Paramedics
29-2043.00Median wage $60,600/yr100,610 employed (US)Rank #910 of 923 scored · top 99% by substitution
Administer basic or advanced emergency medical care and assess injuries and illnesses. May administer medication intravenously, use equipment such as EKGs, or administer advanced life support to sick or injured individuals.
Sub-scores
0–100 · band = confidence interval from rater disagreement
Substitution7
Exposure8
Augmentation38
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
13 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%8
panel mean rating 1.3/5 → substitution pressure 8/100
Technical feasibility todayw 20%9
panel mean rating 1.3/5 → substitution pressure 9/100
Cost vs. human wagew 15%7
panel mean rating 1.3/5 → substitution pressure 7/100
Adoption barriersw 20%inverted — strong barriers lower the score5
panel mean rating 4.8/5 (barrier strength) → substitution pressure 5/100
Sector adoption velocityw 10%7
panel mean rating 1.3/5 → substitution pressure 7/100
Task breakdown (13 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.
Observe, record, and report to physician the patient's condition or injury, the treatment provided, and reactions to drugs or treatment.
51CI 25–76 · exposure 58 · augmentation 75 · 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.
51| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Healthcare and emergency services are adopting clinical documentation AI and EHR automation, but adoption is middling: many systems pilot AI-assisted note generation and vital-sign tracking, yet full displacement of paramedic recording duties remains rare due to regulatory and liability conservatism. Faster uptake in hospital settings than in field-based paramedicine. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | EMS and prehospital care is a physically-oriented, safety-critical sector with historically slow technology adoption; ambient documentation pilots exist but are not yet widespread standard practice. |
| Augmentation potential | claude-haiku-4-5-20251001 | 5/5 | AI substantially augments paramedic productivity by automatically capturing and structuring observations, flagging critical findings, and pre-populating reports—allowing paramedics to focus on patient care and complex decision-making rather than manual documentation. This transforms the workflow while keeping the human fully in the loop. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-powered ambient listening and transcription tools can help paramedics draft reports and structure data faster, but the paramedic must still perform and verify the clinical observation and communication themselves. |
| Task automatability | claude-haiku-4-5-20251001 | 5/5 | Observation, recording, and reporting patient condition, treatments, and reactions are highly standardized, structured tasks well-suited to AI analysis. Current AI systems can parse vital signs, medication logs, and symptom descriptions to generate comprehensive clinical reports faster and at least as accurately as human paramedics, meeting the 50% time-saving threshold. |
| Task automatability | claude-sonnet-5 | 2/5 | Voice-to-text and ambient documentation tools can assist with recording, but the core task requires real-time clinical observation, judgment about relevant findings, and physical presence with the patient that current AI cannot perform end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Regulatory frameworks (HIPAA, state scope-of-practice laws) require human paramedic involvement in direct patient care assessment, but documentation and reporting can be partially automated. Liability concerns and institutional preference for human accountability on critical decisions create moderate friction but do not legally block automation of the recording and reporting component itself. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Clinical reporting to a physician involves licensure, medical liability, and regulatory documentation standards (e.g., EMS protocols, HIPAA), requiring a credentialed provider to observe and attest to patient condition and treatment. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | AI inference and documentation automation cost substantially less than paramedic labor (fully loaded wages ~$40–50/hour), while integration and compliance monitoring add overhead. The all-in cost of AI-assisted recording is typically 20–30% of paramedic wages, placing it favorably in the cost-benefit equation. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI documentation tools reduce some paperwork time but still require paramedic oversight, verification, and the underlying observation/assessment work, so total cost savings versus a paramedic's time are modest, not order-of-magnitude. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Mature medical AI products (EHR integration, automated vital-sign interpretation, clinical note generation) are deployed in production healthcare settings; however, real-world variability in patient presentations and the requirement to integrate with human oversight introduce some friction. Most healthcare systems have partially automated versions of this task but rarely fully autonomous end-to-end execution. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Ambient scribe and documentation products exist in EMS/clinical settings but are narrow in scope (transcription/summarization) and do not reliably handle the full observe-assess-report loop, especially in field conditions with noise and urgency. |
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 | Paramedic services operate in traditional, regulated, often under-resourced sectors with slow digital adoption. While some agencies pilot decision-support tools, field deployment remains sporadic and largely experimental rather than production-scale. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | EMS is a physical, low-digitization field role with minimal AI deployment for hands-on assessment tasks despite some digital charting adoption. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist paramedics by flagging vital-sign anomalies, suggesting differential diagnoses, and recommending protocol checks, improving decision speed and consistency. However, augmentation is constrained by the need for robust connectivity and the paramedic's primary reliance on direct patient assessment. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI-based protocols, decision trees, or documentation aids can support paramedics' judgment, but core physical assessment still relies entirely on human skill. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can analyze vital signs and imaging to support triage decisions, the task fundamentally requires real-time clinical judgment, patient interaction, and dynamic reassessment in unpredictable field conditions. Current AI cannot reliably perform initial scene assessment, differential diagnosis, and procedural prioritization end-to-end without substantial human oversight. |
| Task automatability | claude-sonnet-5 | 1/5 | Triage and assessment require real-time physical examination, hands-on vital sign checks, and situational judgment in unpredictable field environments that current AI cannot perform end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Strong licensing and legal barriers exist: paramedics must be certified and legally liable for initial assessment and treatment decisions. Regulatory bodies and medical directors enforce human accountability, and liability for missed diagnoses creates a high error-cost asymmetry that prevents full substitution. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Licensed EMS personnel are legally required to perform patient assessment and triage decisions, with strict liability, certification, and regulatory oversight. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Integrating AI decision-support into EMS systems requires robust hardware, real-time connectivity, training overhead, and liability management. The cost per assessment remains comparable to or exceeds the marginal cost of a paramedic's evaluation, especially given required human oversight. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI-only substitute for this task, so cost comparison favors the human paramedic who must physically be present. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | AI-assisted diagnostic tools exist in hospital settings, but deployed paramedic field systems remain limited to basic vital-sign interpretation and protocol suggestions. No production systems demonstrate reliable autonomous triage and prioritization in the chaotic pre-hospital environment without medic-led validation. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs autonomous field triage of trauma/illness; AI decision-support tools exist only as adjuncts, not as substitutes for the physical assessment itself. |
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 50 · 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 task is governed by regulatory mandates that cannot be circumvented by technology adoption. Paramedics in all sectors must meet certification requirements regardless of AI availability, and no displacement is occurring. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | EMS and healthcare training sectors are slow to adopt AI-driven certification pathways due to regulatory rigidity and reliance on physical skills assessment. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist by generating study guides, summarizing new developments in the field, or creating practice questions to help paramedics prepare for or review training content more efficiently, though the attendance and learning remain human responsibilities. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help paramedics study, summarize new protocols, and provide personalized quizzes or reminders for renewal deadlines, improving efficiency without replacing the certification process itself. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Training attendance and licensure maintenance inherently require a human learner to be present and engaged. No AI system can substitute for the human's participation in classes or satisfy regulatory continuing education requirements on their behalf. |
| Task automatability | claude-sonnet-5 | 2/5 | AI can deliver e-learning content and quizzes, but the task requires attending accredited training, hands-on skills practice, and often in-person certification exams that AI cannot fully replace. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Regulatory bodies (state EMS boards, medical licensing authorities) require paramedics to personally complete and document continuing education. Legal authorization mandates that a licensed human must attend and complete these training requirements themselves. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Licensure bodies require documented completion of accredited courses and often hands-on skills verification, creating regulatory barriers to fully AI-delivered certification. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | This task requires human time investment (hours of attendance) that cannot be meaningfully reduced by AI. The human paramedic must personally complete the training; AI cannot lower the total cost burden. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI-based e-learning modules can be cheap to deliver, but mandatory hands-on skills labs and proctored certification exams still require human instructors and equipment, keeping overall costs comparable to traditional training. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | While AI can generate educational content or study materials, it cannot attend classes, take certification exams, or fulfill the human presence and participation mandated by licensing bodies. No deployed product performs this task end-to-end. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some online CE platforms use adaptive learning tech, but paramedic recertification still mandates practical skills verification and instructor-led components not handled by deployed AI products. |
Coordinate with treatment center personnel to obtain patients' vital statistics and medical history, to determine the circumstances of the emergency, and to administer emergency treatment.
10CI 0–20 · exposure 13 · augmentation 50 · click for rater detail
Coordinate with treatment center personnel to obtain patients' vital statistics and medical history, to determine the circumstances of the emergency, and to administer emergency treatment.
10| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Emergency medical services are highly regulated, government or hospital-run, and operate in low-digitization field environments. Adoption of AI for core paramedic tasks remains minimal; most technology deployment focuses on dispatch and routing rather than task replacement. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | EMS and emergency medical response are physical, low-digitization sectors with minimal AI agent deployment for direct patient care tasks. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist paramedics by rapidly retrieving patient medical history, suggesting protocols based on symptoms, and streamlining documentation and communication with treatment centers. However, augmentation is limited because the paramedic must remain in full control of clinical decisions and hands-on care delivery. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can assist with documentation, decision-support checklists, or relaying patient history data to treatment centers, but does not materially change the hands-on treatment process. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can help gather and summarize medical history and vital statistics from records, the task requires real-time patient interaction, clinical judgment on emergency circumstances, and hands-on treatment delivery. Current AI cannot independently obtain vitals, assess emergency context through patient interviews, or administer physical treatments, so automatability falls well short of 50% time savings at equal quality. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires real-time physical assessment, hands-on emergency treatment, and dynamic verbal coordination under crisis conditions that current AI cannot perform end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Paramedics operate under strict licensure requirements, scope-of-practice laws, and medical oversight regulations that mandate a licensed human perform or directly supervise emergency treatment. Legal liability for patient outcomes and the requirement for physical medical intervention by a licensed provider create hard barriers to automation. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Paramedic certification, medical licensure, liability for treatment decisions, and legal requirements for direct patient care create hard barriers to substitution. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI systems capable of medical data analysis and note-taking are available but remain relatively costly to integrate into ambulance operations, and they cannot replace the paramedic's physical presence and clinical expertise, making the cost-benefit ratio unfavorable for substitution. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute performing the physical treatment and coordination, so cost comparison favors the human paramedic entirely. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Deployed products can assist with data retrieval and documentation, but no current AI system reliably performs the full task of coordinating with treatment centers, determining emergency circumstances through patient assessment, and administering treatment. Emergency medicine demands continuous human decision-making and physical intervention that AI cannot replicate in production. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs on-scene emergency treatment coordination and administration; this remains firmly in the physical, human-performed domain. |
Coordinate work with other emergency medical team members or police or fire department personnel.
3CI 0–5 · exposure 5 · augmentation 38 · click for rater detail
Coordinate work with other emergency medical team members or police or fire department personnel.
3| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Emergency services remain highly conservative in adopting autonomous AI systems, especially for core command-and-control functions. Adoption is limited to back-office dispatch support, not front-line coordination replacement. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Emergency medical services are a low-digitization, physically embedded sector with minimal AI agent deployment in operational coordination roles. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could provide marginal assistance via real-time data display (unit locations, available resources) or protocol reminders, but current systems offer limited augmentation to human coordinators managing dynamic, multi-team situations in the field. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-enabled dispatch systems, radio transcription, and situational data sharing tools can support communication and information flow, improving coordination efficiency without replacing the human role. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Coordinating with multiple emergency personnel requires real-time communication, contextual judgment, and dynamic decision-making in high-stakes situations. Current AI cannot autonomously participate in this interpersonal coordination, which fundamentally depends on human presence and authority at the scene. |
| Task automatability | claude-sonnet-5 | 1/5 | Real-time, high-stakes coordination among first responders at dynamic emergency scenes requires physical presence, situational judgment, and split-second decisions that current AI cannot perform end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Coordination of emergency response is legally and operationally controlled by authorized personnel (incident commanders, paramedic supervisors); liability and safety regulations require qualified humans to direct and sign off on emergency scene coordination. AI cannot replace this gatekeeping role. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Emergency medical response is heavily regulated, requires licensed personnel, involves life-safety liability, and mandates human authority and accountability at scene. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | AI systems capable of even partial coordination assistance require significant infrastructure, training, and oversight, making them costlier than paying paramedics to coordinate directly with each other and other personnel. |
| 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. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While AI can assist with dispatch scheduling or suggest communication protocols, no deployed system actually performs live coordination between emergency teams. Some decision-support tools exist, but they do not reliably substitute for human-led coordination in production emergency settings. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product autonomously coordinates paramedics with police/fire personnel in live emergency response; dispatch software assists but does not replace this task. |
Operate equipment, such as electrocardiograms (EKGs), external defibrillators, or bag valve mask resuscitators, in advanced life support environments.
3CI 0–5 · exposure 5 · augmentation 50 · click for rater detail
Operate equipment, such as electrocardiograms (EKGs), external defibrillators, or bag valve mask resuscitators, in advanced life support environments.
3| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Emergency medical services remain highly human-centric and are slow to digitize. Real-world EMS adoption of advanced automation is minimal; paramedics remain the bottleneck and gold standard for field care, with no evidence of displacement by AI systems. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | EMS and prehospital care is a low-digitization, physically demanding field with minimal AI agent deployment for hands-on procedures. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist by offering real-time EKG interpretation prompts or protocol reminders, improving decision speed and consistency. However, the paramedic retains full manual and clinical control; assistance is advisory rather than transformative to the core equipment-operation task. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-enabled devices can assist with EKG interpretation or defibrillator guidance (e.g., automated rhythm analysis), providing decision support while the paramedic performs the physical task. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Operating advanced life support equipment requires real-time clinical decision-making, physical manipulation of devices in chaotic environments, and rapid adaptation to patient condition changes. Current AI systems cannot autonomously perform these interconnected functions end-to-end in prehospital or emergency settings. |
| Task automatability | claude-sonnet-5 | 1/5 | This is hands-on physical operation of medical devices on patients in emergency, often mobile settings, requiring physical dexterity and real-time clinical judgment that no current AI system can perform end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Paramedics operate under strict medical and legal licensing requirements; protocols demand a licensed provider to assess the patient, make treatment decisions, and be accountable for outcomes. Liability and regulatory frameworks create hard barriers to autonomous operation of life-support equipment. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Emergency medical care requires licensed personnel, direct physical intervention, and legal/medical liability structures that mandate human operation of life-support equipment. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The cost of deploying autonomous systems capable of reliable equipment operation in emergency environments, including safety redundancy and regulatory compliance, far exceeds the cost of a trained paramedic, making this economically unfeasible today. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute performing the physical operation, so cost comparison favors the human paramedic entirely. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While AI can interpret EKG strips in controlled lab settings and guide protocol selection, deployed systems do not reliably operate physical equipment (defibrillators, bag valve masks) or manage the full task in the high-stakes, variable conditions paramedics face. No production system handles the embodied, real-time aspects. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product operates EKGs, defibrillators, or bag valve masks autonomously on patients in the field; these remain human-operated devices with at most software-assisted interpretation. |
Administer first aid treatment or life support care to sick or injured persons in prehospital settings.
1CI 0–3 · exposure 0 · augmentation 38 · click for rater detail
Administer first aid treatment or life support care to sick or injured persons in prehospital settings.
1| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Adoption of AI in prehospital care remains minimal; most EMS systems are in early stages of exploring decision-support tools, and actual operational displacement is negligible despite digitization of some protocols. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | EMS is a physical, decentralized, low-digitization field with minimal AI deployment for direct patient care; adoption of AI here is essentially nonexistent. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist with protocol guidance, vital-sign interpretation, and dispatch optimization, helping paramedics make faster decisions, though the core physical interventions remain entirely human-performed. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can assist with decision-support tools (triage protocols, dispatch optimization, documentation) but offers little direct help during the hands-on delivery of first aid or life support. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | First aid and life support require physical intervention (CPR, wound care, intubation) and real-time clinical judgment in unpredictable field conditions; no current AI system can perform these hands-on interventions or reliably replace human decision-making in dynamic emergency settings. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires hands-on physical intervention (CPR, airway management, IV insertion, medication administration) in unpredictable field environments—current AI cannot perform physical medical procedures. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Emergency medical care is heavily regulated; only licensed paramedics and certain medical professionals are legally authorized to administer life support and first aid, creating hard legal and licensure barriers to automation. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Paramedics must be licensed/certified, life support care is heavily regulated, and liability for medical errors is severe, requiring authorized human providers by law. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | AI systems cannot physically deliver care, so direct cost comparison is not meaningful; any AI support would be additive to human paramedic costs rather than substitutive. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute performing this physical task, so cost comparison favors humans entirely; any AI-robotic equivalent would be far more expensive and unproven. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | While AI can support triage and protocol decision-making in controlled settings, no deployed product reliably administers actual first aid or life support care; the task fundamentally requires physical presence and embodied competence. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product performs hands-on prehospital emergency care; robotics for such dynamic, high-stakes physical intervention remains research-stage at best. |
Administer drugs, orally or by injection, or perform intravenous procedures.
0CI 0–0 · exposure 0 · augmentation 25 · click for rater detail
Administer drugs, orally or by injection, or perform intravenous procedures.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Paramedics work in field, acute, and pre-hospital settings with high physical and environmental variability; adoption of automation in these contexts remains minimal, and no industry trend toward autonomous drug administration exists. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Emergency medical services are a physically-grounded, low-digitization sector with no meaningful movement toward AI or robotic execution of hands-on clinical procedures. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI offers limited assistance—e.g., decision support on drug dosing or IV site selection—but the physical act of administration itself resists augmentation; the paramedic must remain the primary actor. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can support dosage calculations, drug interaction checks, or protocol reminders via decision-support tools, but it does not materially transform the hands-on execution of injections or IV procedures. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Physically administering drugs or performing IV procedures requires precise manual dexterity, physical presence, and real-time adaptation to patient anatomy and condition—tasks that current AI systems cannot perform end-to-end without a human operator. |
| Task automatability | claude-sonnet-5 | 1/5 | Administering drugs and IV procedures requires physical manipulation of a patient's body, real-time clinical judgment, and manual dexterity that current AI systems cannot perform without robotic embodiment far beyond deployed capability. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Hard legal and regulatory barriers exist: only licensed healthcare professionals (paramedics, nurses, physicians) are authorized to administer medications and perform invasive procedures; liability and malpractice law reinforce this requirement. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Drug administration and IV procedures are tightly regulated medical acts requiring licensure, protocol adherence, and legal accountability, with severe liability for errors, making non-human performance essentially prohibited. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Current robotic systems capable of precise medical manipulation cost hundreds of thousands to millions of dollars per unit, plus maintenance and oversight, far exceeding the wage cost of a trained paramedic. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute performing this physical task, so any hypothetical AI-robotic solution would be far more expensive than a trained paramedic given current technology costs. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product performs autonomous drug administration or IV insertion in clinical settings; this remains exclusively a human-performed clinical task despite decades of robotic research. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product autonomously administers drugs or performs IV insertion in prehospital emergency settings; this remains purely hypothetical for field paramedicine. |
Comfort and reassure patients.
0CI 0–0 · exposure 0 · augmentation 25 · click for rater detail
Comfort and reassure patients.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | There is no measurable adoption of AI to replace paramedic reassurance in any sector; the task is fundamentally tied to licensed human presence in emergency response. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | EMS and prehospital care is a low-digitization, physically-embodied sector with minimal AI adoption for direct patient interaction tasks. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI might provide paramedics with suggested language or protocols to guide communication, but the core value of this task—human presence and authentic reassurance—cannot be augmented by current AI systems. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI could provide minor background support (e.g., protocol reminders, translation aids) but does not meaningfully enhance the human act of comforting a patient. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Comforting and reassuring patients requires genuine human empathy, presence, and emotional intelligence that AI systems cannot authentically provide. Current AI lacks the embodied, relational capacity to perform this task meaningfully. |
| Task automatability | claude-sonnet-5 | 1/5 | Comforting and reassuring an injured or acutely distressed patient requires physical presence, tone of voice, empathy, and real-time human judgment that current AI cannot deliver at the scene of an emergency.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Legal, professional, and ethical standards require a licensed paramedic to be physically present with patients. Patient safety and liability frameworks make human accountability non-negotiable in emergency care. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Direct human contact, trust-building, and licensed medical care during emergencies are legally and practically required, making substitution essentially impossible. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The task exists precisely because human presence has irreducible value; automating it would require either replacing the human entirely or deploying AI at costs that exceed the value of removing a paramedic from the role. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute performing this function, so cost comparison favors the human paramedic entirely. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI product reliably performs genuine emotional comfort and reassurance in a clinical emergency setting. Chatbots can simulate supportive language but cannot replicate the presence and trust-building that defines this task. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product provides in-person emotional reassurance during medical emergencies; this remains a purely human, physically embodied interaction. |
Instruct emergency medical response team about emergency interventions to ensure correct application of procedures.
0CI 0–0 · exposure 0 · augmentation 25 · click for rater detail
Instruct emergency medical response team about emergency interventions to ensure correct application of procedures.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Emergency medical services remain human-led by necessity and regulation. Despite some adoption of decision-support tools, actual team instruction and leadership remains a human responsibility with no meaningful displacement trend. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | EMS and field emergency medicine are low-digitization, physically embedded environments with minimal AI agent deployment for real-time team direction. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could provide real-time reference material or decision-support suggestions to assist a human instructor, but the core task—authoritative leadership during emergencies—cannot be meaningfully augmented by current systems without undermining clinical safety and accountability. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI-based training simulations, protocol checklists, and decision-support apps can help paramedics prepare and reference procedures, but offer limited real-time assistance during live team instruction in emergencies. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Instructing a team during high-stakes emergency interventions requires real-time judgment, authority, accountability, and adaptive communication based on unpredictable clinical situations. Current AI cannot reliably lead or coordinate live emergency teams, nor can it take responsibility for procedure outcomes. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires real-time, hands-on clinical leadership during live emergencies with unpredictable physical and physiological variables; AI cannot direct human teams performing physical medical interventions in the field. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Emergency medical practice is heavily regulated; only licensed paramedics/physicians can legally direct emergency interventions and sign off on procedures. Liability and legal accountability are non-delegable to automation. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Paramedic certification, medical liability, and legal scope-of-practice requirements mandate a licensed human directing emergency care, making this a hard regulatory and liability barrier. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | A human paramedic team lead must be present regardless of any AI assistance; there is no cost substitution path. The loaded wage of the human instructor cannot be displaced by inference costs. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute performing this function, so cost comparison favors the human paramedic entirely; any AI-assisted training tools do not replace the real-time instructional task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI system can serve as the authoritative instructor for emergency medical teams during active incidents. This requires a licensed, accountable human making real-time decisions and bearing legal/clinical responsibility. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product directs or corrects a live emergency response team's physical interventions during an active medical emergency; this remains firmly outside current product capability. |
Perform emergency cardiac care, such as cardioversion and manual defibrillation.
0CI 0–0 · exposure 0 · augmentation 25 · click for rater detail
Perform emergency cardiac care, such as cardioversion and manual defibrillation.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | There is zero adoption of AI for autonomous performance of emergency cardiac care interventions, and there are no pilots or production deployments attempting to replace paramedics for manual defibrillation or cardioversion. The task remains exclusively within human professional scope. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | EMS and prehospital emergency care is a highly physical, low-digitization sector with minimal AI agent deployment for hands-on interventions. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI can provide marginal assistance (e.g., ECG rhythm classification, prompting for protocol steps, alerting to deviations) but the core manual and decisional work remains with the paramedic. Current systems offer limited augmentation because the task is so time-critical and physically dependent that decision support is secondary to immediate human action. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI-enabled defibrillators can provide rhythm analysis and shock/no-shock guidance, offering some decision support, but the core physical task and judgment remain human-driven with limited AI enhancement. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Emergency cardiac care requires real-time physical intervention on a patient's body (manual defibrillation, cardioversion), immediate clinical judgment under uncertainty, and rapid adaptation to changing vital signs. Current AI systems cannot physically perform these interventions or make the split-second, context-dependent decisions required in a cardiac emergency. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a hands-on physical emergency intervention requiring real-time clinical judgment, physical equipment placement, and immediate action on a live patient; no AI system can perform this end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Hard legal and regulatory barriers exist: only licensed paramedics and physicians are authorized to perform defibrillation and cardioversion under medical law. Liability, patient safety requirements, and explicit professional licensing requirements prevent any automation or substitution by non-human agents. |
| Adoption barriers | claude-sonnet-5 | 5/5 | This is a licensed medical intervention with strict scope-of-practice regulations, life-or-death liability, and legal requirements that only certified personnel administer emergency cardiac care. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The cost comparison is moot because no AI system can perform this task at all. A paramedic performing emergency cardiac care cannot be replaced by current technology, so the AI cost is infinite relative to the human wage for performing this specific intervention. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI substitute performing this physical task, so cost comparison favors the human paramedic entirely; AI cannot deliver the output at any cost. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product can autonomously perform manual defibrillation or cardioversion on a patient. These tasks require licensed human paramedics with specialized training and certification; no AI system operates independently in this capacity in any healthcare setting. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs manual defibrillation or cardioversion autonomously; automated external defibrillators exist but require human application and are not AI-driven decision agents replacing paramedics. |
Perform emergency invasive intervention before delivering patient to an acute care facility.
0CI 0–0 · exposure 0 · augmentation 25 · click for rater detail
Perform emergency invasive intervention before delivering patient to an acute care facility.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Emergency medical services remain heavily reliant on human paramedics; no measurable shift toward AI-driven invasive interventions exists, and regulatory and safety constraints make rapid adoption implausible. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | EMS and prehospital emergency care is a low-digitization, physically demanding field with minimal AI deployment for hands-on interventions; adoption of physical automation here is essentially nonexistent. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could marginally assist via real-time decision support (suggesting procedures, guiding technique, monitoring) but the paramedic must perform the intervention; current tools offer limited augmentation for the core physical act itself. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can assist with decision support, protocol reminders, or vital sign monitoring during the encounter, but it does not meaningfully enhance the physical execution of the invasive procedure itself. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Emergency invasive interventions (intubation, chest decompression, surgical airway, emergency transfusion) require real-time physical manipulation, rapid decision-making under uncertainty, and immediate adaptation to patient response. Current AI cannot perform physical procedures or operate medical equipment autonomously. |
| Task automatability | claude-sonnet-5 | 1/5 | Emergency invasive interventions (e.g., intubation, needle decompression, IV/IO access) require physical manual dexterity, real-time sensory judgment, and hands-on manipulation of a patient's body—no current AI system can perform this physical act at all. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Invasive emergency medicine is legally and clinically restricted to licensed healthcare professionals (paramedics, physicians, nurses) acting under strict protocols; liability, malpractice exposure, and informed consent requirements create insurmountable legal barriers to autonomous AI substitution. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Invasive medical procedures are tightly regulated, requiring licensed, certified personnel under medical protocols and physician oversight, with high liability for errors—strong legal and professional barriers prevent non-human substitution. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The infrastructure and oversight required for any autonomous invasive intervention would vastly exceed the cost of a trained paramedic, who can be deployed immediately for multiple tasks across an entire shift. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI substitute performing this physical task, so cost comparison is moot—human paramedics are the only option, making AI effectively infinitely costlier (i.e., not viable) for this task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product performs invasive emergency medical interventions autonomously in clinical practice. AI systems lack the embodied capability, real-time sensing, and regulatory clearance to execute such high-stakes procedures independently. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs physical invasive medical procedures in the field; this remains far outside the scope of any commercial AI or robotic system available to EMS today. |
Perform emergency pharmacological interventions.
0CI 0–0 · exposure 0 · augmentation 38 · click for rater detail
Perform emergency pharmacological interventions.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Paramedic services remain largely human-staffed emergency response operations with limited digitization of field protocols. Adoption of AI agents to replace paramedic functions in emergency settings is virtually nonexistent; regulatory and operational constraints prevent rapid shifts. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | EMS is a physical, hands-on field profession with minimal AI-driven displacement of actual drug administration; adoption is confined to decision-support tools, not the intervention itself. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI-assisted clinical decision support (e.g., protocol reminders, drug interaction checking, dosing calculators) can assist paramedics in confirming decisions and reducing errors. Such tools are beginning to appear in some EMS systems, providing meaningful but limited augmentation to the human's core task. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can assist with dosage calculators, protocol lookup, or documentation, but offers limited real-time support during the actual physical administration of interventions. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Emergency pharmacological interventions require real-time clinical assessment, judgment about patient-specific contraindications, and precise administration under dynamic, life-threatening conditions. Current AI cannot reliably perform the full intervention chain—diagnosis confirmation, drug selection, dose calculation, IV/IM/IO access, and administration—without human execution. |
| Task automatability | claude-sonnet-5 | 1/5 | Administering medications in emergency field settings requires physical presence, hands-on assessment, and real-time judgment that current AI cannot perform end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Pharmacological administration by paramedics is tightly regulated; laws and protocols mandate that a licensed, certified paramedic must authorize and perform medication administration. Liability, legal accountability, and scope-of-practice restrictions create hard barriers to substitution. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Administering controlled substances and emergency drugs requires licensure, medical protocols, and legal authority; only certified paramedics/EMTs may perform this under medical direction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | AI cannot replace the paramedic performing this task—the human remains essential. Integration costs for AI-assisted decision support would be overhead, not a cost reduction, making total cost per intervention higher than human-alone delivery. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI substitute performing this physical task, so cost comparison favors the human paramedic by default. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI system performs end-to-end emergency drug administration in production. Clinical decision support exists, but the physical act of administering medications, managing airway access, and responding to patient response requires trained human paramedics; no autonomous system does this reliably today. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product administers emergency pharmacological interventions in the field; this remains entirely a research-stage or non-existent capability for physical drug administration. |
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.