. |
Compile data and create statistical reports on library usage.
79CI 72–86 · exposure 80 · augmentation 88 · importance 3.5/5 · click for rater detail
Compile data and create statistical reports on library usage.
79| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Library systems are moderately digitized but adoption of automated BI varies widely; some large systems use advanced analytics while many smaller libraries still rely on manual compilation. Adoption is growing but not yet ubiquitous. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Libraries are moderately digitized institutions with some analytics adoption, but many still rely on manual spreadsheet compilation, placing them mid-tier in adoption speed. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI-powered analytics platforms augment librarian judgment by providing real-time dashboards, trend detection, and pattern discovery that help inform collection and programming decisions beyond raw statistical output. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI and BI tools substantially speed up data compilation and report generation while technicians retain oversight for interpretation and quality control. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Current AI and business intelligence tools can reliably extract library usage data from systems, aggregate it, and generate standardized reports with significant time savings. The task involves mostly structured data manipulation and template-based reporting, with minimal subjective judgment required. |
| Task automatability | claude-sonnet-5 | 4/5 | Compiling usage data and generating statistical reports is a structured, data-processing task well suited to AI/BI tools that can query databases, aggregate figures, and produce formatted reports with minimal human input.: |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few regulatory or legal barriers exist; libraries have internal authorization over their own usage data. Main friction is organizational (staff familiarity with tools, IT infrastructure setup) rather than fundamental prohibition. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing, liability, or human-contact requirements apply to internal statistical reporting; it's a purely administrative function. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Automated data pipeline and report generation (via cloud BI tools or open-source solutions) costs a fraction of a library technician's loaded wage per report cycle, especially when amortized across multiple reports. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated reporting pipelines run at a fraction of the labor cost of manual compilation, though some setup and data-cleaning oversight cost remains. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Production-grade systems (BI platforms like Tableau, Power BI, automated SQL reporting, and Python-based ETL tools) demonstrably perform this task at scale in real library systems and organizations today with high reliability. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Mature BI and reporting tools (e.g., Excel automation, Power BI, library ILS analytics modules) already generate usage statistics reports in production at many institutions today. |
Retrieve information from central databases for storage in a library's computer.
77CI 65–90 · exposure 83 · augmentation 75 · importance 3.6/5 · click for rater detail
Retrieve information from central databases for storage in a library's computer.
77| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Library systems are moderately digitized but heterogeneous; while data automation is common in corporate IT, library adoption of advanced retrieval automation remains in the pilot-to-early-production phase due to legacy systems and resource constraints. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Libraries, especially public and small institutional ones, are generally slow technology adopters with limited budgets, so uptake of automation for this task remains modest despite technical feasibility. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can assist technicians by automating routine retrieval cycles while humans handle exception handling, data quality validation, and schema reconciliation, substantially raising technician productivity in hybrid workflows. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI and scripting tools can significantly speed up a technician's retrieval and cataloging workflow, letting them focus on exceptions and quality control rather than repetitive lookups. |
| Task automatability | claude-haiku-4-5-20251001 | 5/5 | Retrieving structured information from central databases and storing it in library systems is a straightforward data extraction and transfer task that current AI agents can perform end-to-end with significant time savings using APIs, database connectors, and automation scripts. |
| Task automatability | claude-sonnet-5 | 4/5 | This is a structured, repetitive data retrieval and storage task involving standardized database queries, which current AI/automation tools can perform with substantial time savings once configured. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Minimal legal barriers exist; the main friction points are IT governance, data security compliance, and organizational processes around system access and validation, but these do not require human sign-off by law. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing or legal requirement mandates human performance of database retrieval; the main friction is integration with legacy systems and internal IT approval processes. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Automated database retrieval via APIs and scheduled jobs costs orders of magnitude less than hiring a technician—infrastructure costs are minimal compared to annual loaded wages for the role. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated scripts and API-based retrieval are far cheaper per record processed than manual technician labor once the integration is built, though setup costs exist. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Mature products (RPA tools, database management systems with API integrations, and ETL platforms) reliably perform database retrieval and data synchronization tasks in production at scale across many organizations, including libraries. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Library management systems and database integration tools already automate much of this via APIs and scripts, but many library systems still require manual verification and handling of edge cases like format mismatches. |
Compose explanatory summaries of contents of books and other reference materials.
74CI 67–81 · exposure 67 · augmentation 88 · importance 3.1/5 · click for rater detail
Compose explanatory summaries of contents of books and other reference materials.
74| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Libraries are transitioning to digital systems but remain relatively conservative adopters of AI. Pilot projects exist, but widespread production deployment of AI summarization in library systems remains limited. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Libraries and educational institutions are moderate adopters of AI, with some pilots for cataloging and summarization but not yet widespread production use. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI summaries can significantly assist librarians by providing drafts that speed composition, allowing humans to focus on refinement, accuracy verification, and tailoring to patron needs rather than starting from scratch. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI drafts summaries that technicians can quickly review and edit, substantially boosting throughput while retaining human quality control. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI can generate summaries of book contents at scale with tools like GPT-4, but quality varies with source material complexity and the need for domain accuracy. Human review and refinement are typically required, limiting time savings to roughly 50% rather than full automation. |
| Task automatability | claude-sonnet-5 | 4/5 | LLMs can read or ingest text and produce accurate summaries with minimal editing, meeting the ≥50% time-saving bar for most reference materials. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Library technicians are not licensed professionals, and there are no regulatory restrictions on AI-generated summaries. The main friction is organizational preference for human quality control and patron trust, not legal barriers. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing or legal requirement mandates a human write these summaries; libraries can freely adopt AI tools for this purpose. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Inference cost for summarizing reference materials is orders of magnitude cheaper than librarian labor, with minimal integration overhead for batch processing of texts. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | AI summarization costs pennies per document versus staff time, making it far cheaper than manual summary writing. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Production systems (ChatGPT, Claude, specialized summarization APIs) reliably generate summaries of published materials at scale. However, library-specific accuracy requirements and occasional errors in factual detail prevent a perfect 5 rating. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Summarization is a mature, widely deployed AI capability (e.g., in document management and library software), though quality varies with specialized or obscure texts. |
Enter and update patrons' records on computers.
73CI 70–76 · exposure 75 · augmentation 50 · importance 4.0/5 · click for rater detail
Enter and update patrons' records on computers.
73| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Libraries have historically lagged in automation adoption relative to finance and professional services sectors. While digitization is widespread, production deployment of AI agents for record management remains limited; most libraries rely on manual or legacy RPA rather than modern AI systems. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Public and academic libraries have adopted self-service and automated ILS record systems over recent decades, but many smaller libraries still rely on manual staff entry, giving a moderate, uneven adoption pace. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI assists library technicians by auto-populating fields from external sources, checking for duplicates, and flagging data quality issues, thereby raising throughput and accuracy. However, the task is primarily transactional, so augmentation gains are moderate compared to more judgment-heavy library work. |
| Augmentation potential | claude-sonnet-5 | 3/5 | Existing library software assists technicians with templates, autofill, and validation checks, meaningfully speeding up record entry and reducing errors while staff still oversee accuracy. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Most of this task is straightforward data entry and updating, which AI can perform reliably with structured forms and databases. Current systems can validate patron information, update records, and flag inconsistencies with minimal human oversight, achieving substantial time savings. Some complexity around edge cases (name variations, address disambiguation) may require occasional human review, preventing a perfect 5. |
| Task automatability | claude-sonnet-5 | 4/5 | Data entry and record updates in library management systems follow structured, repetitive workflows that can be largely automated via forms, barcode scanning, self-service kiosks, and API integrations with minimal human intervention. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Low regulatory barriers exist; library patron records are not heavily licensed activities, and there is no statutory requirement for human sign-off. Main friction is organizational preference for human oversight of patron data and modest IT infrastructure constraints in some libraries, but these are surmountable. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing or legal requirement mandates human entry of patron records; the main friction is privacy/data protection concerns and organizational IT change management rather than hard regulatory barriers. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | AI inference and integration cost for patron record entry is minimal—pennies per transaction—compared to loaded hourly wages for library technicians ($15–25/hour), making automation an order of magnitude cheaper at scale. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated data entry via self-service terminals and software integrations costs far less per transaction than staff time, though initial system setup and occasional exception handling add some cost. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Library management systems (ILS platforms) increasingly include automated patron record management features, and general-purpose data-entry AI (form-filling agents, RPA) has demonstrated reliability in production environments. Deployed systems handle straightforward record updates consistently, though integration varies by library IT infrastructure. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Library patron management systems (ILS software like Koha, Sierra, self-checkout kiosks) already automate most record creation and updates in production libraries today, though some manual entry and correction still occurs. |
Verify bibliographical data for materials, including author, title, publisher, publication date, and edition.
71CI 65–76 · exposure 70 · augmentation 75 · importance 3.6/5 · click for rater detail
Verify bibliographical data for materials, including author, title, publisher, publication date, and edition.
71| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Large academic and research libraries have piloted or deployed automated verification (e.g., batch ISBN lookups, vendor data matching), but small and mid-size libraries still rely on manual processes. Adoption is steady but not universal. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Libraries and archival institutions are typically slower adopters of AI tooling compared to fast-moving information/finance sectors, with automation of cataloging tasks still emerging unevenly. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI-assisted verification (highlighting mismatches, suggesting corrections, auto-populating fields) significantly speeds a technician's review workflow. The human remains in loop for judgment calls on ambiguous or rare records, making this a strong augmentation use case. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI and automated database matching significantly speed up verification of bibliographic fields, letting technicians focus on exceptions and quality control rather than manual lookup. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Verification of bibliographic data against database records and external sources is highly structured, rule-based work. Current AI systems (OCR + LLMs + API lookups) can reliably extract and match author, title, publisher, date, and edition at scale, with human spot-checking achieving >50% time savings on large batches. |
| Task automatability | claude-sonnet-5 | 4/5 | Verifying bibliographic metadata against catalogs like WorldCat, ISBN databases, and publisher records is a structured lookup/comparison task well-suited to automation with existing APIs and AI tools, though occasional edge cases (rare editions, foreign titles) require human judgment. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Library workflows are largely unregulated; no legal requirement mandates human verification of metadata. Main friction is organizational inertia and librarian preference for human review of edge cases, not licensing or liability barriers. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement mandates a human perform this task, but institutional workflows and quality-control norms in libraries create some friction against full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Automated verification via batch APIs and OCR pipelines costs pennies per record. A library technician costs $20–30/hour loaded; the cost differential is at least one order of magnitude in favor of AI for high-volume verification work. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated lookups against bibliographic databases are extremely cheap compared to a technician manually cross-referencing each field, though some oversight cost remains for exceptions. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Deployed library management systems (e.g., Alma, Koha integrations) and OCR vendors already perform automated bibliographic verification in production. Error rates on standard print metadata are low; edge cases (rare editions, variant titles) require human review but most routine verification is operationalized. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Library systems already use automated cataloging tools (e.g., OCLC, MARC record matching) that check bibliographic data, but full end-to-end verification without human review is not yet standard practice in most libraries. |
Order all print and non-print library materials, checking prices, figuring costs, preparing order slips, and making payments.
67CI 65–70 · exposure 70 · augmentation 75 · importance 3.7/5 · click for rater detail
Order all print and non-print library materials, checking prices, figuring costs, preparing order slips, and making payments.
67| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Libraries lag behind corporate procurement in AI adoption; most operate with partial digitization and heterogeneous vendor ecosystems. While purchasing departments elsewhere adopt automated procurement rapidly, library systems adoption remains slower due to budget constraints and lower digital maturity. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Libraries are a slow-adopting, often underfunded sector with legacy systems, so despite technical feasibility, actual deployment of AI-driven acquisitions remains limited and gradual. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can substantially augment librarians by automating price/availability lookups, flagging budget overruns, and suggesting optimal vendors, freeing staff to focus on collection development strategy and material selection. Human judgment on content and curation remains valuable while routine logistics are AI-assisted. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools can meaningfully speed up price comparisons, cost calculations, and order slip generation, letting technicians focus on selection judgment and vendor relationships. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | AI can automate most of this task end-to-end: scraping supplier catalogs, comparing prices, generating purchase orders, and integrating with payment systems. The core workflow (identify item → check price → create order slip → process payment) is highly structured and repetitive, achievable with current RPA and agent tools, likely saving >50% of human time. |
| Task automatability | claude-sonnet-5 | 4/5 | Ordering, price checking, cost calculation, and order slip preparation are structured, repetitive data-entry tasks that AI-integrated procurement/ILS systems can largely automate today, though final approval and vendor exceptions still need oversight. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few hard regulatory barriers exist; most libraries are public or nonprofit institutions without licensing requirements for purchasing. Primary friction is organizational (legacy systems, vendor relationships, desire for human curation) rather than legal, making adoption straightforward once integrated. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement for this clerical task, though payment authorization and budget accountability create moderate organizational controls before full automation of payments. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | AI-driven procurement (including integration, vendor APIs, and payment processing) costs substantially less than the loaded wage of a library technician conducting manual catalog searches, price comparisons, order entry, and payment reconciliation. Amortized per-task cost is typically an order of magnitude lower. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated procurement software and API-based price lookups are far cheaper per transaction than technician labor once implemented, though setup and vendor integration add some cost. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Production systems exist for procurement automation, including vendor management platforms, invoice processing, and payment integration tools used by libraries and enterprises. While some edge cases (unusual materials, complex licensing) require human review, the routine ordering path is reliably deployable today. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Library management systems (e.g., integrated library systems with acquisitions modules) offer automated ordering workflows, but many libraries still rely on manual review and vendor-specific quirks limit full reliability. |
Conduct reference searches, using printed materials and in-house and online databases.
67CI 62–72 · exposure 70 · augmentation 100 · importance 3.6/5 · click for rater detail
Conduct reference searches, using printed materials and in-house and online databases.
67| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Libraries and information services are adopting AI-powered search and discovery tools, but adoption remains uneven: well-funded institutions pilot AI integration while smaller libraries lag. The sector is digitizing rapidly but adoption of full automation is still in the pilot-to-early-production phase. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Libraries and educational institutions are adopting AI search tools at a moderate pace, with pilots more common than full production deployment in public/institutional libraries. |
| Augmentation potential | claude-haiku-4-5-20251001 | 5/5 | AI dramatically augments reference librarians by instantly scanning multiple databases, suggesting search terms, and ranking results by relevance, enabling staff to handle more complex queries and focus on specialized research guidance rather than manual searching. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI search assistants substantially speed up reference queries, suggest sources, and summarize results, greatly boosting technician productivity while they verify and contextualize findings. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Current AI systems can search databases, retrieve documents, and synthesize reference material with high accuracy for most query types. While some nuanced reference requests may require human judgment, the core search and retrieval functions meet or exceed the 50% time-saving threshold, especially for standardized reference tasks. |
| Task automatability | claude-sonnet-5 | 4/5 | Reference search using databases is largely retrieval and synthesis, tasks that AI search/RAG systems and chatbots handle well, though verifying obscure printed materials and specialized library systems still needs human oversight. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Libraries have established workflows and patron expectations for human-assisted reference work, creating organizational friction. Some institutions have licensing constraints on database access, and patrons may prefer human-verified results, though no hard legal requirement mandates human search conduct. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement for this task, though some in-house/proprietary database access controls and institutional procurement policies create moderate friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | AI inference costs for conducting reference searches are minimal (cents per query), while the loaded wage for a library technician conducting searches runs $25–50+ per hour. The cost ratio heavily favors automation once integrated into library systems. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated database and online searches run at a fraction of the cost of a technician's time once integrated, though initial setup and licensing for in-house databases add some cost. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Production systems including vector databases, large language models, and semantic search tools already perform reference searches reliably across online databases and digital collections. Library-integrated search systems and AI-powered discovery tools are deployed at scale in real organizations, though some edge cases still require human verification. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | AI-powered search and library discovery tools (e.g., AI-enhanced catalog search, research assistants) exist and are used, but reliability on niche archival or printed-only sources remains limited in production. |
Design posters and special displays to promote use of library facilities or specific reading programs at libraries.
65CI 56–74 · exposure 58 · augmentation 88 · importance 3.4/5 · click for rater detail
Design posters and special displays to promote use of library facilities or specific reading programs at libraries.
65| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Libraries are comparatively laggard sectors—typically smaller, public institutions with limited IT infrastructure, tight budgets, and slower digital adoption. Pilot use of AI design tools exists, but production deployment remains rare; many libraries still rely on staff or contractors for poster design. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Public libraries are typically under-resourced, slow-moving institutions with limited IT/design budgets and low digitization pressure, so AI design tool adoption for this niche task lags more digitized sectors. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI excels at rapidly generating design alternatives, suggesting layouts, and drafting promotional copy, which meaningfully accelerates a library technician's workflow. A technician can curate and refine AI outputs faster than creating designs from scratch, substantially raising productivity while maintaining human judgment on final messaging and library mission fit. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI design tools dramatically speed up ideation, layout, and image generation for library technicians who still refine messaging and finalize the display, making this a strong augmentation case. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI can generate poster designs, select layouts, and draft promotional copy with moderate setup, but the task requires iterative creative refinement, brand alignment, and understanding of specific library programs that demand human direction. Roughly half the work—initial concept generation, layout suggestions, copy drafting—can be automated; the final curated output typically requires human review and revision. |
| Task automatability | claude-sonnet-5 | 4/5 | AI image and design generation tools can produce complete poster drafts and layouts from a text prompt, and combined with templating tools this covers most of the creative and production work with minor human review. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Libraries have no legal requirement for human-created promotional materials, and there is minimal liability risk in using AI-generated designs. Some organizational friction exists around brand guidelines and staff comfort with AI tools, but no hard regulatory or licensing barriers prevent adoption. |
| Adoption barriers | claude-sonnet-5 | 1/5 | There is no licensing, regulatory, or safety requirement tied to designing library promotional materials, so nothing legally blocks AI-assisted or AI-generated output. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | AI design generation costs are very low (pennies to dollars per poster), while a library technician's fully loaded cost is $40–60 per hour. Even accounting for oversight and iteration, AI-assisted generation is substantially cheaper than hiring a graphic designer or technician to create displays from scratch. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | AI-assisted design tools cost a few dollars per month versus staff time spent on manual graphic design, making this drastically cheaper at scale. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | AI image generation tools (DALL-E, Midjourney) and design templates exist and produce usable outputs, but current systems struggle with text legibility, brand consistency, and nuanced program messaging. Products work in practice but with notable quality gaps and require significant human editing and oversight. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Products like Canva Magic Design, Adobe Firefly, and general image generators are deployed and used for marketing collateral, but outputs still often need human editing for text accuracy, branding consistency, and library-specific context. |
Compile bibliographies and prepare abstracts on subjects of interest to particular organizations or groups.
64CI 56–72 · exposure 58 · augmentation 88 · importance 2.5/5 · click for rater detail
Compile bibliographies and prepare abstracts on subjects of interest to particular organizations or groups.
64| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Libraries and information services sectors have slower digitization and automation adoption compared to finance or tech; most library operations remain human-centered. AI-driven bibliographic and abstracting tools are emerging but not yet deeply embedded in production workflows at scale. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Libraries and information services are moderately digitized and adopting AI research tools, but full workflow automation is still emerging rather than deeply entrenched. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can significantly assist library technicians by drafting bibliographies and abstracts that humans refine and validate, reducing manual compilation time and allowing focus on quality curation and subject-matter judgment. Language models excel at rapid draft generation and source matching, raising technician output substantially when paired with human oversight. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI tools substantially speed up literature searching, citation compilation, and abstract drafting, letting technicians focus on curation, relevance judgment, and quality control. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI can auto-generate bibliographies from source material and produce draft abstracts using language models, achieving partial automation of compilation and summarization. However, curating bibliographies for specific organizational needs, validating source relevance, and ensuring abstract quality for specialized groups typically requires human judgment and domain expertise. |
| Task automatability | claude-sonnet-5 | 4/5 | LLMs can search literature databases, extract citations, and generate abstracts with substantial time savings, though verification of accuracy and citation correctness still needs human review. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few formal barriers exist; library roles are not licensed professions and AI use faces minimal regulatory restriction. Some organizational preference for human curation and institutional friction around task reassignment exist, but nothing legally prevents substitution of AI-assisted or autonomous bibliography and abstract generation. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing or regulatory requirement mandates a human library technician for compiling bibliographies or abstracts; it's an administrative/informational task. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | AI inference for generating drafts is inexpensive per task, with minimal computational overhead for bibliographic compilation and abstract generation relative to librarian labor. Integration and validation overhead is moderate, but the raw cost per output is substantially below human technician wages. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | AI-assisted literature search and abstract drafting is dramatically cheaper per bibliography than manual compilation, though some human oversight cost remains. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Products exist (citation managers with AI features, abstracting tools, LLM-based summarization) that can assist with these tasks, but they produce material error rates in relevance judgment and require substantial human validation. Deployed systems lack the nuanced understanding of organizational context and subject matter expertise needed for fully reliable autonomous output. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Tools like reference managers with AI summarization, Elicit, and citation-generating LLMs are deployed, but hallucinated citations and incomplete coverage remain material error sources in production use. |
Process interlibrary loans for patrons.
60CI 52–67 · exposure 58 · augmentation 75 · importance 3.7/5 · click for rater detail
Process interlibrary loans for patrons.
60| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Libraries are moderately digitized but often lag in automation compared to corporate sectors; pilot projects and partial automation are common, but comprehensive AI-driven ILL processing remains uncommon in production at scale across the sector. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Libraries are generally slower technology adopters with constrained budgets, and while ILL software is standard, deeper AI-driven automation is not yet widely deployed in this sector. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI systems can augment technicians by auto-populating forms, flagging priority requests, predicting availability, and managing status workflows, substantially raising technician productivity while keeping humans in the loop for exceptions and patron relations. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI and existing library systems already meaningfully speed up request processing, tracking, and communication, letting technicians handle higher volumes with less manual tracking effort. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Interlibrary loan processing involves routine administrative steps (search requests, data entry, status tracking, notification) that are readily automatable; however, exceptions (complex requests, verification issues) and patron communication nuance keep it from reaching full 5-level automation. |
| Task automatability | claude-sonnet-5 | 3/5 | Much of the interlibrary loan workflow (request submission, tracking, status updates via systems like ILLiad or WorldShare) can be automated by existing library software, but exceptions, physical item handling, and communication with partner institutions still need human oversight. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | No licensing or regulatory requirement mandates human performance of ILL processing; some organizational preference for human oversight and community trust may slow adoption, but barriers are primarily operational friction rather than legal. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement for this task, though institutional agreements, library policies, and reliance on partner institution systems create moderate procedural friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | API calls and workflow automation are inexpensive once integrated into existing library systems; the cost per transaction is orders of magnitude below the loaded wage of a technician performing manual data entry and tracking. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Existing ILL management software is relatively cheap to run compared to staff time for the digital/administrative portion, but physical handling and exception resolution keep overall costs comparable to human labor. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Existing library management systems and RFP/ILL software automate parts of the workflow, but end-to-end handling requires integration across multiple systems and human judgment on edge cases; production reliability is high for routine cases but spotty for complex scenarios. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Library automation systems already handle much of ILL request routing and tracking in production, but full end-to-end automation (verification, shipping, exception handling) still requires staff intervention regularly. |
Collect fines and respond to complaints about fines.
58CI 51–65 · exposure 53 · augmentation 63 · importance 3.7/5 · click for rater detail
Collect fines and respond to complaints about fines.
58| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Libraries have slow IT adoption relative to other sectors, budgetary constraints, and cultural emphasis on service; automated fine systems are emerging but widespread production deployment of AI complaint handling remains limited. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Libraries are public-sector, often under-resourced institutions with slower technology adoption compared to finance or professional services sectors. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist technicians by drafting responses, flagging duplicate complaints, suggesting appropriate fine waivers based on policy, and organizing complaint data, meaningfully improving processing efficiency while preserving human judgment on disputes. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI/chatbots and automated systems can handle routine fine payments and FAQs, freeing technicians to focus on complex or sensitive complaints, meaningfully boosting productivity. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | Current AI systems can handle routine fine collection notices and automated complaint responses for standard scenarios, but require human judgment for dispute resolution, special circumstances, and customer escalation, limiting full automation to perhaps 40-60% of typical cases. |
| Task automatability | claude-sonnet-5 | 4/5 | Fine collection and standard complaint responses are largely rule-based (payment processing, waiver policies) and can be handled by automated systems, chatbots, and payment portals with high time savings. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | While not legally restricted, libraries often prefer human judgment for fairness and customer relationships; institutional resistance to replacing face-to-face complaint handling and patron trust concerns present moderate friction. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement, but organizational policy often mandates human discretion for waivers, appeals, or patron relations, creating moderate friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | Automated fine notices and chatbot responses cost a fraction of human technician time per interaction, though oversight and exception handling require some human involvement, making the ratio favorable at approximately 1:5 or better. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated payment processing and templated complaint responses are far cheaper than staff time per transaction, though escalations still require human cost. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While basic chatbots exist for complaint handling, production systems for library fine collection lack the contextual flexibility and sensitivity needed for real disputes; most deployments remain limited to notification rather than full complaint resolution. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Many library systems already deploy self-service kiosks, online payment portals, and automated notices, but nuanced complaint handling (disputes, hardship cases) still often routes to staff. |
Issue identification cards to borrowers.
57CI 25–90 · exposure 62 · augmentation 50 · importance 4.0/5 · click for rater detail
Issue identification cards to borrowers.
57| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Library systems have historically been slow to adopt AI and automation broadly, and card issuance remains largely manual or semi-automated at most institutions. Adoption is concentrated in larger urban systems, not representative of the sector's overall velocity. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Many libraries have adopted self-service kiosks and online registration, but many smaller or under-resourced libraries still rely on manual processes. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI could assist by auto-populating forms from scanned IDs or suggesting corrections to manually entered data, moderately improving librarian workflow. However, the task itself is short enough that augmentation gains are incremental rather than transformative. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI/automation can speed up verification and data entry for staff, but the task is simple enough that augmentation offers only modest incremental value. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Issuing library cards requires verifying identity documents, collecting personal information, and often photographing borrowers—steps that involve variable human interaction and judgment. Current AI cannot reliably perform the full multi-step process without human verification of identity and signature capture, though parts like data entry could be automated. |
| Task automatability | claude-sonnet-5 | 5/5 | Issuing library cards is a simple data-entry/verification task easily handled by self-service kiosks, online registration forms, and automated ID systems that already meet or exceed the time-saving threshold. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Identity verification and card issuance often require a human staff member to physically verify documents and sign off on access credentials for legal and accountability reasons. Many libraries have policies mandating human review before access is granted. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing or legal requirement mandates a human issue library cards, though some libraries prefer in-person verification of identity/address for fraud prevention. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | The task is labor-light (issuing a card takes minutes), so the base human wage per instance is already low. AI infrastructure and integration costs would be comparable to or exceed the direct labor savings from automating a simple, quick task. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated kiosks and online registration cost a fraction of a cent per transaction compared to staff time spent on manual card issuance. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While some libraries use self-service kiosks or automated data entry systems, these typically require human staff oversight for identity verification and final approval. No end-to-end AI product reliably issues library cards independently in production at scale. |
| Technical feasibility today | claude-sonnet-5 | 5/5 | Self-checkout kiosks, online card registration, and automated patron management systems are already deployed at scale in public and academic libraries. |
Review subject matter of materials to be classified and select classification numbers and headings according to classification systems.
54CI 44–65 · exposure 58 · augmentation 75 · importance 3.7/5 · click for rater detail
Review subject matter of materials to be classified and select classification numbers and headings according to classification systems.
54| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Public and academic libraries have been slow to adopt automation at scale; most deployment is pilot-stage or limited to routine cataloging, and budget constraints and cultural attachment to human curation slow velocity in a sector with relatively low digitization maturity. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Libraries and cultural heritage institutions are generally slower adopters of AI compared to finance or tech sectors, with pilots more common than full production deployment. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI classification suggestions and metadata extraction significantly assist technicians by pre-populating candidates and flagging ambiguities, allowing them to review and refine rather than classify from scratch, thereby meaningfully raising per-technician throughput while maintaining human judgment. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can suggest candidate classification numbers and headings, significantly speeding up the technician's review and decision process while the human retains final judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI can extract metadata and suggest classification numbers via automated tagging and machine learning models trained on library classification systems, reducing manual review work by 40–60%. However, domain expertise in nuanced subject categorization and system-specific headings still requires human oversight for accuracy. |
| Task automatability | claude-sonnet-5 | 4/5 | Classifying materials by subject and assigning standardized headings (e.g., Dewey, LCSH) is largely pattern-matching over text content, which current LLMs can do well when given the text and classification schema, saving substantial time though not fully eliminating review.The task's rule-based nature makes it highly amenable to automation. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Libraries have institutional workflows and stakeholder preference for consistent human judgment on subject matter; there are no strict legal bars to automation, but organizational friction (staff retraining, system integration) and the desire for accuracy on specialized materials create meaningful adoption friction. |
| Adoption barriers | claude-sonnet-5 | 2/5 | There is little regulatory or licensing requirement for who can classify library materials, though some institutions have internal quality standards and reliance on established schemas that create moderate organizational friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI classification tools reduce per-item processing cost, but library technician wages are modest, and the overhead of AI system maintenance, training data curation, and human review overhead mean total cost savings remain marginal rather than substantial. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated subject classification via AI/NLP tools costs a small fraction of a technician's hourly wage once integrated, though initial setup and periodic human verification add some cost. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Deployed products (e.g., library automation platforms with ML-driven classification helpers, DOI-based metadata extraction) exist and perform reasonably on standard materials, but error rates remain material—especially for interdisciplinary or specialized content—and integration with legacy library systems is often narrow and inconsistent. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Some library systems and metadata tools use AI-assisted classification suggestions, but most production cataloging still relies on human review and correction, so deployed reliability is moderate rather than fully mature. |
Claim missing issues of periodicals and journals.
50CI 23–77 · exposure 50 · augmentation 63 · importance 3.4/5 · click for rater detail
Claim missing issues of periodicals and journals.
50| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Library systems remain relatively low-digitization, non-competitive sectors with limited incentive to adopt AI agents for back-office tasks. Adoption of automation in library operations is slow, pilot-heavy, and concentrated in large academic institutions; small public and school libraries dominate the workforce. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Library and archival services are a mixed-digitization sector; many libraries already use automated claiming modules, but smaller institutions still do this manually, giving moderate overall adoption. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI could meaningfully assist by automatically flagging missing issues against hold records, summarizing circulation patterns to prioritize claims, and generating standardized claim letters. A human technician could then review AI-flagged gaps and send claims more efficiently, raising productivity on the routine parts while maintaining judgment over which gaps matter. |
| Augmentation potential | claude-sonnet-5 | 4/5 | Even where not fully automated, AI/software tools significantly reduce technician effort by flagging gaps and drafting claim communications, letting staff focus on exceptions. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Claiming missing periodical issues requires checking inventory records, identifying gaps, contacting publishers, and following up on orders—tasks involving routine database queries and standardized correspondence. While AI could assist with flagging gaps and drafting routine inquiries, the process involves publisher-specific procedures, verification of actual receipt, and judgment calls about which claims are worth pursuing, limiting full automation well below the 50% threshold. |
| Task automatability | claude-sonnet-5 | 4/5 | This is a rule-based, repetitive workflow (identify gaps in serial holdings, generate and send claim requests to vendors/publishers) that maps well onto automation via ILS/ERM system rules and scripted claiming logic.integration."},"note":ignore |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Libraries typically operate within vendor contracts and licensing agreements that specify claim procedures; many publishers require direct contact from authorized library staff or account representatives. Organizational policies around inventory accountability and vendor relationships create friction, and librarians are often required by contract terms to initiate and sign off on claims. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing, legal, or human-judgment requirement is attached to this clerical task; libraries already delegate it to automated systems routinely. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Current AI and automation tools still require significant integration overhead, data cleanup, and human oversight to connect library systems with publisher claim systems. The cost of setup and maintenance is likely comparable to or higher than employing a technician part-time on this routine but irregular task. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated claiming via existing library software requires minimal marginal compute cost compared to staff time spent manually tracking and issuing claims, though some setup/maintenance cost exists. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No mature products exist that perform this task end-to-end in library production systems. General-purpose workflow automation tools and email systems can handle parts of it, but there is no specialized system deployed at scale that reliably tracks missing issues, interfaces with library management systems, and manages multi-vendor publisher claims autonomously. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Library systems (e.g., ILS/ERM modules like ExLibris Alma, Sierra) already have automated serials claiming features that detect missing issues and generate claims automatically in production use at many libraries. |
Answer routine telephone or in-person reference inquiries, referring patrons to librarians for further assistance, when necessary.
43CI 30–56 · exposure 38 · augmentation 63 · importance 4.3/5 · click for rater detail
Answer routine telephone or in-person reference inquiries, referring patrons to librarians for further assistance, when necessary.
43| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Library systems are typically underfunded, conservative, and lag in digital transformation. While some libraries have piloted chatbots, meaningful production-level displacement of reference technicians remains limited and slow across the sector. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Public and educational library sectors are generally slow adopters of AI due to budget constraints and traditional service models, with pilots more common than full deployment. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist technicians by pre-answering common questions, suggesting referral pathways, and retrieving relevant library resources quickly, moderately improving their productivity on routine inquiries while they handle complex cases and patron interaction. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI search tools, chatbots, and knowledge bases can meaningfully speed up how technicians locate answers and triage inquiries, improving efficiency while humans still manage escalations and personal interaction. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can handle some routine reference questions via text or phone interfaces, real-world inquiries often require context-sensitive judgment about when to escalate to librarians, understanding of patron intent, and handling edge cases. Current systems cannot reliably manage the full referral decision-making at 50% time savings without human oversight. |
| Task automatability | claude-sonnet-5 | 3/5 | Many routine reference questions (hours, availability, basic lookups) can be handled by chatbots or voice AI, but distinguishing routine from complex queries and escalating appropriately requires judgment that current systems handle imperfectly. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Public libraries have some expectation of human customer service, and organizations often prefer human-library technician presence for patron satisfaction. However, there are no hard legal or licensing barriers preventing AI deployment for routine reference work. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement blocks AI handling routine inquiries, though patron preference for human interaction and library service norms create moderate friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI voice systems and chatbots have low marginal cost, but integration into library systems, training on local collections, and human oversight of escalation decisions add significant overhead. Setup and maintenance costs remain comparable to part-time library technician wages. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated chat/voice systems for FAQ-type reference questions are inexpensive to run compared to staff wages, though initial setup and maintenance add some cost. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Chatbots and IVR systems can answer basic queries, but deployed library systems still rely heavily on human technicians for nuance, patron interaction quality, and accurate escalation decisions. No mature production system has demonstrably replaced this task end-to-end in library settings. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Library chatbots and AI-driven catalog search tools exist and are deployed in some public and academic libraries, but coverage of in-person and nuanced phone inquiries remains limited and often supplements rather than replaces staff. |
Process print and non-print library materials to prepare them for inclusion in library collections.
41CI 30–52 · exposure 38 · augmentation 63 · importance 4.1/5 · click for rater detail
Process print and non-print library materials to prepare them for inclusion in library collections.
41| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Libraries are typically underfunded, conservative in technology adoption, and concentrated in public and academic sectors with slower IT investment cycles; while barcode and basic digital systems exist, deep AI adoption for material processing remains limited. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Libraries (public/educational sector) are historically slow adopters of new technology due to budget constraints and legacy systems, so despite available tools, deep production-scale AI adoption remains limited. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist technicians by automating metadata lookup, suggesting classifications, and flagging condition issues for human review, meaningfully reducing manual data entry and research time while preserving human judgment on complex or damaged materials. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-assisted cataloging tools, metadata suggestion, and batch record matching meaningfully speed up technicians' work even though humans still handle physical processing and verification. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While basic metadata entry and barcode scanning can be automated, the task requires judgment about physical condition assessment, categorization decisions, and handling of varied media types that current AI struggles with end-to-end. Significant human oversight remains necessary for quality control and exception handling. |
| Task automatability | claude-sonnet-5 | 3/5 | Cataloging metadata lookup, barcode/label generation, and record entry can be substantially automated with existing systems, but physical processing (covering, stamping, shelving prep) still requires manual handling., so only part of the workflow meets the threshold. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Library systems have established workflows, staff expertise, and organizational inertia favoring human technicians; however, there are no hard legal or licensing barriers preventing automation of routine cataloging and material preparation tasks. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement blocks automation of cataloging; some organizational friction exists around workflow standards and quality control, but no legal mandate requires a human to perform this task. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Current AI automation (barcode systems, basic data entry) saves only incremental cost—perhaps 10–20% of labor—and requires infrastructure investment and ongoing human oversight that partially offset savings, making the all-in cost ratio close to or potentially higher than human labor for comprehensive processing. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Software-based cataloging automation is cheap relative to labor, but the physical handling component still requires paid human time, keeping overall cost roughly comparable rather than order-of-magnitude cheaper. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Narrow automation exists for barcode generation and basic cataloging data entry, but no deployed products reliably handle the full pipeline of inspecting, assessing condition, and preparing diverse print and non-print materials autonomously. Human technicians must still verify and manage most steps. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Integrated library systems already automate copy cataloging, MARC record retrieval, and batch processing, but physical item preparation and edge-case cataloging still require staff intervention in production settings. |
Organize and maintain periodicals and reference materials.
39CI 30–49 · exposure 38 · augmentation 50 · importance 3.5/5 · click for rater detail
Organize and maintain periodicals and reference materials.
39| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Libraries are traditionally slower adopters of automation technology; while digital cataloging is standard, AI-driven organization and maintenance of physical materials remains in pilot phases in most public and academic institutions. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Libraries (public/academic institutions) are typically slower adopters of AI-driven physical automation compared to fast-moving information-sector firms, though some large systems have implemented automated sorting. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist library technicians by automating cataloging, detecting condition issues via image analysis, and recommending organization schemes, meaningfully reducing manual search and administrative overhead while the technician retains curatorial control. |
| Augmentation potential | claude-sonnet-5 | 3/5 | Software tools assist with cataloging, tracking overdue/returned periodicals, and generating shelving lists, improving technician efficiency without replacing physical organizational tasks. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While cataloging and metadata tagging can be partially automated with OCR and classification systems, the physical organization and maintenance of periodicals—shelving, condition assessment, handling delicate materials—remain largely manual tasks that current AI cannot perform end-to-end with 50% time savings. |
| Task automatability | claude-sonnet-5 | 3/5 | Cataloging metadata, sorting, and tracking periodicals can largely be automated via ILS software and barcode/RFID systems, but physical shelving, condition checks, and handling exceptions still require human presence. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Library operations face modest adoption friction: institutional preferences for human curation, stakeholder trust in physical collections care, and the need for domain expertise in archival standards create some resistance, though no hard legal barriers exist to AI-assisted organization. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement for this task, but institutional budgets, existing physical infrastructure, and legacy collections create organizational friction to full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI solutions for library automation (cataloging software, RFID systems) still require substantial upfront investment and ongoing human supervision; the loaded cost of library technicians remains competitive or lower for mixed manual-digital workflows. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Automation systems (RFID, ILS) require significant upfront capital and maintenance, and physical shelving/retrieval still needs paid staff, so cost savings versus a library technician's wage are moderate at best. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Some library management systems can automate cataloging workflows and bar-code tracking, but few production systems reliably handle the full spectrum of organization and maintenance tasks including physical arrangement, binding decisions, and condition monitoring without significant human oversight. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Library management systems (ILS) with automated check-in, serials tracking, and RFID sorting are deployed in many libraries, but full end-to-end automation of physical organization is not standard everywhere. |
Catalogue and sort books and other print and non-print materials according to procedure and return them to shelves, files, or other designated storage areas.
36CI 24–49 · exposure 38 · augmentation 50 · importance 4.0/5 · click for rater detail
Catalogue and sort books and other print and non-print materials according to procedure and return them to shelves, files, or other designated storage areas.
36| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Library technician roles are in laggard sectors: small to medium institutions, physical operations, and low digitization intensity. While some large research libraries pilot automation, most public and school libraries rely on human labour due to cost and organizational inertia. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Libraries are a low-digitization, budget-constrained, physically-oriented sector where automation adoption (beyond basic ILS software) has been slow and uneven, especially for physical sorting/shelving. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI-assisted barcode scanning, catalog lookup tools, and shelf-reading suggestions can meaningfully improve a technician's speed and accuracy on portions of the task (identifying call numbers, flagging misplaced items). The human remains essential for judgement, handling special collections, and exception resolution. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-assisted cataloguing tools (auto-classification, metadata suggestion, barcode/RFID systems) meaningfully speed up the cataloguing portion of this task, though they don't touch the physical shelving component. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can read barcodes and identify some materials, the full task requires physical manipulation (shelving, handling diverse formats), spatial reasoning about library layouts, and handling exceptions or damaged items. Current robotics + vision systems exist in narrow lab settings but cannot reliably execute the complete workflow at scale with 50% time savings. |
| Task automatability | claude-sonnet-5 | 3/5 | Cataloguing metadata generation can be substantially automated with existing systems (barcode scanning, MARC record lookup, AI classification tools), but the physical sorting, shelving, and storage placement remains a manual, embodied task that current AI cannot perform end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Libraries operate under collection management policies and institutional workflows, and there is moderate organizational friction around automation. However, no legal licensing requirement or regulatory barrier exists; the main friction is procedural and institutional adoption preferences. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing or legal requirement mandates a human perform this task, but physical handling of materials and specific institutional storage/classification schemes create moderate organizational friction against full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | A library technician performs this task at $20–30k annually loaded cost. Current robotic systems (hardware + integration + maintenance) cost hundreds of thousands of dollars and require significant human oversight, making the AI cost-per-task substantially higher. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Software for cataloguing metadata is cheap, but physical shelving/sorting still requires human labor or expensive robotic infrastructure, keeping overall cost comparable to or only modestly cheaper than human labor for most libraries. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Shelf-reading and inventory robots exist in some libraries (e.g., robotic scanning), but end-to-end cataloguing, sorting, and returning materials to correct locations remains mostly manual in production systems. Deployed solutions handle only subsets (scanning, not shelving) with ongoing human oversight. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Library automation systems (RFID sorting, integrated library systems with automated cataloguing suggestions) are deployed in many libraries, but full end-to-end automation including physical shelving is rare and mostly limited to well-funded institutions with conveyor/robotic sorting. |
Help patrons find and use library resources, such as reference materials, audio-visual equipment, computers, and other electronic resources and provide technical assistance when needed.
33CI 30–35 · exposure 25 · augmentation 63 · importance 4.2/5 · click for rater detail
Help patrons find and use library resources, such as reference materials, audio-visual equipment, computers, and other electronic resources and provide technical assistance when needed.
33| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Public libraries and academic institutions have been slow to adopt AI-driven reference services; most AI adoption remains pilot-stage or limited to basic chatbots, with production deployment of comprehensive AI technician replacements rare. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Libraries are typically slow, under-resourced adopters of AI technology, with pilots being more common than full production deployment. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI-powered search, recommendation engines, and chatbots can significantly augment technician productivity by pre-filtering resources, suggesting answers to FAQs, and diagnosing basic technical issues before human escalation, while patrons and staff remain central to the interaction. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI search tools, chatbots, and knowledge bases can meaningfully help technicians quickly locate resources or answer common questions, augmenting but not replacing their role. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can assist with locating digital resources and answering reference queries, the task fundamentally requires human interaction to understand patron needs, diagnose technical issues, and provide in-person guidance—capabilities that remain beyond autonomous AI systems today. |
| Task automatability | claude-sonnet-5 | 2/5 | While AI chatbots can answer some directional/reference queries, the task requires in-person physical assistance (equipment, computer troubleshooting) and personalized guidance that current AI cannot fully replace end-to-end.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Library services are public-facing and often embedded in institutional governance that values human service, but there are no hard legal licensing barriers to partial automation of reference lookups or resource-discovery workflows. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing barrier, but there is a human-contact/service expectation and reliance on in-person troubleshooting that creates moderate organizational friction against full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | The combination of human interaction, physical presence, and real-time technical support required makes current AI solutions more expensive than lower-wage library technician labor when all integration and oversight costs are included. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI tools are cheap for basic Q&A, but the hands-on technical troubleshooting and physical navigation assistance still require a human, keeping blended costs relatively high for full task substitution. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Chatbots and search systems exist for library resource discovery, but they operate at narrow scope and lack the contextual understanding to troubleshoot equipment issues or adapt to diverse patron needs; no mature product reliably handles the full breadth of this task in production library environments. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some libraries deploy chatbots or search assistants, but no deployed product reliably handles the full range of in-person technical assistance and physical resource help. |
Provide assistance to teachers and students by locating materials and helping to complete special projects.
33CI 30–35 · exposure 25 · augmentation 50 · importance 3.9/5 · click for rater detail
Provide assistance to teachers and students by locating materials and helping to complete special projects.
33| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Library systems are typically underfunded and conservative in technology adoption; while some libraries pilot digital tools, they remain focused on human-led service delivery and have not shown rapid uptake of autonomous AI agents for patron assistance. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Libraries and schools are generally slower adopters of AI tools compared to finance or tech sectors, with pilots more common than widespread production deployment. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can helpfully suggest resources, summarize research, and flag relevant materials to accelerate librarian workflows, but the interpretive work of understanding project scope and guiding students remains human-centric and benefits most from skilled technician insight. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI search and recommendation tools can meaningfully speed up locating materials and brainstorming project ideas, augmenting the technician's efficiency while they retain the interpersonal and physical task components. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can search library catalogs and suggest materials, the task requires understanding nuanced project requirements, navigating physical locations, and providing contextual guidance that demands human judgment and real-time adaptation to specific student/teacher needs. |
| Task automatability | claude-sonnet-5 | 2/5 | AI can help locate digital resources or answer reference questions, but physical shelving, in-person guidance, and hands-on project assistance require human presence and cannot be fully automated end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Schools and libraries value human interaction with students and have organizational inertia around staffing; however, no legal licensing barrier exists and budget pressure could drive adoption of AI-assisted or automated systems over time. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No strict licensing requirement, but institutional preference for human staff interacting with students, safeguarding/supervision norms in schools, and reliance on physical library operations create moderate friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI systems for library assistance require ongoing training, maintenance, and integration costs that approach or exceed the part-time wage of library technicians, especially when accounting for the human oversight needed to handle edge cases and complex requests. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | While AI search assistance is cheap, the task still requires a human on-site for physical retrieval and personalized guidance, keeping overall cost comparable to or only slightly cheaper than staffing. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Library discovery systems and chatbots exist but struggle with context-dependent query interpretation and lack integration with the physical workspace; no deployed product reliably handles the full end-to-end assistance workflow across diverse project types. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Library chatbots and search tools exist and assist with catalog lookups, but reliable deployed systems that handle personalized project assistance and physical material retrieval in schools/libraries remain narrow and immature. |
Maintain and troubleshoot problems with library equipment, including computers, photocopiers, and audio-visual equipment.
33CI 30–35 · exposure 25 · augmentation 50 · importance 3.7/5 · click for rater detail
Maintain and troubleshoot problems with library equipment, including computers, photocopiers, and audio-visual equipment.
33| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Library systems are moderately digitized but tend to adopt technology slowly; many libraries still operate with legacy equipment and limited IT infrastructure. Adoption of AI maintenance tools is in early pilot phases, not widespread production deployment across the sector. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Libraries are a low-digitization, resource-constrained sector with slow technology adoption for equipment maintenance workflows. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI-powered troubleshooting guides, symptom checkers, and documentation assistants can meaningfully help technicians diagnose equipment faster and reduce service tickets. However, augmentation remains partial since the technician must still perform hands-on repairs and validate AI suggestions against physical systems. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-powered troubleshooting guides, chatbots, and diagnostic knowledge bases can meaningfully speed up problem identification even though physical repair remains manual. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI could assist in diagnosing some equipment problems through symptom analysis or documentation review, the task requires hands-on physical intervention (replacing parts, adjusting hardware, reconnecting cables) that current AI systems cannot perform. Remote support and diagnostic chatbots exist but cannot fully replace the technician's physical presence and tactile problem-solving. |
| Task automatability | claude-sonnet-5 | 2/5 | Physical troubleshooting and hands-on equipment repair requires manual diagnosis and intervention that current AI cannot perform autonomously, though AI can assist with diagnostic guidance and knowledge lookup.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Library operations often have service agreements and warranty requirements that mandate authorized technicians, creating some friction. However, most public and academic libraries can adopt remote diagnostic support or AI-assisted troubleshooting without legal barriers, so barriers are moderate rather than hard. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement, but physical presence and hands-on manipulation of equipment create practical barriers to full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI diagnostic tools have low inference costs but require significant integration overhead and still necessitate a human technician for actual repairs. The loaded cost of a library technician (~$35–45k annually) is competitive with the combined cost of AI systems plus technician oversight, making AI not yet cost-advantageous. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI diagnostic assistance is cheap, but the physical labor of maintenance and repair still requires a human technician, so overall cost savings are limited. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | AI-powered diagnostic tools and troubleshooting guides exist in limited form, but no deployed product reliably handles the full range of library equipment problems independently. Most solutions require human technicians to interpret results and perform repairs, making end-to-end automation unfeasible today. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | No deployed product autonomously maintains or repairs physical library equipment; some chatbot-based IT helpdesk tools exist for basic software issues but not for hands-on hardware fixes. |
Operate and maintain audio-visual equipment, such as projectors, tape recorders, and videocassette recorders.
24CI 14–35 · exposure 20 · augmentation 38 · importance 2.8/5 · click for rater detail
Operate and maintain audio-visual equipment, such as projectors, tape recorders, and videocassette recorders.
24| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Libraries are traditionally low-digitization, risk-averse sectors with limited automation investment; physical equipment maintenance still relies on human technicians. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Libraries are a low-digitization, low AI-adoption sector, and this specific physical maintenance task sees essentially no AI deployment. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI-assisted diagnostics (fault detection, predictive maintenance alerts) and remote monitoring dashboards can meaningfully support technicians, though core operation remains human-driven. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI could help with troubleshooting guides or manuals lookup, but offers minimal direct assistance to the physical operation and repair work itself. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Basic equipment operation (power on/off, input switching) is automatable, but troubleshooting physical failures, tape loading, and maintaining proper connections requires human judgment and dexterity that current AI cannot reliably handle end-to-end. |
| Task automatability | claude-sonnet-5 | 2/5 | This is a physical, hands-on task involving equipment setup, operation, and repair that current AI cannot perform end-to-end; no software-based automation reduces the physical labor involved.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Libraries often require trained staff to handle expensive equipment, and liability concerns around damage to audio-visual materials create institutional friction against full automation. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing barrier exists, but the physical nature of equipment handling and troubleshooting creates a practical barrier since robots/AI cannot yet substitute for manual dexterity. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI systems for physical equipment operation would require expensive robotics and integration; a technician's loaded wage is likely lower than the total cost of deployment and maintenance. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | AI has no capability to replace the physical labor of operating/maintaining equipment, so a human is required regardless of cost comparison. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed AI systems reliably operate and maintain this diverse array of physical A/V equipment autonomously; remote monitoring and basic diagnostics exist but full operation remains beyond current products. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product operates or maintains physical AV hardware; this remains a manual technical task requiring physical presence. |
Sort and deliver library mail and packages.
24CI 15–33 · exposure 13 · augmentation 13 · importance 3.0/5 · click for rater detail
Sort and deliver library mail and packages.
24| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Library systems are slow adopters of automation; mail and package sorting is routine, low-priority work that libraries typically assign to existing staff. No meaningful production adoption of AI-driven mail delivery in libraries is evident. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Libraries are a low-digitization, low-automation-investment sector, and physical mail handling is not a target of current AI/robotics adoption trends. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | Barcode scanning and sorting software can assist in tracking and organizing mail flow, but the task itself is largely manual and routine; AI assistance would be marginal and focused only on inventory/routing optimization. |
| Augmentation potential | claude-sonnet-5 | 1/5 | AI offers essentially no meaningful assistance for the physical acts of sorting and delivering mail and packages. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Sorting mail by destination and scanning packages can be partially automated with barcode readers and sorting systems, but the delivery component—navigating varied library layouts, handling exceptions, and managing human interaction—requires significant manual effort. The task does not meet the 50% time-saving threshold with current off-the-shelf AI. |
| Task automatability | claude-sonnet-5 | 1/5 | Physical sorting and delivery of mail and packages requires manipulation and mobility that current AI systems cannot perform; this is a physical logistics task, not a cognitive/digital one. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Library operations are not heavily regulated for mail handling itself, but human contact for package accountability, hand-off protocols, and organizational preference for staff presence create moderate friction to full automation. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing or legal requirement mandates a human specifically for mail sorting, but physical presence and building navigation create practical barriers to substitution. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Barcode scanners and basic sorting equipment exist cheaply, but integrated robotic mail-handling and delivery systems with navigation are expensive relative to the loaded wage of a library technician, especially for small to medium libraries. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI-only solution for physical mail sorting/delivery, so AI costs are effectively irrelevant or would require expensive robotic infrastructure exceeding human labor costs. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Some warehouse sorting robots exist, but reliable deployment of end-to-end mail/package sorting and delivery in a library environment with human interaction is not demonstrated in production. Most automation remains laboratory or narrow pilot stage. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product autonomously sorts and physically delivers mail within library settings; any automation here would require robotics, which remains research-stage for this context. |
Check for damaged library materials, such as books or audio-visual equipment, and provide replacements or make repairs.
23CI 10–35 · exposure 13 · augmentation 38 · importance 3.7/5 · click for rater detail
Check for damaged library materials, such as books or audio-visual equipment, and provide replacements or make repairs.
23| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Library institutions tend toward laggard digitization and slow capital investment in automation. Pilot projects on damage detection exist but production deployment in library systems remains sparse and limited to larger institutions; most small and mid-size libraries rely entirely on manual inspection. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Libraries are a low-digitization, low-budget sector with minimal AI adoption for physical maintenance tasks, and no meaningful robotic automation trend exists here. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI-powered image flagging or damage severity scoring could assist technicians by prioritizing inspection tasks or suggesting repair versus replacement decisions, reducing time spent on visual assessment. However, the final judgment and repair execution remain human, providing moderate augmentation value. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI could help track inventory of damaged items or flag items needing repair via a database, but it offers minimal assistance to the actual physical inspection and repair work. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Detecting damage requires visual and tactile inspection that current computer vision can partially address (visual defects in book covers or media), but assessing structural integrity, binding quality, or equipment functionality demands nuanced judgment. End-to-end automation with equal quality remains infeasible; a human must still verify and decide on repair versus replacement. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires physically handling books and equipment to inspect for damage and perform manual repairs, tasks that current AI systems cannot execute since they lack embodiment for physical manipulation. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Library organizations value human judgment on condition and repair decisions, and there is modest organizational friction in adopting automated inspection. No hard legal barrier exists, but customer trust in human oversight of collection maintenance and the decision to discard or repair creates practical friction. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement exists, but the physical nature of the work (handling materials, using repair tools) creates a practical barrier to any non-physical AI system performing it directly. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Implementing computer vision and repair logistics automation requires substantial infrastructure investment and oversight. For a routine inspection and decision task performed by a modestly paid technician, the all-in cost of vision systems, integration, and human review likely exceeds or approaches the human wage, with limited savings. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI system that can perform this physical inspection/repair task, so the cost comparison favors humans by default since AI cannot substitute at all. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While image recognition can flag visible damage on book covers or cases, no deployed library system reliably performs the full inspection task autonomously. Existing tools are narrow (damage detection only) and require human verification; they do not operate independently at scale in production library workflows. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs physical inspection and repair of library materials; this remains a manual, physical task requiring human dexterity and judgment. |
Plan and conduct children's programs, community outreach programs, and other specialized programs, such as library tours.
16CI 5–28 · exposure 13 · augmentation 63 · importance 3.5/5 · click for rater detail
Plan and conduct children's programs, community outreach programs, and other specialized programs, such as library tours.
16| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Library systems have historically resisted full automation of public-facing community programs; adoption data shows libraries use AI for back-office tasks (cataloging, scheduling) but retain human program leadership as a core mission-critical function and community touchpoint. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Public libraries are a slow-moving, low-digitization, community-service sector with minimal AI deployment for direct patron-facing programming. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can significantly assist planning (suggest age-appropriate content, generate activity frameworks, manage registrations, draft promotional materials), allowing technicians to design richer programs and reach more communities. The human technician remains essential for delivery and relationship-building, but productivity gains in prep and outreach are substantial. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help technicians brainstorm program themes, draft outreach flyers, create activity content, or plan tour scripts, meaningfully aiding preparation even though delivery remains human. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can help draft program curricula, schedule logistics, and generate marketing materials, the core task—conducting live programs and engaging children or community members—requires real-time human interaction, relationship-building, and responsive adaptation to audience needs. AI cannot meaningfully replace the in-person facilitation component that defines this task. |
| Task automatability | claude-sonnet-5 | 1/5 | Planning and conducting in-person children's programs, tours, and community outreach requires physical presence, live facilitation, and real-time human interaction that AI cannot perform end-to-end today. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Library programs, especially children's programming, often require background checks, institutional accountability, and organizational trust in the person conducting activities. Substituting human program leaders with AI faces strong organizational and reputational friction, and caregivers expect human facilitation of educational content for minors. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing requirement exists, but the task involves direct interaction with children and community members, creating strong safety, trust, and organizational expectations for human staff. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI tools for program planning and outreach support are inexpensive, but they augment rather than replace the technician. The cost of AI oversight and quality control (ensuring programs meet community needs) approaches human wage for the specialized skillset required. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | Since AI cannot perform the in-person facilitation component, a human must still be paid to deliver the program, so there is no meaningful cost substitution. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Deployed AI systems can assist with planning (calendar tools, content suggestions) but no mature product reliably conducts live children's or community programs autonomously. Chatbots and scheduling tools exist, but they do not perform the full task as typically understood in library work. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product conducts library tours or live children's programming; AI is at best used for backend planning materials, not the actual delivery of the task. |
Train other staff, volunteers, or student assistants and schedule and supervise their work.
16CI 10–21 · exposure 5 · augmentation 50 · importance 3.6/5 · click for rater detail
Train other staff, volunteers, or student assistants and schedule and supervise their work.
16| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Library systems are traditionally conservative in automation, with strong cultural emphasis on human mentorship and relationship-based work. Adoption of AI for staff training and supervision remains minimal and experimental across the sector. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Libraries are a low-digitization, public-sector environment with slow AI adoption for management and supervisory functions. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist by auto-generating training schedules, suggesting best practices for onboarding, and providing record-keeping; however, the human technician remains essential for actual mentoring, performance evaluation, and adaptive coaching. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help create training materials, schedules, and performance tracking, meaningfully assisting the supervisor but not replacing the interpersonal supervisory role. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Training and supervision inherently require human judgment, relationship-building, and adaptive response to individual performance variations. Current AI cannot reliably conduct interactive training sessions or make real-time supervisory decisions that meet a 50% time-saving threshold without human oversight. |
| Task automatability | claude-sonnet-5 | 1/5 | Training, scheduling, and supervising people requires interpersonal judgment, motivation, and situational adaptation that current AI cannot perform end-to-end., only fragments like scheduling logistics are automatable. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Library institutions have ingrained expectations that management and training of staff require human judgment and accountability. Some organizational friction exists, but no hard legal barrier prevents experimentation with AI-assisted scheduling or training frameworks. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing requirement, but organizational norms and accountability for supervising people create moderate friction against full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The minimal automatable portion (basic scheduling) is already cheap; the irreplaceable supervisory and training components require human presence, making end-to-end AI replacement economically infeasible compared to a library technician's loaded wage. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | While scheduling software is cheap, the human supervisory and training components still require paid staff time, so overall cost savings from AI are limited. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While AI can assist with scheduling (calendar tools) and generate training materials, no deployed product reliably handles the core supervisory and adaptive training functions independently. Existing systems lack the contextual judgment and interpersonal responsiveness this task demands. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product autonomously trains and supervises human staff; scheduling tools exist but supervision and training of people remain human-led in practice. |
Deliver and retrieve items throughout the library by hand or using pushcart.
15CI 15–15 · exposure 0 · augmentation 13 · importance 4.2/5 · click for rater detail
Deliver and retrieve items throughout the library by hand or using pushcart.
15| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Library systems are typically public or non-profit institutions with constrained budgets, low digital-transformation momentum, and physical-world dependencies; adoption of physical automation is minimal and experimental. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Libraries are a low-digitization, low-tech-investment sector with minimal deployment of robotic automation for physical material handling. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could assist with inventory tracking and predictive shelving logistics, but the core task of physically moving items remains human-driven; modest augmentation potential through better route-planning tools rather than transformative productivity gain. |
| Augmentation potential | claude-sonnet-5 | 1/5 | AI offers essentially no assistance for the physical act of carrying or pushing items through a library space. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires physical navigation through variable library layouts, manipulation of diverse items, and responsive human interaction (shelving, retrieval). Current AI cannot perform the full end-to-end physical task; robotic systems exist but are extremely limited and far from production-ready in typical library environments. |
| 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 deployed for this purpose.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Libraries have some digital-inventory systems and organizational interest in automation, but there are no hard legal or licensing barriers to deploying automation itself; however, human-contact requirements (assisting patrons) and the need for reliable physical performance provide moderate friction. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement exists, but physical workspace constraints, safety around patrons, and cost of robotic retrofitting create practical friction against automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Any current robotic or autonomous system capable of library item movement would cost far more to acquire, maintain, integrate, and oversee than hiring a library technician at typical wages. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | Any robotic solution capable of this would require significant capital investment in hardware, navigation systems, and maintenance, far exceeding the cost of a human technician for this simple physical task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product reliably performs autonomous item delivery and retrieval in live libraries at scale. Research-stage mobile manipulation exists, but no mature production system handles the variability of book types, patron requests, and dynamic library environments. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No mainstream deployed product performs library item retrieval and delivery via pushcart; warehouse robotics exist but are not adapted to library floor navigation and shelving at scale. |
Take actions to halt disruption of library activities by problem patrons.
3CI 0–5 · exposure 0 · augmentation 25 · importance 3.8/5 · click for rater detail
Take actions to halt disruption of library activities by problem patrons.
3| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Libraries are traditionally low-digitization environments with strong human-contact requirements. Adoption of AI for behavioral management remains negligible, and organizational culture prioritizes human relationship-building with patrons. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Libraries are a low-digitization, physically-grounded environment with minimal AI adoption for behavioral/security management tasks. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could potentially assist by identifying early warning signs of disruption via security cameras or traffic patterns, but the core task of de-escalation and intervention remains fundamentally human-centric with limited scope for AI enhancement. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI could help with alert systems, logging incidents, or providing de-escalation scripts, but offers limited real-time assistance during active disruptions. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires real-time judgment, interpersonal de-escalation, situational awareness, and physical presence to manage disruptive behavior—capabilities that current AI systems cannot perform. AI cannot autonomously respond to or resolve patron conflicts in a library setting. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires real-time physical presence, judgment, de-escalation, and often direct human interaction to manage disruptive individuals; current AI cannot perform this end-to-end.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Strong legal and organizational barriers exist: library staff have authority to enforce conduct policies, and liability for safety depends on qualified human judgment. Removing human discretion from disruption management exposes institutions to liability and conflicts with patron safety obligations. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Handling disruptive patrons often involves safety, legal liability, and institutional policy requiring trained staff to intervene, creating strong barriers to automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Deployment of AI for this task would require expensive surveillance, anomaly detection, and robotics infrastructure alongside human oversight, making it far more costly than trained library staff managing disruptions directly. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute for this physical, interpersonal task, so cost comparison favors the human by default. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI product reliably handles real-time behavioral disruption management in libraries or similar public settings. This task fundamentally requires human judgment and physical intervention that AI cannot provide at production scale. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product handles in-person conflict de-escalation or physical security intervention in libraries today. |
Related occupations — Educational Instruction & Library
How to read this
A high substitution score does not mean this job disappears — it means a large share of its current tasks face replacement pressure, so the mix of tasks is likely to change. High augmentation alongside substitution typically means the occupation reorganizes around the protected tasks. Wide confidence intervals mean the rater panel disagreed: treat those scores as open questions, not verdicts.
What would change this score
New model capabilities (automatability, feasibility), falling inference costs (cost ratio), regulation and licensing shifts (barriers), and measured sector adoption (velocity) all re-enter at every index release. Each release is recomputed, versioned and kept queryable — scores are claims with a date on them, not permanent labels.