Emergency Medical Technicians
29-2042.00Median wage $44,470/yr180,510 employed (US)Rank #904 of 923 scored · top 98% by substitution
Assess injuries and illnesses and administer basic emergency medical care. May transport injured or sick persons to medical facilities.
Sub-scores
0–100 · band = confidence interval from rater disagreement
Substitution8
Exposure7
Augmentation36
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
12 rated tasks, binned by substitution score.
Position among all scored occupations
Distribution of 923 occupation scores; the marker is this occupation.
Tasks with substitution ≥ 70
0%
Run 1.0.0-draft.1 · computed 2026-08-05 · rater panel: claude-sonnet-5, claude-haiku-4-5-20251001 · intervals span rater disagreement.
Why this score
The five weighted dimensions of the composite, averaged across this occupation's tasks (importance-weighted, panel mean). Exact weights and formulas: /api/v1/methodology.
Task automatabilityw 35%7
panel mean rating 1.3/5 → substitution pressure 7/100
Technical feasibility todayw 20%7
panel mean rating 1.3/5 → substitution pressure 7/100
Cost vs. human wagew 15%6
panel mean rating 1.3/5 → substitution pressure 6/100
Adoption barriersw 20%inverted — strong barriers lower the score10
panel mean rating 4.6/5 (barrier strength) → substitution pressure 10/100
Sector adoption velocityw 10%7
panel mean rating 1.3/5 → substitution pressure 7/100
Task breakdown (12 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.
Maintain vehicles and medical and communication equipment, and replenish first aid equipment and supplies.
30CI 25–35 · exposure 25 · augmentation 50 · click for rater detail
Maintain vehicles and medical and communication equipment, and replenish first aid equipment and supplies.
30| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Emergency services adoption of automation remains slow; many departments rely on traditional maintenance schedules and in-house technicians. Inventory management software exists but hasn't driven meaningful displacement of maintenance technician roles. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | EMS agencies are a low-digitization, resource-constrained sector with slow technology adoption for logistics and maintenance tasks, though basic inventory apps are gradually being introduced. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can meaningfully assist by tracking inventory levels, predicting maintenance needs, and scheduling service windows, helping technicians work more efficiently. However, the core task of hands-on equipment maintenance and repair remains human-centered. |
| Augmentation potential | claude-sonnet-5 | 3/5 | Inventory management apps, predictive maintenance alerts, and digital checklists can meaningfully streamline tracking and reordering supplies, improving efficiency even though physical tasks remain human-performed. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Maintaining vehicles and equipment requires physical inspection, hands-on repairs, and judgment about when replacement is needed. While AI could assist with inventory tracking and scheduling, end-to-end automation of mechanical maintenance and equipment diagnostics remains beyond current capability without significant human oversight. |
| Task automatability | claude-sonnet-5 | 2/5 | Physical inspection, cleaning, restocking supplies and vehicle maintenance require manual handling in a real ambulance and cannot be done end-to-end by current AI; software can only assist with tracking/scheduling.dev |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Equipment maintenance in emergency services is constrained by regulatory compliance, vehicle certification requirements, and legal liability if systems fail during emergency response. Most jurisdictions require licensed or certified technicians to sign off on vehicle maintenance and medical equipment readiness. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement specifically for restocking/maintenance, but some equipment checks (e.g., medical device functionality) may have compliance/safety protocols requiring trained personnel sign-off. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI-assisted inventory management and scheduling systems exist at modest cost, but cannot replace the technician labor required for actual vehicle maintenance, equipment repair, and supply chain execution. Total cost savings remain marginal compared to technician wages. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI-based inventory/maintenance tracking software is cheap, but it doesn't replace the labor of physically restocking and inspecting equipment, so overall cost savings versus a human EMT doing this are limited. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed AI system can autonomously maintain vehicles, diagnose medical equipment failures, or perform physical repairs. AI tools can track inventory and flag supplies needing replenishment, but actual maintenance execution and equipment servicing require human technicians. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Inventory management software and fleet maintenance tracking systems exist and are used, but the actual physical replenishment and equipment checks remain manual, so no deployed product performs the full task. |
Observe, record, and report to physician the patient's condition or injury, the treatment provided, and reactions to drugs or treatment.
25CI 25–25 · exposure 25 · augmentation 50 · click for rater detail
Observe, record, and report to physician the patient's condition or injury, the treatment provided, and reactions to drugs or treatment.
25| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | While some EMS agencies pilot automated vital-sign logging, the sector as a whole (many small rural and municipal services) lags in adoption of AI integration; production deployment of AI-driven observation and reporting remains rare. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | EMS and prehospital care are physical, low-digitization environments with slow AI tool adoption relative to hospital-based clinical documentation systems. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can usefully assist EMTs by auto-logging vitals, flagging abnormal values, and drafting structured reports, thereby reducing administrative burden and improving data quality, though the core clinical observation remains human-driven. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-assisted transcription, symptom-checklists, and decision-support tools can help EMTs structure and communicate observations more efficiently, though the human remains central to sensing and judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can assist with structured data recording and flagging vital signs, the task fundamentally requires real-time clinical observation, judgment about injury severity, and nuanced patient assessment that current AI systems cannot reliably perform end-to-end in the field without continuous human oversight. |
| Task automatability | claude-sonnet-5 | 2/5 | Documentation and reporting can be partly assisted by AI (voice-to-text, structured note generation), but real-time patient observation, clinical judgment about condition/injury severity, and verbal handoff communication under field conditions require human sensory and cognitive presence. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Legal and liability barriers are substantial: an EMT's observations and reports are medico-legal documents; malpractice risk, regulatory requirements (state EMT licensure), and physician reliance on human judgment create strong institutional friction against full automation. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Clinical reporting to a physician involves licensure, liability, and legal documentation requirements that mandate a certified human EMT perform and attest to the observation and reporting. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Integration of AI monitoring and documentation systems into ambulances has meaningful setup and oversight costs that may reduce but not substantially undercut the loaded wage of an EMT performing observation and reporting. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI transcription/documentation tools are cheap, but the core task—physical observation and judgment-based reporting—still requires the EMT, so cost savings apply only to a small documentation slice. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Deployed products can capture and log vital signs automatically, but no production system reliably observes complex injuries, detects subtle treatment reactions, or generates clinically accurate reports without substantial EMT intervention and verification. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Ambient documentation tools exist in clinical settings (e.g., ED scribes) but are not widely deployed in prehospital EMS environments for real-time field reporting to physicians with reliability. |
Assess nature and extent of illness or injury to establish and prioritize medical procedures.
13CI 0–25 · exposure 13 · augmentation 38 · click for rater detail
Assess nature and extent of illness or injury to establish and prioritize medical procedures.
13| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Emergency medicine and pre-hospital care remain heavily regulated, conservative sectors with strong human-contact requirements. Adoption of AI triage support is slow; most systems remain pilot-stage in academic or large urban systems rather than widespread production deployment. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | EMS is a physical, low-digitization field-based sector with minimal AI agent deployment for autonomous clinical decision-making; adoption is largely limited to documentation and dispatch support. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI-powered symptom and vital-sign analysis tools can assist EMTs by suggesting differential diagnoses and flagging critical findings, improving decision speed and consistency. However, augmentation is partial because the EMT's physical examination and scene judgment remain irreplaceable. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can assist with protocol lookup, documentation, and decision-support suggestions, but it plays a minor role compared to the hands-on sensory and judgment-based nature of the assessment itself. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can assist with symptom analysis and triage protocols, the task requires real-time physical examination, vital sign collection, and dynamic clinical judgment that cannot be fully automated. Current AI lacks the embodied sensorimotor capability and cannot independently examine patients in chaotic field conditions, leaving only narrow diagnostic support possible. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires real-time, hands-on physical assessment of a patient in unpredictable field conditions, involving palpation, auscultation, vital signs, and rapid triage decisions that current AI cannot perform physically or perceptually end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Strong regulatory and liability barriers exist: EMTs are licensed professionals, malpractice liability is high, medical protocols require human accountability, and patients in emergency settings legally and ethically expect human clinical judgment. No autonomous system could substitute without explicit regulatory approval per clinical guidelines. |
| Adoption barriers | claude-sonnet-5 | 5/5 | EMS protocols legally require a certified/licensed EMT or paramedic to perform patient assessment and triage, with strict liability, certification, and scope-of-practice regulations. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Current AI systems that assist with triage require integration infrastructure, oversight, and validation; combined with the need for human EMTs to remain on scene for physical care, the all-in cost savings are minimal compared to a trained EMT's loaded wage. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI system capable of independently performing this task, so no meaningful cost comparison favors AI; any assistive tool adds cost on top of the required human EMT. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Deployed AI products exist for symptom checkers and triage support, but none reliably perform independent patient assessment in emergency field settings at production scale. Medical diagnosis AI remains research-heavy with significant error rates, and no mature system can replace the EMT's physical examination and real-time prioritization. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs autonomous field triage and prioritization of emergency medical procedures; clinical decision-support tools exist only as advisory aids to a human EMT, not substitutes. |
Attend training classes to maintain certification licensure, keep abreast of new developments in the field, or maintain existing knowledge.
13CI 0–25 · exposure 13 · augmentation 63 · click for rater detail
Attend training classes to maintain certification licensure, keep abreast of new developments in the field, or maintain existing knowledge.
13| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | This is a mandated, human-centric compliance activity in a highly regulated sector. No adoption of AI automation is possible because the task is legally anchored to human participation and credential maintenance. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | EMS and healthcare training sectors are slower adopters of AI-driven learning platforms compared to information/finance sectors, though some blended e-learning is emerging. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI could assist by curating learning materials, summarizing new developments, or organizing continuing education content, but the actual attendance and knowledge acquisition must remain with the human. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can meaningfully assist via personalized quizzes, simulation scenarios, adaptive learning modules, and rapid access to updated medical guidelines, enhancing study efficiency. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task fundamentally requires human participation in formal training and certification processes. AI cannot attend classes, complete accreditation requirements, or satisfy the legal/regulatory mandate for an EMT to personally maintain their license. |
| Task automatability | claude-sonnet-5 | 2/5 | AI can deliver e-learning content, quizzes, and study aids but the actual attendance, hands-on skills practice, and certification testing require physical human participation, limiting full automation of the task.6 |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Regulatory and legal barriers are absolute: EMTs are required by law to maintain active certification through accredited training programs. Only the licensed individual can fulfill this requirement; no automation or delegation is legally permitted. |
| Adoption barriers | claude-sonnet-5 | 4/5 | EMT certification is governed by state/national licensing bodies (NREMT) requiring accredited coursework, practical skill verification, and proctored exams that AI cannot legally substitute for. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | This task has no meaningful cost comparison to AI, since the human must perform it. The human cost (time in training) is unavoidable and non-substitutable. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI-based study aids are cheap, but certified training still requires accredited instructors, proctored exams, and hands-on simulation labs, keeping overall costs comparable to traditional training delivery. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product can complete certification maintenance for a human; this is a human-centric compliance and learning activity by definition. AI cannot obtain or renew credentials on behalf of licensed practitioners. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Online CE platforms and AI tutoring tools exist for medical training, but few products autonomously deliver certified EMT training end-to-end without instructor oversight and hands-on skill verification. |
Drive mobile intensive care unit to specified location, following instructions from emergency medical dispatcher.
4CI 0–9 · exposure 8 · augmentation 25 · click for rater detail
Drive mobile intensive care unit to specified location, following instructions from emergency medical dispatcher.
4| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Emergency medical services remain highly regulated, conservative, and reliant on human judgment and physical presence. Adoption of autonomous driving in this sector is negligible; pilot programs are virtually nonexistent in production EMS fleets. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Emergency medical services are a physically-demanding, safety-critical sector with minimal AI/autonomous vehicle adoption; this is a laggard use case even within a laggard sector. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | GPS navigation and real-time traffic optimization could assist drivers, but the core task—safe emergency operation under dispatcher control—offers limited room for AI assistance without the human remaining fully in active control of the vehicle. |
| Augmentation potential | claude-sonnet-5 | 2/5 | GPS navigation and dispatch-routing software already assist drivers with route optimization, but this offers only modest incremental help to the core driving task itself. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Current AI can handle some navigation and route planning, but cannot reliably handle real-time emergency driving decisions, pedestrian/obstacle detection, vehicle control under high-stress conditions, or communication with dispatch. The safety-critical nature and unpredictability of emergency driving prevent end-to-end automation meeting the 50% time-saving bar. |
| Task automatability | claude-sonnet-5 | 1/5 | Driving an emergency vehicle through traffic under time pressure while responding to dispatcher instructions requires physical control and real-time judgment that current AI/autonomous systems cannot reliably perform in this safety-critical context. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Regulatory, liability, and safety barriers are insurmountable: vehicles carrying patients in critical condition must have human operators legally responsible for safe transport. Insurance, medical licensing, and emergency service protocols all mandate human control and accountability. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Emergency vehicle operation requires licensed, trained personnel, involves major liability exposure, and is heavily regulated, with legal requirements for human operators in emergency response vehicles. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | A trained EMT driver costs $40–50k annually in salary. Autonomous vehicle systems capable of safe emergency driving would cost far more in hardware, insurance, liability coverage, and oversight infrastructure than employing human drivers today. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute performing this function, so cost comparison favors the human driver by default; any autonomous vehicle solution would require far more expensive specialized hardware and validation. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed autonomous vehicles reliably perform emergency medical transport in production. While some self-driving systems exist in limited settings, none have demonstrated safe, consistent operation for urgent dispatch-driven mobile ICU transport with medical personnel aboard. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product autonomously drives ambulances to emergency scenes; autonomous vehicle technology remains research/limited-pilot stage even for standard driving, let alone emergency response driving. |
Administer first aid treatment or life support care to sick or injured persons in prehospital settings.
4CI 0–7 · exposure 5 · augmentation 50 · click for rater detail
Administer first aid treatment or life support care to sick or injured persons in prehospital settings.
4| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | EMS systems are geographically dispersed, resource-constrained, and heavily regulated; adoption of AI-augmented tools (e.g., dispatch optimization, protocol reminders) is slow and fragmented, with no evidence of material displacement of EMT roles. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | EMS is a physical, low-digitization field with essentially no AI displacement of hands-on emergency care occurring in practice. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can support EMTs through real-time protocol reminders, symptom checkers, and dispatch optimization, modestly raising effectiveness; however, the core physical and interpersonal demands of patient care limit how much AI can enhance productivity without the human fully engaged. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can assist via decision-support tools, triage protocols, dispatch optimization, and documentation, but does not materially transform the moment-to-moment physical care delivery. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Administering first aid and life support requires physical intervention (bandaging, CPR, airway management, medication delivery), real-time clinical judgment under uncertainty, and rapid adaptation to uncontrolled environments—tasks that current AI cannot perform end-to-end without human execution of critical physical actions. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires hands-on physical intervention (CPR, wound care, airway management) in unpredictable field environments—no current AI system can perform physical emergency medical care. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Legal and regulatory frameworks strictly require licensed EMTs and paramedics to deliver prehospital care; liability, licensure, and duty-of-care standards create hard barriers that prevent autonomous or AI-only service delivery. |
| Adoption barriers | claude-sonnet-5 | 5/5 | EMT certification, medical licensure, liability for patient outcomes, and legal scope-of-practice rules mandate a credentialed human perform this care. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | AI systems cannot physically administer care, so the cost comparison is not applicable; human EMTs remain essential and their loaded wages exceed the cost of decision-support software, making AI a complement rather than substitute. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute performing this physical 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 assist with diagnostic decision-support and protocol guidance in controlled settings, no deployed product reliably performs actual first aid or life support delivery in prehospital emergency conditions; current systems lack the embodied capability and real-world reliability required. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product provides physical prehospital emergency care; this remains purely a research/robotics frontier far from field deployment. |
Comfort and reassure patients.
3CI 0–5 · exposure 0 · augmentation 13 · click for rater detail
Comfort and reassure patients.
3| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | EMS and emergency healthcare remain heavily regulated, human-contact-dependent sectors with slow AI adoption; there is minimal observed displacement of reassurance tasks by AI systems in production. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | EMS is a physical, high-touch, low-digitization field with minimal AI adoption for direct patient interaction and comfort during emergencies. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could marginally assist EMTs by providing talking points or suggested phrases, but current systems offer minimal meaningful productivity enhancement for a task requiring authentic human presence and emotional attunement. |
| Augmentation potential | claude-sonnet-5 | 1/5 | AI offers essentially no real-time assistance for calming or reassuring a patient in an active emergency response context. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Comforting and reassuring patients requires genuine emotional connection, empathetic presence, and contextual human judgment that current AI systems cannot meaningfully replicate. This task is fundamentally relational and cannot be automated with the 50% time-saving threshold met. |
| Task automatability | claude-sonnet-5 | 1/5 | Comforting and reassuring a frightened or injured patient in person requires physical presence, tone, touch, and real-time human empathy that no AI system can deliver during an emergency response.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Strong legal, regulatory, and ethical barriers protect this task: patients expect and often require human contact during emergencies, liability for emotional harm would fall on organizations deploying automated systems, and professional standards mandate human presence. |
| Adoption barriers | claude-sonnet-5 | 4/5 | While not formally licensed as a discrete task, patient reassurance during emergency care is embedded in EMT scope of practice and requires physical presence and trained interpersonal judgment, creating strong practical barriers to substitution. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The cost of attempting to automate this task through AI agents would exceed the cost of human EMTs performing it, given the limited feasibility and high error-cost asymmetry of AI-only reassurance. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI substitute performing this task, so cost comparison favors the human by default since the AI alternative doesn't exist for this in-person function. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product reliably performs genuine patient comfort and reassurance; such systems exist only in research or limited prototype form and cannot substitute for human presence in real emergency medical settings. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product provides in-person emotional reassurance to patients during EMS calls; this remains entirely a human function performed on-scene. |
Decontaminate ambulance interior following treatment of patient with infectious disease, and report case to proper authorities.
3CI 0–5 · exposure 0 · augmentation 25 · click for rater detail
Decontaminate ambulance interior following treatment of patient with infectious disease, and report case to proper authorities.
3| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | EMS operates in fragmented, resource-constrained settings with minimal digital automation; decontamination is a low-volume, high-stakes physical task in an industry with slow technology adoption and strong human-oversight norms. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | EMS and emergency medical services are a physically-oriented, low-digitization sector with minimal AI adoption for hands-on tasks like vehicle decontamination. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could assist with generating case reports for regulatory submission or reminding technicians of decontamination protocols, but offers minimal productivity gain given the task's straightforward, legally-mandated nature and the small time budget available for augmentation. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI could help by generating or auto-filling the required case report to authorities based on structured inputs, but offers no assistance with the physical decontamination process itself. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Decontamination requires physical manipulation in a confined, variable space (ambulance interior) with safety-critical pathogen handling that demands real-time judgment about contamination zones. No current AI system can autonomously perform the physical cleaning task or reliably assess whether decontamination is complete. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires physical cleaning of a vehicle interior with specialized disinfection procedures, which current AI systems cannot perform as they lack physical embodiment for such tasks.the reporting component could be partially assisted but the core decontamination is physical labor. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Hard regulatory barriers: health codes, occupational safety regulations, and disease-reporting statutes require documented human accountability for both decontamination completion and case reporting to public health authorities. A licensed medical professional must legally certify these actions. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Infectious disease decontamination is governed by strict health and safety regulations, OSHA/CDC protocols, and reporting to public health authorities often has legal mandates requiring human accountability and certification. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | AI systems capable of physical decontamination would require robotics hardware orders of magnitude more expensive than the labor cost of a technician performing the cleaning; human decontamination remains the cheaper option. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI substitute for the physical cleaning task, so AI cost is not applicable/comparable; any robotic cleaning solutions would be far more expensive than human labor for this task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI product performs autonomous ambulance decontamination in production. The task combines hazmat-level safety requirements, variable surface geometry, and health-code compliance that remain entirely dependent on human or specialized equipment execution. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product performs physical ambulance decontamination; this remains purely a manual task performed by EMTs following infection control protocols. |
Communicate with dispatchers or treatment center personnel to provide information about situation, to arrange reception of survivors, or to receive instructions for further treatment.
0CI 0–0 · exposure 0 · augmentation 38 · click for rater detail
Communicate with dispatchers or treatment center personnel to provide information about situation, to arrange reception of survivors, or to receive instructions for further treatment.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Emergency medical services remain highly regulated, human-centric sectors with strong legal requirements for human accountability. Adoption of AI for direct dispatch communication is virtually non-existent in production, and regulatory barriers make rapid adoption unlikely. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | EMS is a physical, low-digitization emergency response sector with minimal AI deployment for live field-to-hospital communication tasks. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI might assist EMTs by drafting or suggesting key information points to communicate or by transcribing radio traffic, but the interactive nature of dispatch communication—receiving dynamic instructions and coordinating care—requires human judgment and real-time responsiveness that AI tools offer only marginal support for. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-enabled dispatch systems, voice-to-text documentation, and decision-support tools can assist by pre-populating information or summarizing situations, but the core live communication remains human-driven. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task fundamentally requires real-time two-way communication with humans (dispatchers, treatment center personnel) to convey dynamic, high-stakes situational information and receive contextual instructions. Current AI cannot autonomously conduct such critical interactive communication in emergency settings where judgment and adaptability are essential. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires real-time, high-stakes verbal communication under dynamic field conditions with physical presence at an emergency scene; no AI system can perform this end-to-end today.5, |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | This task has hard regulatory and liability barriers: emergency dispatch and patient care coordination are governed by federal and state regulations that require licensed personnel to handle critical communications. Malpractice and error-cost asymmetry are severe, and legal authority to direct emergency response cannot be delegated to AI. |
| Adoption barriers | claude-sonnet-5 | 5/5 | EMTs are licensed medical responders whose real-time clinical communication and decision relay is legally and professionally mandated to be performed by certified personnel, with high liability for errors. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Implementing AI to replace human dispatch communication would require extensive infrastructure, integration with emergency systems, and liability oversight—making it more expensive than the current cost of human dispatchers and EMT communication, which is already integrated into existing systems. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute performing this task independently, so cost comparison favors the human by default since AI cannot deliver the output at all. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product reliably performs autonomous emergency dispatch communication or treatment coordination. While AI can draft messages or transcribe speech, it cannot replace the interactive, decision-making dialogue required to arrange patient reception and receive treatment directives in real-time emergency contexts. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product autonomously communicates with dispatchers/hospitals on behalf of an EMT to coordinate patient care; this remains firmly human-performed. |
Coordinate work with other emergency medical team members or police or fire department personnel.
0CI 0–0 · exposure 0 · augmentation 25 · click for rater detail
Coordinate work with other emergency medical team members or police or fire department personnel.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Emergency services remain one of the most conservative and human-centric sectors; adoption of AI for core coordination functions is negligible and faces strong institutional, legal, and safety resistance. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Emergency medical services are a physical, safety-critical, low-digitization sector with minimal AI agent adoption for real-time field coordination. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | Current AI could marginally assist with information routing or radio logging, but coordination inherently demands human presence, judgment, and accountability, limiting meaningful augmentation potential. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI-enabled dispatch systems, radios, and CAD software can support situational awareness and communication logistics, but they only marginally assist the core interpersonal coordination task itself. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Coordination of emergency personnel requires real-time human judgment, situational awareness, communication adaptation, and authority-based decision-making that current AI cannot reliably perform end-to-end in high-stakes environments. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires real-time physical presence, verbal coordination under dynamic emergency conditions, and split-second decision-making among human responders at a scene, which current AI cannot perform end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Emergency response coordination is legally and operationally required to involve licensed/credentialed personnel; liability, chain-of-command accountability, and regulatory requirements mandate human decision-makers in coordination roles. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Emergency response coordination involves licensed personnel, legal chain-of-command, liability for patient and public safety, and mandated protocols requiring human judgment and authority. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | AI coordination systems would require expensive infrastructure, training, integration, and 24/7 oversight, making them far more costly than the marginal labor of a human EMT already present at the scene. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute performing this coordination, so cost comparison favors the human entirely; any AI-assisted dispatch tools still require full human execution on-site. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI systems can independently coordinate emergency response teams or manage inter-agency communication during active emergency incidents; this remains a human-only function in all production emergency services. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product coordinates live, on-scene multi-agency emergency response actions among EMTs, police, and fire personnel; this remains a human-only function today. |
Immobilize patient for placement on stretcher and ambulance transport, using backboard or other spinal immobilization device.
0CI 0–0 · exposure 0 · augmentation 13 · click for rater detail
Immobilize patient for placement on stretcher and ambulance transport, using backboard or other spinal immobilization device.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Emergency medical services operate in low-digitization, high-physical-contact environments with strong regulatory oversight and slow technology adoption; no meaningful displacement of EMTs by automation is occurring. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | EMS and prehospital emergency care is a low-digitization, physically-demanding field with minimal automation of hands-on patient care tasks. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI offers minimal assistance for the immobilization task itself; training simulations or post-incident documentation could be assisted, but the core manual and judgmental work remains largely non-augmented by current systems. |
| Augmentation potential | claude-sonnet-5 | 1/5 | AI offers no meaningful assistance to the physical act of immobilizing a patient on a backboard; any AI role would be limited to unrelated documentation or dispatch support. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires physical manipulation of a patient's body, precise handling of medical equipment, and real-time assessment of spinal injury severity. Current AI systems cannot perform end-to-end physical manipulation in uncontrolled environments like emergency scenes. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a hands-on physical task requiring manual manipulation of a patient's body, application of straps, and adaptive physical technique that current AI and robotics cannot perform. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Strong legal and regulatory barriers exist: EMTs must be licensed; patient safety and liability create high error-cost asymmetry; medical protocols require trained human judgment; and direct physical contact with vulnerable patients is both legally and ethically restricted to qualified humans. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Requires licensed, certified EMT personnel physically present to safely immobilize and move patients; strong liability, safety, and regulatory requirements mandate human performance. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The cost of a humanoid robot capable of safe patient handling, combined with integration, liability insurance, and oversight, would vastly exceed the loaded wage of an EMT performing this critical task. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI system capable of substituting for this physical task, so any comparison favors the human EMT by default. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI product can reliably perform patient immobilization. This demands embodied robotics with dexterous manipulation, situational awareness, and medical judgment—all beyond current production systems. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs physical spinal immobilization; this remains entirely a manual, in-person clinical/physical skill. |
Perform emergency diagnostic and treatment procedures, such as stomach suction, airway management, or heart monitoring, during ambulance ride.
0CI 0–0 · exposure 0 · augmentation 50 · click for rater detail
Perform emergency diagnostic and treatment procedures, such as stomach suction, airway management, or heart monitoring, during ambulance ride.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Emergency medical services are heavily regulated, conservative in adoption, and require immediate human presence on-scene. Regulatory, liability, and safety constraints mean adoption of autonomous or AI-driven emergency procedures remains negligible, with only incremental adoption of diagnostic aids. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Emergency medical services are a highly physical, safety-critical field with minimal AI adoption for hands-on procedures; digitization here lags far behind office-based professional sectors. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI-assisted diagnostics (e.g., automated cardiac rhythm analysis, real-time vital sign alerts, or remote physician consultation) can meaningfully assist the EMT in decision-making during procedures. However, the human EMT remains in full control of execution, limiting augmentation to decision support rather than transformation. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-enabled monitoring equipment (e.g., smart heart monitors, decision-support algorithms) can assist EMTs in interpreting vitals or flagging anomalies, but the physical procedures themselves remain manually performed. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Emergency diagnostic and treatment procedures require real-time physical manipulation (airway management, stomach suction), immediate clinical judgment under uncertainty, and hands-on patient contact. Current AI systems cannot physically perform these interventions or reliably make split-second decisions in unstable environments without human oversight. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires physical manipulation of a patient's body in a moving vehicle under time pressure, involving fine motor skills, tactile assessment, and split-second judgment calls that no current AI system can perform end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Strict regulatory and licensing requirements mandate that only certified EMTs can perform emergency medical procedures; legal liability for errors is high and falls on the licensed provider. Patient safety and duty-of-care laws create hard barriers to automation or delegation to unlicensed systems. |
| Adoption barriers | claude-sonnet-5 | 5/5 | EMTs must be licensed/certified and legally authorized to perform these invasive procedures, with strict protocols, medical direction oversight, and liability requirements that mandate human execution. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Emergency medical procedures require licensed personnel whose presence is non-negotiable for legal and safety reasons. AI monitoring or decision support would only supplement, not replace, the paramedic cost, making the all-in cost of AI integration exceed the cost of human provision. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI substitute performing this physical task, so no meaningful cost comparison exists; the human EMT is the only viable option. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI product can autonomously perform physical emergency procedures like airway management or stomach suction. While remote monitoring and diagnostic assistance exist, the core task—executing emergency treatment during transport—remains entirely human-dependent in practice. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs physical emergency medical procedures like airway management or stomach suction; this remains squarely in the domain of trained human responders. |
Related occupations — Healthcare Practitioners & Technical
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.