Teaching Assistants, Special Education
25-9043.00Rank #697 of 923 scored · top 76% by substitution
Assist a preschool, elementary, middle, or secondary school teacher to provide academic, social, or life skills to students who have learning, emotional, or physical disabilities. Serve in a position for which a teacher has primary responsibility for the design and implementation of educational programs and services.
Sub-scores
0–100 · band = confidence interval from rater disagreement
Substitution19
Exposure17
Augmentation40
Substitution — the headline: capability discounted by cost, barriers and adoption.
Exposure — technical capability alone, regardless of whether anyone deploys it.
Augmentation — how much AI assists without replacing. High here + moderate substitution = a changing job, not a disappearing one.
Tasks on the substitution scale
30 rated tasks, binned by substitution score.
Position among all scored occupations
Distribution of 923 occupation scores; the marker is this occupation.
Tasks with substitution ≥ 70
3%
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%18
panel mean rating 1.7/5 → substitution pressure 18/100
Technical feasibility todayw 20%16
panel mean rating 1.6/5 → substitution pressure 16/100
Cost vs. human wagew 15%17
panel mean rating 1.7/5 → substitution pressure 17/100
Adoption barriersw 20%inverted — strong barriers lower the score29
panel mean rating 3.8/5 (barrier strength) → substitution pressure 29/100
Sector adoption velocityw 10%12
panel mean rating 1.5/5 → substitution pressure 12/100
Task breakdown (30 tasks)
Substitution pressure per task, weighted by O*NET importance in the composite. Expand a task for the full rater audit trail — every rating, every model, every rationale.
Take class attendance and maintain attendance records.
84CI 76–92 · exposure 87 · augmentation 75 · importance 3.6/5 · click for rater detail
Take class attendance and maintain attendance records.
84| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | K–12 and higher education institutions have adopted digital attendance systems and biometric tools widely; many school districts now use automated check-in as standard practice, indicating rapid and broad adoption in the education sector. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Many schools have adopted digital attendance systems, but special education classrooms often still rely on paraprofessional-driven manual processes due to bundled in-person duties. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI attendance tools augment teaching assistants by flagging chronic absences, generating alerts, and enabling real-time dashboards that help staff identify at-risk students; the assistant remains in the loop to investigate and intervene. |
| Augmentation potential | claude-sonnet-5 | 4/5 | Digital tools substantially reduce time spent on this specific task, freeing the aide to focus on student support duties. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Attendance tracking is highly structured and repetitive: recording names, times, and status (present/absent/late). AI-powered systems can automate roll-call via facial recognition, QR codes, or RFID integration and automatically update digital records with >50% time savings at equal quality. |
| Task automatability | claude-sonnet-5 | 5/5 | Attendance tracking is a simple, structured data-entry task easily handled by existing digital systems (badges, apps, roll-call software) with minimal human oversight. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Schools can adopt automated attendance without legal barriers, though some prefer human supervision for accountability and some parents object to facial recognition. These are adoption frictions rather than hard legal blocks on automation. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement exists for attendance-taking itself, though special education contexts may require a trained aide present for other simultaneous duties, creating minor organizational friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | AI-based attendance systems cost a few dollars per student per year and require minimal human oversight once deployed, making them an order of magnitude cheaper than paying a teaching assistant hourly wages for this clerical task. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated attendance systems cost a small fraction of the labor time required for a human to manually record and maintain attendance logs. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Deployed products exist and perform reliably in production: attendance management systems (e.g., Classdojo, Skyward, school LMS integrations) with computer vision or mobile check-in handle this task at scale in thousands of schools today with low error rates. |
| Technical feasibility today | claude-sonnet-5 | 5/5 | School administration software and attendance-tracking systems are already deployed at scale in most districts, reliably logging and reporting attendance. |
Grade homework and tests, and compute and record results, using answer sheets or electronic marking devices.
66CI 56–75 · exposure 62 · augmentation 75 · importance 3.8/5 · click for rater detail
Grade homework and tests, and compute and record results, using answer sheets or electronic marking devices.
66| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | K–12 and higher-ed sectors have already widely deployed LMS and automated grading tools; this is among the earliest and most mature EdTech automations, with rapid ongoing adoption across public and private institutions. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | K-12 special education is a moderately digitized but resource-constrained sector; adoption of automated grading tools is real but slower and shallower than in tech-forward industries. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI-powered grading assistants and analytics dashboards help teachers and TAs identify struggling students, flag anomalies, and summarize performance trends, meaningfully raising educator productivity while humans retain pedagogical judgment. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI grading tools substantially speed up scoring and record-keeping for teaching assistants, letting them focus more time on individualized student support while the tool handles routine computation. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | Grading objective questions and recording results can be partially automated with current systems (OMR devices, LMS scoring), but subjective answers, partial credit decisions, and handling edge cases typically require human review. This achieves roughly half the time-saving at equal quality for mixed-format assignments. |
| Task automatability | claude-sonnet-5 | 4/5 | Grading against answer keys and computing/recording scores is a highly structured task that AI-based scanning and grading tools already handle for objective items, with automated essay/short-answer scoring covering much of the rest.4 |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Minimal legal or licensing barriers exist to automated grading of objective items. Some schools have contractual or policy friction favoring human grading, and teacher/parent preference for human review adds modest friction, but nothing prevents technical adoption. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement blocks automated grading, though special education students may need individualized assessment adjustments that create some organizational friction and oversight expectations. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | Electronic grading systems cost a few cents per assignment per student in large deployments, far cheaper than paying a teaching assistant hourly wages to hand-grade and manually record results. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated grading and scoring software is inexpensive per unit compared to the hourly cost of a teaching assistant, especially at scale for standardized tests and objective homework. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Mature, deployed products (Scantron, Canvas, Blackboard, Google Classroom) reliably automate objective-question grading and recording at scale in thousands of schools. Limitations exist for complex rubrics and handwritten subjective work, but core functionality is production-proven. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Products like Scantron-style optical scanners plus AI-assisted grading tools (e.g., Gradescope) are deployed in schools today, but special-education contexts often need customized rubrics and human review for accommodations, limiting full reliability. |
Requisition and stock teaching materials and supplies.
61CI 52–70 · exposure 58 · augmentation 63 · importance 2.9/5 · click for rater detail
Requisition and stock teaching materials and supplies.
61| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Schools are laggards in digitization relative to corporate sectors; many still rely on manual requisitioning and offline inventory tracking, with limited adoption of autonomous procurement systems. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | K-12 special education support functions are a low-digitization, resource-constrained sector with slow uptake of automated procurement tools. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI systems can significantly assist by tracking inventory levels, recommending supplies based on curriculum needs, automating purchase orders, and alerting staff to stock shortages—transforming the speed and accuracy of supply management while humans retain oversight. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-based inventory tracking and reorder reminders can meaningfully assist staff in managing supplies, though physical handling remains manual. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Requisitioning and stocking involve predictable, routine tasks—inventory tracking, purchase orders, and record-keeping—that current AI systems can handle end-to-end with substantial time savings. However, some manual physical stocking and context-specific educational supply decisions may require human judgment, preventing a full 5 rating. |
| Task automatability | claude-sonnet-5 | 3/5 | Ordering and inventory tracking can largely be automated with software (procurement systems, reorder triggers), but physical stocking and vendor coordination still need human execution. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few legal or regulatory barriers exist; school purchasing policies and supply-chain integrations create some friction, but nothing prevents automation or delegation to an AI system managing the requisition workflow. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing or regulatory requirement ties this task to a certified human; it's purely administrative and logistical. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | Inventory and procurement automation tools cost far less than the loaded wage of a teaching assistant for these administrative tasks, making the AI approach substantially cheaper once deployed. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Basic inventory/ordering software is cheap relative to staff time, but full automation including physical restocking still requires human labor, keeping costs comparable overall. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | ERP and inventory management systems partially automate requisitioning, but they require human configuration and oversight; no fully autonomous end-to-end product currently manages both the educational context assessment and physical stocking without human intervention. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Inventory/procurement software is widely deployed in schools and districts, but many special education classrooms still rely on manual, ad hoc supply requests rather than integrated AI-driven systems. |
Laminate teaching materials to increase their durability under repeated use.
61CI 24–97 · exposure 58 · augmentation 0 · importance 2.8/5 · click for rater detail
Laminate teaching materials to increase their durability under repeated use.
61| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Schools and educational institutions have widely adopted laminating machines for decades; this is a standard practice in most special education and general classroom settings, reflecting deep, sustained adoption. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Special education support roles and physical material prep tasks show minimal AI adoption, as this sector is low-digitization for hands-on classroom tasks. |
| Augmentation potential | claude-haiku-4-5-20251001 | 1/5 | Lamination offers no meaningful human augmentation; the machine either does the task or it doesn't. There is no collaborative or assistive dimension where AI enhances human judgment. |
| Augmentation potential | claude-sonnet-5 | 1/5 | AI provides no meaningful assistance for the physical act of laminating materials; it's a manual craft task unrelated to AI-augmentable cognitive work. |
| Task automatability | claude-haiku-4-5-20251001 | 5/5 | Lamination is a fully automatable mechanical process: feeding material into a laminating machine, setting temperature/speed, and collecting output requires no human judgment or decision-making. Commercial laminating machines can process materials end-to-end with minimal human intervention, easily exceeding 50% time savings. |
| Task automatability | claude-sonnet-5 | 2/5 | Laminating is a physical, hands-on task involving machine operation and manual handling of materials, which current AI systems and robots cannot perform end-to-end without specialized robotic hardware not typically deployed in schools.aaa |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | There are minimal regulatory or licensing barriers to automating lamination. Some organizational friction may exist (equipment cost, familiarity with machines), but nothing legally prevents or requires human involvement. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing, regulatory, or liability barriers exist; the barrier is purely physical/robotic capability, not legal or organizational restriction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Laminating machines have low per-task costs (film and electricity) amounting to cents per document, while a human performing this task manually costs $15–30 per hour in loaded wages. AI/mechanical automation is orders of magnitude cheaper. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI-based substitute for this physical process, so a human (or simple mechanical laminator) remains the only cost-effective option. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Laminating machines are mature, widely deployed technology found in schools, print shops, and offices worldwide. Modern thermal and cold laminators perform this task reliably and consistently at scale, with minimal error rates. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No consumer or classroom product uses AI to physically laminate materials; this remains a manual task done by staff using laminating machines. |
Prepare lesson outlines and plans in assigned subject areas and submit outlines to teachers for review.
45CI 34–56 · exposure 38 · augmentation 75 · importance 3.4/5 · click for rater detail
Prepare lesson outlines and plans in assigned subject areas and submit outlines to teachers for review.
45| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Education sectors show slower AI adoption than information/finance sectors, with many schools remaining cautious about delegating curriculum planning to automated systems. Pilots are growing but production deployment of AI lesson planning remains limited, particularly in special education where customization and legal compliance are critical. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | K-12 special education is a historically slow-adopting, resource-constrained sector with limited AI tool integration compared to corporate or professional services sectors. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can substantially assist teaching assistants by generating initial outlines, suggesting accommodations for special needs, and automating formatting, freeing time for refinement and personalization. This augmentation significantly raises productivity while keeping the TA in a reviewing and customizing role. |
| Augmentation potential | claude-sonnet-5 | 4/5 | Generative AI can meaningfully speed up drafting of lesson outlines and structures, letting the teaching assistant focus on tailoring content to individual student needs. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI can generate draft lesson outlines and basic plans using curriculum standards and learning objectives, achieving partial automation. However, alignment with specific teacher requirements, special education accommodations, and school-specific standards typically requires human review and iteration, limiting time savings to roughly 50% with significant setup. |
| Task automatability | claude-sonnet-5 | 2/5 | AI can draft generic lesson outlines but adapting them to specific special education needs, IEP goals, and individual student accommodations requires human judgment and contextual knowledge AI lacks reliable access to.dry. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | No legal license is required for outline preparation, but educational institutions often have oversight processes, curriculum review requirements, and teacher sign-off mandates that create organizational friction. Special education contexts carry additional compliance and IEP requirements that increase oversight burden. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No strict licensing requirement blocks AI-assisted drafting, but IEP compliance, legal documentation standards, and teacher sign-off create moderate organizational friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | Inference cost for generating lesson outlines is very low (pennies per task), while a teaching assistant's time spent on outline preparation costs significantly more. Integration and oversight add modest overhead, but the cost advantage remains substantial. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI drafting assistance is cheap per use, but the human review, customization for individual student needs, and teacher oversight still consume significant paid staff time, keeping overall costs comparable. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Products like ChatGPT and specialized educational AI tools can generate lesson outlines reliably, but they lack deep context about individual student needs, IEP requirements, and institutional constraints. Current systems produce usable drafts that require material teacher review rather than production-ready plans. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Tools like ChatGPT and lesson-planning apps exist and are used informally, but no deployed product reliably handles special-education-specific curriculum adaptation at scale in production workflows. |
Prepare lesson materials, bulletin board displays, exhibits, equipment, and demonstrations.
42CI 25–60 · exposure 38 · augmentation 63 · importance 3.0/5 · click for rater detail
Prepare lesson materials, bulletin board displays, exhibits, equipment, and demonstrations.
42| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Schools and special education programs are slower adopters of AI; most remain reliant on manual material preparation with limited automation infrastructure. Adoption remains at pilot stage, with teacher preference for hands-on preparation persisting in most districts. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | K-12 special education settings show moderate AI tool adoption for content creation, but overall education sector adoption of AI for physical classroom prep remains uneven and pilot-stage in many districts. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can meaningfully assist by generating draft content, suggesting design layouts, and automating formatting, allowing teaching assistants to focus on customization and physical assembly. The human remains central to ensuring appropriateness for special education contexts. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI substantially speeds up brainstorming, drafting, and designing lesson materials and visual layouts, meaningfully boosting productivity even though physical assembly remains manual. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | AI can generate text and simple graphics for lesson materials, but preparing physical displays, exhibits, and demonstrations requires manual setup, spatial reasoning, and hands-on assembly that current AI cannot perform end-to-end. Partial automation of content creation does not meet the 50% time-saving threshold for the full task. |
| Task automatability | claude-sonnet-5 | 3/5 | AI can generate lesson content, worksheets, and even design suggestions for bulletin boards, but physical assembly, printing, cutting, and installing displays/equipment requires human physical labor that AI cannot perform. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Special education contexts typically require customized materials tailored to individual student needs, and educators have strong preferences for human-crafted, contextually appropriate displays. Professional judgment and legal liability around student-specific accommodations create friction against full automation. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing or legal barriers prevent using AI to help design lesson materials or displays; this is a low-stakes creative/preparatory task. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | While AI-assisted content drafting reduces some labor, the manual assembly, printing, mounting, and demonstration setup still require significant human time. Overall cost savings are modest compared to the loaded wage of a teaching assistant for the complete task. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI can cheaply generate content and design templates, but the physical labor component (assembling displays, setting up equipment) still requires paid human time, keeping overall cost comparable rather than dramatically lower. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | AI tools can draft lesson content and design templates, but no deployed product reliably handles the full workflow of creating, organizing, and physically preparing diverse classroom materials and demonstrations. Most production use is limited to content suggestions rather than complete material readiness. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Tools like ChatGPT, Canva AI, and educational content generators are widely used to draft materials and visual designs, but they don't handle the physical setup portion of the task at all. |
Maintain computers in classrooms and laboratories, and assist students with hardware and software use.
30CI 25–35 · exposure 25 · augmentation 50 · importance 3.5/5 · click for rater detail
Maintain computers in classrooms and laboratories, and assist students with hardware and software use.
30| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | School IT adoption is slow relative to other sectors; most districts still rely on dedicated technicians or teaching assistants for hands-on support, with limited piloting of AI-driven IT triage in classroom environments. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | K-12 special education support roles are in a low-digitization, high-human-contact sector with minimal AI agent deployment for IT support tasks. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI-powered troubleshooting guides, remote diagnostics, and knowledge bases can assist teaching assistants in resolving common software issues faster and help them triage student problems; however, the augmentation is partial since physical repairs and adaptive student support remain human-dependent. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI chatbots and knowledge bases can help TAs troubleshoot software issues faster, but hardware fixes and hands-on student assistance are not meaningfully augmented. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can troubleshoot some hardware/software issues through chatbots or remote diagnostics, the task requires physical maintenance (e.g., replacing components, managing lab equipment) and in-person student assistance that cannot be fully automated. AI could handle ~20-30% of basic troubleshooting but not hands-on repairs or adaptive student support. |
| Task automatability | claude-sonnet-5 | 2/5 | Basic troubleshooting guidance could be AI-assisted, but physical hardware maintenance and hands-on student support in a special education classroom require in-person presence and adaptability AI cannot replicate. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Schools require human staff physically present for student safety, duty-of-care liability, and direct support; district IT governance, device inventory management, and special education accommodations all depend on trained human judgment and accountability that cannot be fully automated. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement, but the need for physical presence, supervision of special education students, and hardware handling creates practical barriers to full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI support systems (chatbots, remote diagnostics) carry integration and licensing costs, but do not eliminate the need for qualified technical staff or teaching assistants present in the classroom for physical repairs and hands-on student guidance, making total cost comparable or higher. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Physical maintenance and hands-on assistance still require a human present; AI can only reduce time on diagnostic/troubleshooting steps, so overall cost savings are limited. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Limited deployed products reliably perform the full task; remote support tools and knowledge bases exist but lack the physical intervention capability and must be paired with human technicians. Schools rarely deploy autonomous systems for hardware maintenance and student support at scale. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | IT chatbots and remote support tools exist for generic tech support, but no deployed product autonomously maintains classroom hardware or assists special-needs students with devices in real time. |
Use computers, audio-visual aids, and other equipment and materials to supplement presentations.
30CI 25–35 · exposure 30 · augmentation 75 · importance 3.5/5 · click for rater detail
Use computers, audio-visual aids, and other equipment and materials to supplement presentations.
30| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | School districts and special education departments are comparatively slow in adopting AI automation, with adoption primarily in pilot stages; funding constraints, regulatory caution, and institutional resistance to replacing human support roles in special education remain significant. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | K-12 special education is a historically slow-adopting, underfunded, and highly in-person sector with limited AI tool integration at scale. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can meaningfully assist TAs by rapidly generating or curating supplementary materials, organizing digital resources, and suggesting adaptive media—tasks that would otherwise consume preparation time—while the TA retains the critical roles of in-classroom delivery and responsive student interaction. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can meaningfully help teaching assistants generate visual aids, adapt materials for different learning needs, and prepare multimedia content, enhancing the human-led presentation process. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can help select or prepare supplementary materials (audio-visual aids, slides), the actual deployment and real-time integration of these aids during a lesson with special education students—who require adaptive, responsive support—requires human judgment and classroom presence that current AI cannot fully automate end-to-end. |
| Task automatability | claude-sonnet-5 | 2/5 | The physical operation of equipment and in-person supplementing of lessons for special education students requires human presence, adaptability, and real-time responsiveness that AI cannot fully replace, though AI can help create materials used in the process. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Special education operates under strong regulatory frameworks (IDEA, IEPs) that typically require a licensed or certified human to deliver and adapt instructional support; using AI to fully replace this function faces both legal and contractual barriers tied to student protections and documented human accountability. |
| Adoption barriers | claude-sonnet-5 | 3/5 | While not licensed work itself, special education contexts involve IEP compliance, safeguarding, and direct student interaction needs that create organizational and legal friction against full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI tools for content generation have relatively low per-task cost, but the need for human oversight, customization for specific students, and integration into lesson workflows means total cost savings remain modest compared to a human TA's full salary. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | A human aide is still required to physically operate and integrate equipment during instruction, so AI mainly reduces prep costs rather than replacing the labor cost of task execution. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Products exist to generate slides, select media, and organize digital materials, but none reliably handle the dynamic, individualized adaptation required in special education settings where student needs vary significantly and real-time instructional decisions are critical. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | No deployed product autonomously operates classroom AV equipment or supplements live instruction for special education students; existing tools only assist with content creation, not the task itself. |
Tutor and assist children individually or in small groups to help them master assignments and to reinforce learning concepts presented by teachers.
25CI 25–25 · exposure 25 · augmentation 63 · importance 4.1/5 · click for rater detail
Tutor and assist children individually or in small groups to help them master assignments and to reinforce learning concepts presented by teachers.
25| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Educational technology adoption, while growing, remains slow and cautious. School budgets are tight, pilot programs are common but scaling is rare, and institutional resistance to replacing human interaction with special-needs children is substantial. Most adoption remains assistive rather than substitutive. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | K-12 special education is a low-digitization, high human-contact sector with slow AI adoption relative to white-collar industries, though some ed-tech pilots exist. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can meaningfully assist teaching assistants by generating personalized worksheets, explaining concepts in multiple modalities, providing real-time learning analytics, and flagging knowledge gaps—allowing the assistant to focus on relationship-building, behavioral support, and nuanced judgment. Such augmentation can significantly boost productivity. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI tools can generate practice materials, simplify content, or provide supplementary explanations that assist teaching assistants, but the core hands-on tutoring and behavioral support remain human-led. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can generate practice problems and explanatory content, tutoring fundamentally requires real-time assessment of a child's understanding, adaptive questioning, and emotional responsiveness. Current AI systems cannot reliably replicate the individualized diagnosis and dynamic adjustment that defines effective small-group or one-on-one tutoring, especially for children with special education needs. |
| Task automatability | claude-sonnet-5 | 2/5 | Tutoring special education students requires reading nonverbal cues, behavioral management, and adapting to individual disabilities in real time, which current AI cannot do end-to-end; some drill/practice content generation can be automated but not the core interpersonal tutoring. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Strong barriers exist: schools require background-checked, trained staff for child interaction; special education law (IDEA) mandates documented individual education plans and human professional oversight; parental and institutional trust in human educators is high; and potential liability for inadequate learning outcomes discourages full automation. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Special education is governed by IEPs and legal mandates requiring qualified, often certified, human staff to provide individualized support and document compliance, creating strong regulatory and liability barriers. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Integrating and maintaining AI tutoring systems, combined with necessary human oversight to ensure learning outcomes and safety, approaches or exceeds the cost of a teaching assistant wage in most educational settings, especially when accounting for setup, customization, and liability. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | While software tools are cheap per interaction, achieving equivalent quality for special-needs students requires substantial human oversight, narrowing the cost advantage significantly. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | AI-based tutoring tools exist (e.g., conversational tutors, adaptive platforms), but they show material limitations in depth, context-awareness, and handling the behavioral and emotional dimensions common in special education settings. No production system reliably performs the full tutoring task at the quality and personalization expected of a human teaching assistant. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | AI tutoring products (e.g., adaptive learning apps) exist but are not deployed as reliable substitutes for in-person special education support, especially for students with significant behavioral or cognitive needs. |
Observe students' performance, and record relevant data to assess progress.
25CI 25–25 · exposure 25 · augmentation 50 · importance 4.0/5 · click for rater detail
Observe students' performance, and record relevant data to assess progress.
25| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Education sectors, particularly special education, are laggards in AI adoption. While some districts use learning management systems and automated scoring tools, autonomous observation and assessment replacement remains rare in production; most adoption is limited to supplementary data collection. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | K-12 special education is a slow-adopting, in-person, resource-constrained sector where AI tools for data tracking are only beginning to see pilot use. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist by organizing observational data, flagging outliers in performance metrics, and generating preliminary reports that a teaching assistant reviews, moderately improving the efficiency of record-keeping and data synthesis while the human remains responsible for interpretation. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help by transcribing notes, organizing data into progress reports, and flagging patterns, meaningfully assisting documentation while the human remains responsible for observation and judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | AI can capture and analyze some quantitative data (e.g., test scores, attendance patterns), but observation of nuanced student behavior, social-emotional indicators, and context-dependent performance requires human judgment and presence. The task involves real-time classroom observation that is difficult for current AI to fully replace while maintaining equal quality. |
| Task automatability | claude-sonnet-5 | 2/5 | Direct behavioral observation of students in a physical classroom requires human presence and judgment about nuanced behaviors, so AI cannot fully perform the observation; only data logging/summarization portions are automatable. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Special education assessment is heavily regulated under IDEA and state/district policies; qualified educators must conduct observations and sign assessments. Privacy concerns around student data and video recording in classrooms, plus the requirement for human judgment in IEP development, create substantial legal and organizational barriers. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Special education involves legal requirements (IEP compliance, IDEA regulations) and needs trained personnel to interpret student behavior and needs, creating strong regulatory and human-judgment barriers. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Deploying video analysis or data management systems to replace observation and assessment involves significant infrastructure, integration, and ongoing oversight costs that rival or exceed a teaching assistant's loaded wage, especially when quality and liability are factored in. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Human aides must still be physically present to observe and interact with students, so AI can only reduce documentation time slightly, not replace the bulk of the labor cost. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While AI tools exist for basic data logging and test scoring, no widely deployed product reliably performs comprehensive observation and assessment of special education student progress in real classrooms. Most solutions are narrow (e.g., automated grading) and still require substantial human oversight and interpretation. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some products exist for digitizing IEP progress notes or behavior tracking, but reliable automated observation of student performance in real time is not deployed at scale in special education settings. |
Present subject matter to students under the direction and guidance of teachers, using lectures, discussions, supervised role-playing methods, or by reading aloud.
19CI 14–25 · exposure 20 · augmentation 50 · importance 3.9/5 · click for rater detail
Present subject matter to students under the direction and guidance of teachers, using lectures, discussions, supervised role-playing methods, or by reading aloud.
19| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Schools, especially special education programs, remain low-digitization sectors with strong institutional preference for human staff supervision and direct student contact. Adoption of AI for teaching assistant roles is minimal; pilots are rare and cultural resistance is substantial. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | K-12 special education is a low-digitization, in-person, highly regulated sector where AI adoption for direct instruction remains at the pilot or assistive-tool stage. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI could assist a teaching assistant by generating lecture outlines, providing text-to-speech support for read-aloud tasks, or offering differentiated content suggestions—but the human must remain active in real-time presentation, student engagement, and behavioral response. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help prepare lesson materials, generate reading content, or suggest role-play scenarios and discussion prompts, giving moderate support to the teaching assistant's preparation even though live delivery remains human-led. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can generate lecture content and read text aloud, presenting subject matter to special education students requires real-time adaptation, emotional responsiveness, and behavioral management that current systems cannot reliably perform. The task's requirement for "under the direction and guidance of teachers" and engagement with students with diverse learning needs makes full automation infeasible; AI could support preparation but not replace the human presenter. |
| Task automatability | claude-sonnet-5 | 2/5 | Direct instruction and supervised interaction with special education students requires real-time behavioral management, physical presence, and adaptive human judgment that current AI cannot replicate end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Significant barriers exist: special education requires credentialed or supervised staff in most jurisdictions, students need human interaction and behavioral response, and liability concerns are high if an AI system mishandles a vulnerable student's needs. Schools also face organizational and regulatory friction around automating student-facing instruction. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Special education work often involves legal requirements (IEP compliance, supervision by certified teachers, safeguarding requirements) and strong preference/need for human contact with vulnerable student populations. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI infrastructure (LLMs, text-to-speech, integration) costs would likely exceed or approach the loaded wage of a teaching assistant, particularly when accounting for oversight, customization, and the need for human backup to handle student behavioral and learning needs. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | Since AI cannot perform this in-person, supervised teaching task, there is no viable AI substitute, making the human cost the only real option. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed products reliably perform this task end-to-end in special education settings. Text-to-speech and lecture generation exist, but classroom presentation—especially for special education—demands real-time student monitoring, differentiation, and emotional attunement that current AI systems lack in production environments. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | There are no deployed products that autonomously deliver in-person special education instruction to students; existing AI tools are limited to content-generation support, not live task execution. |
Clean classrooms.
15CI 15–15 · exposure 0 · augmentation 0 · importance 3.3/5 · click for rater detail
Clean classrooms.
15| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | K–12 education is a laggard sector for physical automation due to budget constraints, fragmented procurement, and preference for human-performed support services; classroom cleaning automation adoption is minimal. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Physical custodial tasks in education settings show minimal AI/robotic adoption; schools are low-digitization environments for manual labor tasks like this. |
| Augmentation potential | claude-haiku-4-5-20251001 | 1/5 | Current AI and robotics offer no meaningful assistance to human classroom cleaners, as the task is inherently physical and requires no knowledge-based decision support that AI could enhance. |
| Augmentation potential | claude-sonnet-5 | 1/5 | AI offers essentially no meaningful assistance for the physical act of cleaning a classroom. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Cleaning classrooms involves physical manipulation of diverse objects, furniture, and surfaces in unstructured environments. Current AI systems lack the embodied mobility, dexterity, and real-time spatial reasoning required to perform this task end-to-end. |
| Task automatability | claude-sonnet-5 | 1/5 | Physical cleaning of a classroom requires manual dexterity, mobility, and adaptability to varied clutter and spaces that current AI systems and consumer robots cannot handle end-to-end today. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Schools may prefer human contact for classroom environment management, and there are some safety and liability considerations in special education settings, though no hard licensing requirement prevents robotic substitution. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing or legal requirement mandates a human specifically for cleaning, but school policies, safety supervision near children, and lack of viable automated alternatives create practical friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Robotic cleaning systems capable of classroom-level tasks are expensive to purchase, maintain, and deploy, while the human labor cost for classroom cleaning assistants remains relatively low, making AI economically unfavorable. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | Robotic cleaning solutions capable of the full range of classroom tidying tasks do not exist affordably; specialized robots cost more than the marginal labor cost of a teaching assistant performing this task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI or robotic products reliably perform general classroom cleaning at production scale. While research prototypes exist, they cannot operate autonomously in real classroom conditions with the required thoroughness and safety. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product autonomously cleans classrooms comprehensively; existing robotic vacuums address only narrow floor-cleaning subtasks, not full classroom tidying and sanitation. |
Teach socially acceptable behavior, employing techniques such as behavior modification or positive reinforcement.
14CI 5–23 · exposure 13 · augmentation 50 · importance 4.5/5 · click for rater detail
Teach socially acceptable behavior, employing techniques such as behavior modification or positive reinforcement.
14| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Education, especially special education, is a laggard sector for AI automation; adoption remains nascent and experimental. Institutional conservatism, regulatory friction, funding constraints, and parental/educator resistance to automation in this sensitive role slow deployment significantly. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | K-12 special education is a low-digitization, high-touch, in-person sector with minimal AI agent deployment for direct student behavioral instruction. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI could meaningfully assist TAs by suggesting evidence-based behavior strategies, tracking behavioral patterns over time, generating reinforcement schedules, or flagging trends—but the human TA remains the primary executor and decision-maker. This assistive role is useful but not transformative. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help assistants plan behavior modification strategies, generate reinforcement schedules, or suggest techniques, but cannot perform the moment-to-moment human interaction itself. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI could generate behavior modification frameworks and reinforcement schedules, the core task—teaching socially acceptable behavior—requires real-time, context-sensitive interaction, relationship-building, and adaptive responsiveness to individual emotional states that current AI systems cannot reliably execute end-to-end. Partial automation of planning or monitoring is feasible, but autonomous delivery falls well short of the 50% time-saving threshold. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires live, in-person relational interaction with children with disabilities, real-time behavioral judgment, and physical presence that current AI cannot replicate end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Strong barriers exist: special education is heavily regulated (IDEA, IEPs), requires documented accountability and parental involvement, and the legal and ethical responsibility for behavioral interventions typically rests with licensed educators or supervised professionals. Liability asymmetry is high—behavioral mistakes carry psychological and developmental stakes. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Special education law (IEPs, behavior plans), safeguarding requirements, and the need for trained personnel to supervise vulnerable children create strong structural barriers to automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | The cost of building, deploying, and overseeing an AI system to teach behavior modifications, plus the liability overhead for errors in special education contexts, currently exceeds the loaded wage of a teaching assistant, particularly when accounting for necessary human supervision and intervention. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute performing this task, so any AI cost is not comparable to a human aide's wage for the same output. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed product reliably performs this task independently; research prototypes and limited pilots exist for behavioral tracking or guided reinforcement frameworks, but no production system can substitute for a human TA in teaching behavior in real classroom settings. AI chatbots or monitoring tools operate at narrow scope with material limitations in nuance and context. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product autonomously teaches socially acceptable behavior to special-needs students in classrooms; this remains firmly human-delivered work. |
Distribute teaching materials, such as textbooks, workbooks, papers, and pencils, to students.
14CI 5–23 · exposure 8 · augmentation 0 · importance 3.7/5 · click for rater detail
Distribute teaching materials, such as textbooks, workbooks, papers, and pencils, to students.
14| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | K-12 special education settings are slow adopters of automation, with high reliance on human staff, low digital infrastructure maturity, and organizational inertia. No public adoption data shows meaningful displacement of teaching assistants through robotics or automation in this sector. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Special education classroom aide roles involve hands-on physical and interpersonal work with minimal digitization or AI adoption in this specific physical task. |
| Augmentation potential | claude-haiku-4-5-20251001 | 1/5 | This task does not benefit from AI assistance; a teaching assistant either distributes materials manually or an automated system does it. There is no meaningful way AI augments human performance on material distribution in a classroom context. |
| Augmentation potential | claude-sonnet-5 | 1/5 | AI offers essentially no assistance for the physical act of handing out textbooks, workbooks, or pencils to students. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While physically distributing materials requires robotic manipulation and navigation in dynamic classroom environments—still unreliable at scale—the task has no meaningful cognitive component. Current AI cannot reliably perform end-to-end physical distribution to individual students in a classroom setting, though the task itself lacks complexity. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a physical task requiring handling and distributing physical objects to students in a classroom, which current AI systems cannot perform without robotic embodiment far beyond off-the-shelf availability. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Significant barriers exist: educational institutions typically require human interaction in classrooms for duty-of-care and liability reasons, automation of K-12 classroom tasks faces regulatory and parental resistance, and most schools lack robotic infrastructure. Human presence during material distribution also serves supervisory and behavioral-management functions. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing barrier specifically prevents automating material distribution, but the special education context requires human presence for supervision, safety, and behavioral support, creating practical friction against removing the human. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | The capital cost of a reliable classroom robot plus infrastructure far exceeds the loaded wage of a teaching assistant for this routine task. Even accounting for labor, the hardware and maintenance overhead makes substitution economically unviable at current prices. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI solution performing this physical task, so any hypothetical automation (robotics) would be far more expensive than having a human aide simply hand out materials. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product reliably performs autonomous classroom material distribution in production today. This requires embodied robotics (navigation, grasping, social awareness) that exists only in research and limited pilots, not at classroom scale. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed products distribute physical classroom materials to students; this remains purely a human physical action with no robotics deployment in special education classrooms. |
Assist librarians in school libraries.
13CI 5–21 · exposure 8 · augmentation 25 · importance 3.1/5 · click for rater detail
Assist librarians in school libraries.
13| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | School library environments are low-digitization, budget-constrained settings with strong preference for human professional presence; adoption of AI agents in these contexts is minimal. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Education support roles, especially special education aide functions, show slow, uneven AI adoption compared to knowledge-work sectors. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could assist with routine cataloging queries or material recommendations, but the core task (helping students in a safe, human-centered environment) offers limited scope for meaningful productivity augmentation through current systems. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI tools could help with administrative aspects like cataloging suggestions or resource searches, but offer limited assistance for the core physical/interpersonal library support duties. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Assisting librarians in school libraries requires human interaction with students, physical organization of materials, and contextual understanding of user needs—core manual and interpersonal work that current AI systems cannot perform end-to-end with 50% time savings at equal quality. |
| Task automatability | claude-sonnet-5 | 2/5 | This task involves physical library tasks (shelving, checking out books, assisting students in person) that current AI cannot perform end-to-end; only narrow sub-components like cataloging searches could be assisted.“}, |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | School libraries serve minors and require in-person human oversight; there is organizational expectation and policy preference for human library staff, and child-protection liability creates practical barriers to substitution. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No strict licensing requirement, but school settings often require adult supervision and human interaction with children, creating organizational and safeguarding friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | AI systems that could assist with parts of library work (e.g., cataloging suggestions) still require human oversight and integration costs; the task's low skill premium means automation provides no cost advantage over direct human labor. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | Since AI cannot substitute for the physical, in-person components of this task, there is no viable cost comparison favoring AI today. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI product reliably performs the full scope of library assistance tasks (shelving, student interaction, material curation, information literacy support) in production school settings today. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product performs the physical and interpersonal duties of assisting a school librarian; this remains outside current AI product scope. |
Prepare classrooms with a variety of materials or resources for children to explore, manipulate, or use in learning activities or imaginative play.
12CI 5–19 · exposure 8 · augmentation 25 · importance 3.8/5 · click for rater detail
Prepare classrooms with a variety of materials or resources for children to explore, manipulate, or use in learning activities or imaginative play.
12| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | K–12 education remains a low-digitization, human-intensive sector with limited automation adoption. Special education classrooms especially rely on personalized, hands-on teacher judgment and are among the slowest to adopt AI-driven automation. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Education support and physical classroom management sectors show minimal AI adoption for hands-on tasks, especially in special education contexts requiring physical setup. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could assist by suggesting material combinations, developmental activities, or organization schemes via text or image prompts, but the task's heavy reliance on direct manipulation, sensory judgment, and real-time classroom adjustment limits meaningful augmentation to idea generation alone. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI could help suggest activity ideas, generate lesson materials, or create resource lists, but offers minimal assistance with the actual physical arrangement and adaptation of the classroom environment. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Preparing physical classroom materials requires spatial judgment, fine motor handling, and understanding of age-appropriate developmental needs. While AI could generate lists or plans, the actual arrangement and material selection depends on real-time observation of children and classroom constraints that current systems cannot reliably assess end-to-end. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a physical task requiring hands-on arrangement of materials and physical space, which current AI systems cannot perform end-to-end; no time-saving automation exists for the physical setup itself. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | School districts have strong liability and duty-of-care requirements; classroom safety and age-appropriateness of materials are legally and ethically overseen by educators. Human judgment and legal responsibility for children's safety create substantial regulatory and organizational barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 3/5 | While not licensed work per se, it requires physical presence, judgment about individual children's needs (especially in special education), and hands-on classroom knowledge that creates practical barriers to substitution. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Teaching assistants are paid modest wages (~$28k–$35k annually), and the physical, in-person nature of material preparation means AI cannot reduce total cost below human labor when integration, oversight, and any physical automation hardware are factored in. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | AI has no capability to perform physical classroom setup, so there is no viable AI cost comparison—human labor is the only option. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI product today can autonomously prepare physical classroom spaces with materials. This task fundamentally requires embodied presence, manipulation of objects, and real-time responsiveness to a specific learning environment—capabilities that current AI systems lack. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product physically prepares classroom materials or environments; this remains entirely a human physical labor task. |
Organize and label materials and display students' work in a manner appropriate for their eye levels and perceptual skills.
12CI 5–19 · exposure 8 · augmentation 25 · importance 3.1/5 · click for rater detail
Organize and label materials and display students' work in a manner appropriate for their eye levels and perceptual skills.
12| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Special education services are labor-intensive, relationship-driven sectors with limited AI adoption; classroom material organization is deeply embedded in human TA workflows and shows no meaningful automation uptake. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Special education classroom support roles involving physical setup are low-digitization, low-tech-adoption environments with essentially no AI penetration into this specific activity. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could provide suggestions for layout templates or labeling schemes via image recognition and accessibility guidelines, but the task itself is primarily physical and requires human judgment about individual student needs that AI currently augments only marginally. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI could offer minor suggestions (e.g., layout ideas, accessibility guidelines for visual materials) but provides little direct assistance with the physical execution and personalized judgment required. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI could help generate organizational schemes or label designs, the core task requires physical manipulation of materials, spatial judgment of eye levels, and real-time assessment of individual student perceptual abilities that autonomous systems cannot reliably perform end-to-end in a classroom setting. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a hands-on physical task involving arranging materials and displays in a classroom, tailored to specific students' sensory and perceptual needs, which requires physical manipulation and in-person judgment AI cannot perform. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Strong barriers exist due to the requirement for direct physical classroom presence, intimate knowledge of individual student needs, and legal/institutional expectations that qualified educational staff (not automation) manage learning environments and materials for vulnerable populations. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing requirement specifically bars automation, but the need for physical presence, understanding of individual students' perceptual/sensory needs, and classroom customization creates strong organizational and practical friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The physical and spatial components of material organization, coupled with low task frequency and high customization per student, make any robotic or AI-driven solution far more expensive than a teaching assistant performing the work. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | AI has no viable mechanism to perform this physical task, so any hypothetical AI solution (e.g., robotics) would be far costlier and less practical than the human aide already present. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product autonomously organizes physical materials, labels them, and arranges displays in classrooms tailored to diverse special education student needs; this remains a hands-on human task with no comparable automation in production. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product physically organizes classrooms or arranges student work displays; this remains entirely a human physical activity. |
Instruct and monitor students in the use and care of equipment and materials to prevent injuries and damage.
11CI 0–23 · exposure 13 · augmentation 38 · importance 4.0/5 · click for rater detail
Instruct and monitor students in the use and care of equipment and materials to prevent injuries and damage.
11| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Education, especially special education, remains a low-digitization, human-contact-required sector with slow AI adoption. Budget constraints, resistance to surveillance of minors, and union/staffing traditions keep adoption minimal and pilot-focused. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Special education classroom aide roles are low-digitization, high-physical-presence jobs with minimal AI adoption or piloting for this specific safety-monitoring function. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI could meaningfully assist by generating visual safety guides, flagging potential hazards from video feeds for teacher review, and creating personalized equipment care checklists, thereby improving consistency and reducing instructor cognitive load while keeping the human in full control. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI could help create instructional materials or safety checklists in advance, but it offers little real-time assistance during the actual hands-on monitoring and instruction moment. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | AI could assist with creating instructional materials and monitoring video feeds for equipment misuse, but the task requires real-time physical presence, individual safety judgment, and adaptive correction that current AI cannot reliably perform end-to-end. The irreducible human supervision component prevents meaningful time savings at equal safety quality. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires real-time physical presence, hands-on demonstration, and continuous supervision of students (often with special needs) to prevent injury—something current AI cannot perform in a physical classroom setting. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Schools have duty-of-care and liability obligations that legally and practically require a qualified human present for student safety supervision. Regulatory frameworks (special education law, safety codes) and parental expectations create strong barriers to full automation, even if technical capability existed. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Special education settings have strict legal, safety, and duty-of-care requirements mandating direct human supervision and liability accountability, making substitution essentially prohibited. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Integration of AI monitoring systems (cameras, sensors, inference) plus required human oversight adds cost layers that approach or exceed a teaching assistant's wage, particularly when accounting for liability and the need for human backup safety verification. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute performing this physical supervisory task, so any AI cost comparison is moot; human presence is required and cheaper than any hypothetical automation attempt. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While AI can generate safety instructions and assist with video monitoring systems exist in research, no deployed product reliably monitors and corrects student equipment use in real classrooms without human oversight. Current systems have high error rates in understanding physical safety contexts and student intent. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product provides in-person physical instruction and safety monitoring of students with special needs; this remains entirely research-stage or nonexistent for physical supervision tasks. |
Employ special educational strategies or techniques during instruction to improve the development of sensory- and perceptual-motor skills, language, cognition, or memory.
8CI 0–16 · exposure 5 · augmentation 50 · importance 4.0/5 · click for rater detail
Employ special educational strategies or techniques during instruction to improve the development of sensory- and perceptual-motor skills, language, cognition, or memory.
8| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Special education remains highly staffing-constrained and human-dependent; schools are not rapidly deploying AI agents to deliver core instruction. Adoption is limited to supplementary tools (e.g., material generation) rather than substitution of the instructional role itself. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Special education support services are a low-digitization, high human-contact sector with minimal AI agent deployment in direct instructional delivery. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist by suggesting evidence-based techniques, generating differentiated practice materials, and tracking progress data, which may help a teaching assistant plan and prepare lessons. However, the core delivery of sensory-motor and perceptual intervention remains human-centric and not substantially augmented by current systems. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI tools can help generate individualized activity ideas, materials, or track progress data, offering the teaching assistant planning support, though the core interactive instruction remains manual. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires real-time, adaptive instruction tailored to individual sensory and developmental needs, along with physical positioning, tactile feedback, and moment-to-moment responsiveness. Current AI systems cannot reliably deliver the embodied, individualized intervention and on-the-fly adjustment that defines effective special education instruction. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires hands-on, physically present intervention with children with disabilities, using adaptive strategies in real time—far beyond what any AI system can execute end-to-end today. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Special education instruction is heavily regulated (IDEA, IEPs) and typically requires a human educator to assess individual student needs, modify interventions in real time, and document progress for legal compliance. Liability for developmental harm and requirement for human professional judgment create strong adoption barriers. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Special education law (e.g., IDEA/IEP requirements), duty-of-care obligations, and the need for qualified, often certified personnel working directly with vulnerable children create very strong legal and ethical barriers to automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI tools for generating activity suggestions are cheap, but integration with a specialized teaching assistant's workflow, plus human oversight to ensure appropriateness, adds cost. The loaded human wage for a teaching assistant remains lower than the total cost of AI infrastructure plus human validation. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI substitute performing this physical, relational task, so cost comparison favors the human by default since AI cannot deliver the output at all. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While AI can generate instructional materials and suggest techniques, no deployed system reliably executes the sensory-motor, language, or cognitive interventions themselves with the precision and responsiveness required. Pilot systems exist for isolated skill drill, but production systems performingthis task at scale do not exist. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs live sensory, perceptual-motor, or cognitive-development interventions with special education students; this remains entirely human-delivered in practice. |
Discuss assigned duties with classroom teachers to coordinate instructional efforts.
5CI 5–5 · exposure 0 · augmentation 38 · importance 3.9/5 · click for rater detail
Discuss assigned duties with classroom teachers to coordinate instructional efforts.
5| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Education remains a laggard sector in AI adoption, especially for tasks involving core instructional decision-making. Special education coordinatio is governed by law and professional standards that reinforce human-to-human accountability, slowing any automation impulse. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Special education support roles are low-digitization, high-human-contact settings with minimal AI adoption for interpersonal coordination tasks. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could assist by drafting meeting agendas, summarizing prior coordination notes, or suggesting evidence-based strategies, but the core value of discussion—negotiating priorities, resolving conflicts, building shared understanding—remains inherently human and collaborative. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI tools can help by summarizing IEP goals, drafting talking points, or organizing notes ahead of the discussion, aiding preparation even though the conversation itself remains human-led. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires real-time interpersonal communication, context-sensitive problem-solving, and relationship-building between humans with distinct roles and responsibilities. AI cannot meaningfully participate in or replace collaborative discussion that shapes instructional strategy. |
| Task automatability | claude-sonnet-5 | 1/5 | This is an interpersonal coordination task requiring real-time conversation, relationship building, and shared understanding of a specific child's needs; AI cannot conduct this human collaboration end-to-end.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Teachers and special education assistants are expected to work collaboratively as a matter of professional practice and legal obligation under IDEA; schools retain strong organizational and professional norms against removing human coordination from instruction. Liability concerns around delegating instructional decisions to AI also create friction. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Special education involves legal requirements (IEP compliance, staff accountability) and requires direct human coordination between certified/trained staff, creating strong organizational and regulatory friction against automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The task is fundamentally about human professional collaboration; any AI system attempting to mediate or replace this discussion would require expensive integration, oversight, and likely create more work than it saves compared to direct teacher-TA dialogue. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute performing this interpersonal task, so cost comparison favors the human by default. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product can reliably conduct substantive, bidirectional instructional coordination discussions with teachers. This requires genuine understanding of classroom context, student needs, and professional judgment—beyond current AI capabilities in production education settings. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs live in-person coordination discussions between educators about student-specific instructional duties. |
Organize and supervise games and other recreational activities to promote physical, mental, and social development.
4CI 0–9 · exposure 8 · augmentation 25 · importance 3.6/5 · click for rater detail
Organize and supervise games and other recreational activities to promote physical, mental, and social development.
4| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | School systems, particularly special education programs, are characterized by low automation velocity, risk-averse cultures, heavy regulatory oversight, and strong preference for human contact with vulnerable populations. Adoption of replacement automation in this domain remains minimal. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Special education paraprofessional roles involve hands-on physical care and supervision, a sector with minimal AI agent deployment for this kind of task. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could provide limited assistance through activity planning tools or tracking student engagement, but cannot substantially augment the human supervisor's core role of in-the-moment facilitation, safety monitoring, and developmental support during activities. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI could help plan activity ideas or track developmental progress notes, but offers little real-time assistance during actual supervision and play. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI could help plan activities or suggest games, the core task requires real-time supervision, physical presence, and immediate responsiveness to students' needs and safety—elements that current AI cannot perform end-to-end. AI might assist in activity selection but cannot replace the supervisory and developmental facilitation work. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires physical presence, real-time supervision of children with special needs, and hands-on safety monitoring that no current AI system can perform end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Strong legal and regulatory barriers exist: schools have mandatory duty-of-care obligations, special education requires individualized supervision under IDEA, and liability for student safety during activities creates hard requirements for qualified human oversight. Automation would face regulatory and legal constraints. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Special education settings require certified/trained staff for direct supervision of vulnerable students, with strict duty-of-care, safety, and liability requirements that mandate human presence. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The cost of deploying robotic systems or comprehensive AI monitoring for activity supervision would far exceed the loaded wage of a teaching assistant, making this economically unfeasible as a replacement. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | AI cannot substitute for physical supervision at all, so any cost comparison favors the human by default since the task cannot be performed by AI. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI product reliably performs real-time supervision and facilitation of recreational activities for special education students. This task requires embodied presence, safety monitoring, and adaptive human interaction that current systems cannot deliver in production. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product supervises or physically organizes recreational activities for special education students; this remains outside AI's operational domain. |
Carry out therapeutic regimens, such as behavior modification and personal development programs, under the supervision of special education instructors, psychologists, or speech-language pathologists.
3CI 0–5 · exposure 5 · augmentation 50 · importance 4.1/5 · click for rater detail
Carry out therapeutic regimens, such as behavior modification and personal development programs, under the supervision of special education instructors, psychologists, or speech-language pathologists.
3| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Special education and therapeutic services remain among the slowest-adopting sectors for AI replacement due to regulatory requirements, low digitization, small facility scale, and fundamental preference for human-child therapeutic relationships. Adoption is limited to data tracking and planning aids, not service delivery. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Special education and direct care roles are among the least digitized and slowest to adopt AI given the hands-on, supervised, child-safety-critical nature of the work. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can usefully assist by generating behavior tracking summaries, suggesting intervention modifications based on logged data, and organizing progress documentation, reducing paperwork burden on teaching assistants. However, the core therapeutic work remains human-centered. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI tools can help plan behavior modification programs, generate progress tracking, or suggest personalized strategies, but the actual delivery and adaptation in the moment relies on the human aide. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires real-time adaptive interaction with individual children, therapeutic judgment, and emotional attunement that current AI systems cannot reliably replicate. The personalized behavior modification and progress assessment demands human presence and responsiveness that no end-to-end automation can currently match. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires physical presence, real-time relationship building, and hands-on delivery of therapeutic interventions with children with disabilities—no current AI system can execute this end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Strong legal and regulatory barriers exist: therapeutic interventions with special education students typically require licensed professionals (psychologists, speech-language pathologists) or direct supervision, and liability for therapeutic outcomes falls on credentialed humans. School districts face duty-of-care obligations that preclude autonomous AI. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Special education law, IEP compliance, supervision requirements by licensed professionals, and child safety/liability concerns create hard legal and ethical barriers to non-human delivery. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The cost of developing, validating, and deploying AI for therapeutic regimens—including liability coverage and human oversight—exceeds the relatively modest loaded wage of teaching assistants in most school settings. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | AI cannot substitute for the human labor involved, so there is no viable cost comparison—the human is the only option for delivery. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While AI can assist with therapeutic program planning and data logging, no deployed product reliably executes the hands-on delivery of behavior modification or personal development programs with children. Existing systems lack the embodied presence, real-time behavioral calibration, and regulatory compliance for autonomous therapeutic delivery. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product carries out in-person behavior modification or therapeutic regimens with special needs students; this is fundamentally a physical, interpersonal task. |
Participate in teacher-parent conferences regarding students' progress or problems.
3CI 0–5 · exposure 0 · augmentation 38 · importance 3.4/5 · click for rater detail
Participate in teacher-parent conferences regarding students' progress or problems.
3| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Education, particularly special education, is a highly regulated, risk-averse sector with deep institutional preference for human professionals in family-facing roles. Adoption of AI for parent conferencing is virtually nonexistent. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | K-12 special education is a low-digitization, high human-contact sector with minimal AI adoption for interpersonal conferencing tasks. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could modestly assist by pre-generating summaries of student performance data or flagging key academic metrics before a conference, but the core task of participating in the live discussion remains human-driven and AI offers limited enhancement. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help prepare notes, summarize student progress data, or draft talking points beforehand, but cannot conduct the actual conference interaction. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires live interpersonal negotiation, empathy, and real-time judgment about sensitive family circumstances. AI cannot meaningfully conduct or participate in a parent-teacher conference where nuanced discussion and human relationship-building are central. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires physical/virtual presence, real-time relationship building with parents, and sensitive judgment about a child's special needs that AI cannot perform end-to-end today. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Legal and ethical barriers are high: education law often requires qualified personnel to conduct parent meetings, institutional liability is severe if AI mishandles sensitive family information, and parental expectations strongly favor human interaction on matters affecting their child's education. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Special education involves legal requirements (IEP meetings, parental rights, disability law) and strong preference/requirement for human staff presence and accountability. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The task fundamentally requires human presence and accountability; there is no cost-effective AI substitute for a person sitting at a conference table discussing a child's needs with parents. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute performing this task, so cost comparison favors the human by default. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI system can reliably participate as a conferencing agent in this context. While AI can draft notes or summarize records, actually engaging with parents in a live conference remains beyond production capability. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product substitutes a teaching assistant's participation in parent conferences; this remains a human interpersonal task. |
Attend staff meetings and serve on committees, as required.
3CI 0–5 · exposure 0 · augmentation 25 · importance 2.9/5 · click for rater detail
Attend staff meetings and serve on committees, as required.
3| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Educational institutions and school districts have no incentive or ability to replace actual staff meeting attendance with AI; this remains a firmly human-required activity across all sectors. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | K-12 special education is a low-digitization, human-centric sector with minimal AI adoption for governance and administrative participation tasks like committee service. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could assist by preparing meeting agendas, summarizing prior discussions, or drafting follow-up notes, but these are peripheral to the core task of attendance and participation itself. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can help with note-taking, summarizing meeting minutes, or preparing talking points beforehand, but it doesn't meaningfully transform the core act of attending and participating. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires real-time participation in live meetings and committee work that demands contextual judgment, interpersonal presence, and accountability. AI cannot meaningfully replace the human attendance and voice needed in these collaborative settings. |
| Task automatability | claude-sonnet-5 | 1/5 | Physical attendance, participation, and contribution to staff meetings and committees requires human presence, relational judgment, and real-time interpersonal engagement that AI cannot perform end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Staff meetings and committee service typically require an actual licensed professional to be present and represent organizational interests. Legal and governance structures mandate human accountability in these settings. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Organizational and professional norms require the actual staff member to attend and represent themselves; substituting an AI agent for a person's institutional role and voice is not organizationally acceptable. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The task fundamentally requires human presence and participation, so cost comparison is not applicable; AI cannot perform the core function of 'attending' in any meaningful way. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute performing this task, so cost comparison favors the human by default. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product can reliably attend and participate in staff meetings as a substitute human participant. AI could summarize meetings or draft notes, but cannot serve as a committee member or represent an organization in these contexts. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product attends meetings or serves on committees in a person's stead; this remains fundamentally a human presence and participation task. |
Provide assistance to students with special needs.
0CI 0–0 · exposure 0 · augmentation 50 · importance 4.5/5 · click for rater detail
Provide assistance to students with special needs.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Special education departments operate under regulatory compliance and risk-averse institutional constraints. While some schools adopt assistive technology (speech-to-text, visual supports), end-to-end AI replacement of TA roles remains negligible in production; adoption is concentrated in narrow, supervised tool use rather than agent autonomy. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Special education is a highly relational, in-person, heavily regulated field with minimal AI agent deployment for direct student assistance; adoption is essentially nonexistent for this specific task. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist TAs via communication supports, individualized learning content, or data tracking of student progress, improving their efficiency on administrative and instructional planning tasks. However, the irreducibly human aspects of emotional support, behavioral guidance, and physical care limit augmentation's scope relative to the core task. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help teaching assistants by generating individualized materials, suggesting behavioral strategies, or providing communication aids (e.g., AAC tools), meaningfully supporting but not replacing the assistant's direct role. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires real-time, contextual responsiveness to individual students with diverse disabilities and emotional needs—including behavioral de-escalation, physical assistance, and adaptive communication. Current AI cannot reliably provide the in-person, embodied support and relationship-building essential to special education assistance. |
| Task automatability | claude-sonnet-5 | 1/5 | This task involves direct physical, emotional, and behavioral support for students with disabilities, requiring real-time human presence, adaptability, and trust—none of which current AI can perform end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Special education is heavily regulated under IDEA and Section 504; schools have legal obligations to provide qualified human support. Parent expectations, liability concerns around student safety and welfare, and the requirement for trained human judgment in IEP implementation create strong legal and organizational barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Special education law (e.g., IDEA/IEP requirements) mandates qualified human personnel, physical supervision, and safeguarding responsibilities that legally and ethically cannot be delegated to AI. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Deployed AI systems that address parts of special education support (e.g., communication devices, scheduling tools) add cost rather than reduce it; they do not replace the human TA's wage and typically require human oversight and integration labor. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute delivering this in-person care, so the comparison to human wages is not meaningful—AI cannot replace the output at any cost. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI system reliably performs in-person special education assistance at scale. While AI tools can help with lesson planning or communication aids, they do not perform the core task of providing day-to-day, hands-on assistance to students with special needs in educational settings. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product performs in-person special education assistance; existing tools only support narrow adjacent functions like content adaptation or scheduling, not the task itself. |
Supervise students in classrooms, halls, cafeterias, school yards, and gymnasiums, or on field trips.
0CI 0–0 · exposure 0 · augmentation 25 · importance 4.4/5 · click for rater detail
Supervise students in classrooms, halls, cafeterias, school yards, and gymnasiums, or on field trips.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | School adoption of AI for supervision is negligible in production; budget constraints, liability concerns, and preference for human presence in special education settings ensure slow, shallow uptake. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Education support and special education services are a low-digitization, physically-grounded sector with minimal AI agent deployment for direct student supervision. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could assist with post-hoc incident review (flagging video of concerning behavior) or attendance tracking, but the core supervisory task of live student monitoring and intervention offers limited productivity gain from current AI tools. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI tools like monitoring cameras or scheduling aids offer marginal support, but they do not meaningfully enhance the core act of in-person supervision and safety response. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Classroom supervision requires real-time presence, situational awareness, and immediate response to safety risks and behavioral issues—core competencies that current AI cannot provide. This task fundamentally requires human judgment in unpredictable environments with vulnerable minors present. |
| Task automatability | claude-sonnet-5 | 1/5 | Physical supervision of students with special needs in real-world spaces requires in-person presence, safety intervention, and behavioral judgment that no AI system can perform end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Schools face strict duty-of-care legal obligations; a licensed or hired human must remain legally responsible for student supervision. Liability, child-protection law, and regulatory requirements create hard barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Legal, safety, and duty-of-care requirements mandate a responsible human adult supervisor, especially for special education students, making substitution essentially prohibited. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Even if supervisory robots or networked camera systems could substitute, the installation, maintenance, liability insurance, and continuous monitoring infrastructure would far exceed the wage cost of a teaching assistant. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute providing the physical presence and safety response needed, so cost comparison favors the human by default. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product can autonomously supervise students for safety and behavior management in physical spaces. Existing AI is limited to classroom video analysis in research settings; production systems that could replace human supervision do not exist. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs physical, real-time supervision of students; this remains entirely research-stage or nonexistent for embodied care roles. |
Provide students with disabilities with assistive devices, supportive technology, and assistance accessing facilities, such as restrooms.
0CI 0–0 · exposure 0 · augmentation 25 · importance 4.2/5 · click for rater detail
Provide students with disabilities with assistive devices, supportive technology, and assistance accessing facilities, such as restrooms.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Special education services remain labor-intensive and bound by regulatory requirements for qualified human staff. Adoption of AI in this context is minimal; sectors with high digitization have not meaningfully displaced this function. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Special education support roles involving physical care are among the least digitized and slowest to see any AI-driven change given the hands-on nature of the work. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could modestly assist with scheduling accommodations or documenting device usage, but offers minimal productivity gain in the core task of physically assisting students, operating devices with them, and providing real-time in-facility support. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI-based assistive technologies (e.g., communication devices, adaptive software) can support the broader work of teaching assistants, but this specific task of physical facility/device assistance sees minimal AI augmentation. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires physical presence, real-time responsiveness to individual student needs, direct hands-on assistance, and contextual judgment about appropriate accommodation timing and methods. Current AI systems cannot physically assist students, access facilities, or operate in dynamic classroom environments. |
| Task automatability | claude-sonnet-5 | 1/5 | This task involves direct physical assistance, mobility support, and hands-on help with assistive devices and restroom access, none of which current AI systems can perform without embodiment. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Legal and regulatory barriers are substantial: special education is governed by the Individuals with Disabilities Education Act (IDEA), which mandates qualified human personnel; schools have liability obligations; and human contact, supervision, and judgment are legally required components of special education services. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Physical assistance to vulnerable students, especially involving restroom access, requires trained, often certified personnel, with strict safeguarding, liability, and legal requirements around student care and safety. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The loaded wage of a teaching assistant is modest, and any attempt to replace this function would require robotics capable of safe physical interaction with vulnerable students—a far more expensive and immature technology than employing humans. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI substitute for this physical task, so any AI cost comparison is moot—human labor is the only viable option currently. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI product can autonomously provide physical assistive devices, accompany students to facilities, or deliver real-time in-person support to students with disabilities. This task is fundamentally embodied and requires human presence in a physical space. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product provides physical assistance to students with disabilities or helps them access facilities; this remains firmly in the human physical caregiving domain. |
Enforce administration policies and rules governing students.
0CI 0–0 · exposure 0 · augmentation 25 · importance 4.0/5 · click for rater detail
Enforce administration policies and rules governing students.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Schools are slow to digitize and have not adopted AI for student discipline or behavior enforcement, given regulatory, liability, and safeguarding concerns. Adoption remains negligible. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | K-12 special education is a low-digitization, high human-contact sector with minimal AI deployment for behavioral enforcement tasks. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could potentially assist by analyzing patterns in discipline referrals or flagging policy violations in documented behavior logs, but the core task of in-the-moment enforcement and judgment remains human-centric with limited augmentation potential. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI could help track incidents or policy references for staff, but it offers little real-time assistance for the interpersonal enforcement itself. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Enforcing policies and rules requires real-time judgment about context, student behavior, safety, and appropriateness of interventions. Current AI systems cannot reliably perceive physical classroom dynamics, assess nuance in student interactions, or make moment-to-moment decisions that humans delegate to on-site staff. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires real-time, in-person supervision, judgment about student behavior, and physical presence in a classroom with children with special needs, which current AI cannot perform end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | School discipline and student supervision are legally and organizationally human responsibilities requiring certified or employed staff with accountability. Parents, students, and districts expect a human adult in the role, and liability for disciplinary decisions rests with the institution's human agents. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Enforcing rules with special education students involves legal responsibility, safeguarding duties, and required human authority/certification, making this a hard barrier to automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | AI systems cannot replace the human presence needed for this task; any attempted automation would still require human oversight and intervention, making AI additive rather than substitutive in cost. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute performing this task, so any AI cost would be additive rather than replacing the human's wage. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI product can autonomously monitor and enforce school discipline policies in real classrooms. The task is inherently synchronous, requires human presence, and demands legal/pastoral authority that AI cannot exercise. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product enforces school policies or manages student behavior autonomously; this remains a human-only responsibility in practice. |
Instruct students in daily living skills required for independent maintenance and self-sufficiency, such as hygiene, safety, or food preparation.
0CI 0–0 · exposure 0 · augmentation 25 · importance 3.8/5 · click for rater detail
Instruct students in daily living skills required for independent maintenance and self-sufficiency, such as hygiene, safety, or food preparation.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Special education is a highly regulated, human-intensive sector with strong requirements for certified staff presence and individualized instruction. Adoption of AI to replace instructional roles remains negligible; schools continue to hire and deploy human teaching assistants. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Special education classrooms are a low-digitization, high physical-presence sector with minimal AI agent deployment for direct student instruction of life skills. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could marginally assist with content preparation (e.g., generating visual schedules, creating training videos) or tracking progress, but offers minimal augmentation to the core task of live, adaptive instruction and hands-on skill coaching. The human teaching assistant remains the sole effective actor. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI could help create visual schedules, social stories, or lesson plans for teaching these skills, but it offers little assistance during the actual hands-on instructional moment. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires real-time, in-person physical demonstration, hands-on guidance, behavioral reinforcement, and adaptive response to individual student needs and disabilities. Current AI systems cannot physically demonstrate hygiene or food preparation, nor can they safely supervise students or provide the embodied, context-sensitive instruction this population requires. |
| Task automatability | claude-sonnet-5 | 1/5 | This task requires hands-on, physical demonstration, modeling, and real-time behavioral intervention with students who often have disabilities requiring individualized physical and emotional support that current AI cannot provide. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | This task involves direct instruction of minors with disabilities, duty of care, potential liability for harm, and often legal requirements for certified or qualified personnel to supervise and instruct. Schools face regulatory and liability constraints that strongly protect against substitution by automation. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Special education law (e.g., IDEA) mandates qualified personnel for student support, and safety-critical, hands-on skill instruction for vulnerable populations requires human presence, supervision, and legal accountability. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | AI systems cannot perform this task end-to-end, so direct cost comparison is moot. The per-unit cost of any AI-adjacent tools (video content, chatbots) would not offset the need for a human instructor present for safety, demonstration, and personalized support. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI substitute for this task, so the human cost is the only viable cost; any AI attempt would require robotics and supervision far exceeding the aide's wage. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI product reliably performs live instruction in daily living skills for special education students. While AI can generate educational content or answer questions, it cannot replace the interactive, adaptive, physical presence and real-time behavioral coaching that this task demands. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product performs in-person instruction of hygiene, safety, or food preparation skills to special education students; this remains entirely a human-delivered task. |
Assist in bus loading and unloading.
0CI 0–0 · exposure 0 · augmentation 0 · importance 3.7/5 · click for rater detail
Assist in bus loading and unloading.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | School districts operate in a heavily regulated, tradition-bound sector with low digitization of physical student-handling processes and strong legal constraints on substituting human supervision. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Special education support and school transportation are low-digitization, physically embodied service sectors with minimal AI/robotics adoption for direct child supervision tasks. |
| Augmentation potential | claude-haiku-4-5-20251001 | 1/5 | AI offers no meaningful assistance in physically loading/unloading students or in real-time safety monitoring during bus operations; the task requires human judgment and presence throughout. |
| Augmentation potential | claude-sonnet-5 | 1/5 | AI offers essentially no assistance to a human performing hands-on physical supervision during bus loading and unloading. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Bus loading and unloading of students requires physical presence, real-time judgment about individual needs, safety monitoring, and direct supervision of children—no current AI system can perform these physical and interpersonal requirements end-to-end. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a physical supervision and safety task requiring real-time presence to assist students with disabilities on/off a bus; current AI cannot perform physical assistance or supervision.rated no meaningful automation potential exists today. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Schools face hard legal and liability barriers: staff must physically supervise children boarding/exiting buses, duty-of-care laws require licensed/authorized personnel, and no regulatory framework permits automation of student safety on school transport. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Safety regulations, duty-of-care requirements, and legal liability for supervising children with disabilities during transport make human presence mandatory, creating a hard barrier. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | This task has minimal current automation potential, making cost comparison moot; human labor remains the only viable approach for the foreseeable future. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI substitute performing this physical task, so cost comparison favors the human by default; any robotic solution would be far more expensive than a low-wage aide. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product performs autonomous student bus loading/unloading; this task fundamentally requires human physical presence and real-time in-person supervision for legal and safety reasons. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product performs physical child supervision or mobility assistance during bus loading; this remains entirely outside current product capability. |
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.