Nurse Anesthetists
29-1151.00Median wage $236,590/yr51,840 employed (US)Rank #894 of 923 scored · top 97% by substitution
Administer anesthesia, monitor patient's vital signs, and oversee patient recovery from anesthesia. May assist anesthesiologists, surgeons, other physicians, or dentists. Must be registered nurses who have specialized graduate education.
Sub-scores
0–100 · band = confidence interval from rater disagreement
Substitution9
Exposure9
Augmentation47
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
24 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%11
panel mean rating 1.5/5 → substitution pressure 11/100
Cost vs. human wagew 15%9
panel mean rating 1.4/5 → substitution pressure 9/100
Adoption barriersw 20%inverted — strong barriers lower the score6
panel mean rating 4.7/5 (barrier strength) → substitution pressure 6/100
Sector adoption velocityw 10%11
panel mean rating 1.5/5 → substitution pressure 11/100
Task breakdown (24 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.
Perform or evaluate the results of diagnostic tests, such as radiographs (x-rays) and electrocardiograms (EKGs).
44CI 25–62 · exposure 50 · augmentation 88 · importance 4.4/5 · click for rater detail
Perform or evaluate the results of diagnostic tests, such as radiographs (x-rays) and electrocardiograms (EKGs).
44| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Healthcare is moderately digitized with increasing AI adoption in imaging and monitoring, but deployment remains slower than in tech or finance sectors, with many institutions still in pilot phases or maintaining dual-interpretation workflows rather than full substitution. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Healthcare, especially perioperative/anesthesia care, is a slower-adopting sector for autonomous AI decision tools due to regulatory scrutiny and patient safety concerns, though AI-assisted diagnostics are gradually being piloted. |
| Augmentation potential | claude-haiku-4-5-20251001 | 5/5 | AI substantially augments anesthetist productivity by providing rapid preliminary interpretation, flagging critical findings, and reducing cognitive load during procedure planning and intraoperative monitoring, transforming turnaround time while the clinician retains final judgment. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-based EKG and imaging analysis tools can meaningfully speed up and improve accuracy of test interpretation, supporting the nurse anesthetist while they retain final clinical responsibility. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Current AI systems can reliably detect many abnormalities in radiographs and EKGs with performance at or exceeding human radiologists in controlled settings, enabling significant time savings in image interpretation and preliminary analysis, though some complex or ambiguous cases still require expert review. |
| Task automatability | claude-sonnet-5 | 2/5 | AI can assist in interpreting EKGs and radiographs (pattern recognition) but the clinical integration into anesthesia decision-making and hands-on performance of tests requires human judgment and physical presence, so full end-to-end automation is not achievable today.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Significant barriers exist: clinician oversight and sign-off remain legally and professionally mandated, liability concerns around AI misses persist, regulatory frameworks (FDA, state practice acts) restrict autonomous AI interpretation, and clinical inertia favors human credentialing. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Interpreting diagnostic tests as part of anesthesia care requires a licensed practitioner's clinical judgment and legal accountability for patient safety, creating strong regulatory and liability barriers to full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | AI-based interpretation (model inference plus infrastructure) costs dramatically less per result than a licensed human anesthetist or radiologist reviewing the same image or tracing, with the cost gap widening through scale. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI diagnostic tools have real licensing, integration, and oversight costs that keep them comparable to or only modestly cheaper than the marginal cost of a clinician reviewing results, especially given liability requirements. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Multiple FDA-cleared AI products for radiograph and EKG interpretation are deployed in clinical settings, with demonstrated performance in production environments, though adoption remains uneven and clinician review remains standard of care. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Deployed AI products exist for EKG interpretation and some radiograph triage (e.g., cardiology and radiology AI tools), but they are not integrated into nurse anesthetist workflows as autonomous decision-makers and require clinician oversight. |
Read current literature, talk with colleagues, and participate in professional organizations or conferences to keep abreast of developments in nursing.
33CI 10–55 · exposure 30 · augmentation 63 · importance 4.1/5 · click for rater detail
Read current literature, talk with colleagues, and participate in professional organizations or conferences to keep abreast of developments in nursing.
33| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | This is a discretionary, ongoing professional practice tied to individual competency and motivation, not a routine operational task. Healthcare organizations do not automate professional self-development; adoption of AI to replace it would be negligible. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Healthcare professionals increasingly use AI-powered literature summarization and alert tools, though adoption in continuing education workflows is still moderate compared to fully digitized sectors. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can meaningfully assist by summarizing recent literature, flagging relevant papers, or identifying key conference topics, which would help a nurse stay current more efficiently. However, the human must still read critically, evaluate applicability, and attend conferences for the full value. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI literature summarization, alerting, and knowledge synthesis tools substantially speed up staying current with research, meaningfully augmenting this task even though human networking remains central. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Reading literature and participating in professional conferences require sustained human judgment about relevance, synthesis of nuanced domain knowledge, and genuine professional networking that AI cannot replicate end-to-end. While AI can summarize papers, it cannot independently determine what is clinically important or establish professional relationships. |
| Task automatability | claude-sonnet-5 | 3/5 | AI can summarize literature and surface relevant developments, but the networking, discussion, and conference participation components involve human interaction that AI cannot substitute end-to-end.: |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Professional development and maintaining licensure/competency are regulatory and ethical imperatives for nurses; professional societies require human participation for membership benefits and credentialing. Staying current is a personal professional responsibility that cannot be delegated to an automated system. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement mandates this specific CE activity be done a certain way, though professional certification bodies may require documented CE hours from verified sources, adding mild friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The task is inherently about a professional's own development and network-building; automating it away would eliminate the value. Even where AI assists, the human still performs the core work, so cost savings are minimal and the human wage dominates. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI literature review tools are cheap relative to a nurse anesthetist's time, but the human networking/conference elements still require paid time and travel, keeping overall cost comparable. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | AI can assist with literature summaries and identify relevant papers, but no deployed system can reliably replace the full task of staying current—which inherently requires human curation, judgment, and networking participation. Products exist for literature monitoring but require significant human oversight and decision-making. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Products like AI literature summarizers, research digest tools, and conference recommendation systems exist and are used, but no product replaces the full task including colleague dialogue and conference attendance. |
Request anesthesia equipment repairs, adjustments, or safety tests.
25CI 25–25 · exposure 25 · augmentation 50 · importance 4.0/5 · click for rater detail
Request anesthesia equipment repairs, adjustments, or safety tests.
25| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Healthcare adoption of AI for safety-critical equipment maintenance is cautious and slow; most anesthesia departments still rely on human-initiated requests and manual coordination with biomed services, with minimal production AI deployment. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Healthcare, especially clinical/procedural settings, adopts AI slowly for safety-critical administrative workflows tied to medical equipment compliance. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI could assist by summarizing equipment status, flagging patterns of recurring issues, or auto-populating standard maintenance request forms, materially reducing documentation burden while a human anesthetist retains judgment and sign-off on all requests. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-enabled maintenance tracking and predictive alerts can help flag equipment issues or streamline repair requests, offering moderate assistance to the anesthetist's workflow. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | AI cannot currently end-to-end handle equipment repair requests, which require recognizing equipment faults, making judgment calls about safety-critical adjustments, and liaising with biomedical technicians. Some documentation or email drafting could be assisted, but the core task demands human technical assessment. |
| Task automatability | claude-sonnet-5 | 2/5 | This is a coordination/administrative task requiring judgment about equipment status and communication with technicians, but the physical inspection and decision-making about safety-critical equipment resist full automation.Some routing/logging could be automated but the core judgment cannot. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Anesthesia equipment safety is heavily regulated and audited; many institutions require a credentialed anesthesia professional to certify or approve equipment maintenance requests. Liability and regulatory compliance create strong barriers to fully autonomous automation. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Equipment safety in anesthesia is heavily regulated, and only qualified clinical/biomedical staff can authorize or verify equipment safety and repairs, creating strong liability and regulatory barriers. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | The overhead of human oversight, validation of equipment status, and liability for incorrect requests means AI cost savings would be minimal or negative compared to the loaded wage of a nurse anesthetist managing their own equipment concerns in-house. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | A ticketing/workflow system is cheap, but the clinical assessment of what needs repair or testing still requires the nurse anesthetist's time, limiting cost savings for the full task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No production system reliably performs anesthesia equipment maintenance requisition end-to-end; this requires integration with biomedical departments, equipment registries, and safety protocols that vary by institution. Narrow automation of specific request steps (like form-filling) exists but does not constitute reliable task performance. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | No deployed product autonomously manages anesthesia equipment repair requests; at best, hospital maintenance ticketing systems exist but require human initiation and clinical judgment about the issue. |
Instruct nurses, residents, interns, students, or other staff on topics such as anesthetic techniques, pain management and emergency responses.
25CI 25–25 · exposure 25 · augmentation 63 · importance 3.7/5 · click for rater detail
Instruct nurses, residents, interns, students, or other staff on topics such as anesthetic techniques, pain management and emergency responses.
25| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Healthcare education, particularly anesthesia training, remains highly conservative and slow to adopt AI-driven instruction due to accreditation requirements, liability concerns, and institutional inertia. Adoption is limited to supplementary tools rather than core instruction delivery. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Healthcare education is adopting AI tools like simulation and e-learning gradually, but hands-on clinical training remains largely traditional and slow to change due to accreditation and safety requirements. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist instructors by generating explanatory materials, creating interactive case simulations, providing real-time references, and helping organize curriculum content. However, augmentation remains bounded by the need for human judgment in assessment and clinical reasoning demonstration. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can meaningfully support this task by creating training materials, quizzes, case simulations, and personalized learning content, enhancing an instructor's effectiveness and prep time while they still lead in-person or supervised training. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can generate educational content and explanations of anesthetic techniques, instruction inherently requires real-time interaction, personalized feedback, assessment of learner comprehension, and demonstration of procedures—tasks that demand human judgment and adaptive teaching. AI cannot meaningfully replace the mentoring, corrective feedback, and clinical judgment demonstration that defines effective anesthetic education. |
| Task automatability | claude-sonnet-5 | 2/5 | AI can generate instructional content and simulate case scenarios, but live clinical teaching, hands-on demonstration, and real-time supervision of trainees require human presence and judgment that current AI cannot replicate end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Regulatory bodies (accreditation councils, licensing boards) and institutional policies require qualified human instructors for clinical anesthesia training; liability and patient safety concerns create strong barriers to automated instruction in this high-stakes domain. Legal and professional standards tie educational authority to licensed practitioners. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Clinical education, especially involving anesthetic and emergency procedures, typically requires certified professionals for accreditation, liability, and patient-safety reasons, creating strong institutional and regulatory barriers to full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Developing and maintaining AI-driven anesthesia instruction systems requires substantial domain expertise, curation, and quality assurance. The all-in cost (including oversight, validation, and institutional integration) approaches or exceeds the hourly rate of educators, especially when accounting for the criticality of accuracy in medical education. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI-generated training materials are cheap, but the actual instructional task involves supervised practice and credentialed expertise, so overall cost savings versus a qualified instructor are limited. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Educational chatbots and AI tutoring systems exist but lack the clinical credibility, hands-on demonstration capability, and ability to assess and adapt to diverse learner needs in high-stakes medical education. No deployed AI system reliably replaces or fully mirrors a qualified instructor in anesthesia training. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some AI-based medical education tools (adaptive learning platforms, simulation software) exist and are used for supplemental training, but no deployed product substitutes for a nurse anesthetist directly instructing staff on anesthetic technique or emergency response. |
Assess patients' medical histories to predict anesthesia response.
23CI 20–25 · exposure 25 · augmentation 75 · importance 4.7/5 · click for rater detail
Assess patients' medical histories to predict anesthesia response.
23| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Healthcare adoption of AI for clinical decision-making remains cautious; anesthesia is a high-stakes domain with strong reliance on credentialed human expertise. Pilots are underway but production deployment of autonomous prediction systems for anesthesia risk is rare and slow. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Healthcare overall lags in AI adoption for clinical decision-making tasks with direct patient safety implications, with pilots more common than production deployment for this specific function. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can effectively augment nurse anesthetists by automatically surfacing relevant comorbidities, drug interactions, and published risk scores from patient records, reducing manual chart review time and supporting structured pre-anesthetic assessment without removing human judgment from the decision. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can meaningfully assist by aggregating patient history, flagging drug interactions, and summarizing risk factors, helping the nurse anesthetist work faster while retaining final judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can extract and pattern-match against patient histories in structured data, predicting anesthesia response requires integration of complex, often incomplete or unstructured medical records, comorbidities, and patient-specific factors where misclassification carries high clinical stakes. Current systems cannot reliably do this end-to-end without substantial human review and judgment. |
| Task automatability | claude-sonnet-5 | 2/5 | AI can synthesize chart data and flag risk factors, but integrating clinical judgment about individualized anesthesia response requires nuanced reasoning about drug interactions, comorbidities, and physical exam findings that current systems cannot fully replicate end-to-end.4 |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Licensing and regulatory barriers are substantial: only credentialed nurse anesthetists or physicians can legally administer anesthesia and make pre-anesthetic risk assessments; liability for adverse outcomes falls on the human provider; and institutional governance requires human accountability for anesthesia decisions. |
| Adoption barriers | claude-sonnet-5 | 5/5 | This is a licensed medical judgment task; only credentialed anesthesia providers can legally assess and sign off on anesthesia risk, creating a hard regulatory and liability barrier. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Integration into clinical workflows, validation by anesthetists, liability oversight, and the specialized domain expertise required mean AI assistance is roughly comparable in cost to or may exceed the marginal value of a junior anesthetist's review time given low error tolerance. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI-assisted chart review could reduce some clerical time, but the necessary oversight, verification, and liability review keep costs comparable to or only modestly below the human specialist's time given the stakes involved. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Pilot studies exist showing AI can flag risk factors, but no deployed clinical product reliably predicts individual anesthesia response as a standalone decision tool. Existing systems are narrow (specific drug or population), require human validation, and lack the evidence base for independent clinical deployment. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Clinical decision support tools exist that surface risk flags from EHR data, but no deployed product independently performs comprehensive pre-anesthesia risk assessment reliably at scale without clinician review. |
Monitor patients' responses, including skin color, pupil dilation, pulse, heart rate, blood pressure, respiration, ventilation, or urine output, using invasive and noninvasive techniques.
16CI 7–25 · exposure 22 · augmentation 75 · importance 4.9/5 · click for rater detail
Monitor patients' responses, including skin color, pupil dilation, pulse, heart rate, blood pressure, respiration, ventilation, or urine output, using invasive and noninvasive techniques.
16| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Adoption in perioperative settings is slow. While operating rooms are digitized, regulatory caution, malpractice liability, and the legal requirement for continuous provider oversight constrain velocity. Pilots of supplementary monitoring tools exist but broad production adoption of autonomous monitoring remains rare. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Healthcare, especially perioperative/anesthesia care, is a slow-adopting sector for autonomous AI decision-making due to safety-critical regulation, though monitoring tech itself is ubiquitous. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can meaningfully assist by flagging anomalies in vital signs, alerting to trends, and synthesizing continuous data streams, freeing the anesthesia provider to focus on intervention and clinical judgment. Real-time data integration and pattern detection augment provider situational awareness and vigilance during long procedures. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-enhanced monitors with predictive analytics (e.g., hypotension prediction indices) already help anesthetists anticipate complications and prioritize attention, meaningfully augmenting vigilance. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can assist with reading vital signs from continuous monitors and alerts, the task requires real-time clinical judgment about patient responses, risk stratification, and immediate intervention decisions that depend on context anesthesia providers possess. Full end-to-end automation without anesthesia provider oversight would not meet the 50% time-saving bar because continuous human vigilance and intervention capability are safety-critical. |
| Task automatability | claude-sonnet-5 | 1/5 | This is real-time, hands-on clinical monitoring during anesthesia requiring physical patient contact, invasive line management, and split-second judgment; no current AI system performs this end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | State licensing laws and federal regulations mandate that a qualified anesthesia provider (MD, CRNA, or AA) must personally monitor the patient and be physically present during anesthesia. Liability and patient-safety standards create hard legal barriers to any form of full automation or unsupervised AI delegation. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Anesthesia monitoring during procedures legally requires a licensed anesthesia provider present and responsible for patient safety, making this a hard regulatory and liability barrier. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI monitoring systems are costly to implement and integrate into operating-room infrastructure, and they do not eliminate the anesthesia provider, who remains legally and practically required. The all-in cost (software, hardware, integration, oversight) likely exceeds the marginal productivity gain. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI substitute for the human anesthetist performing this monitoring role; monitoring hardware is already standard equipment, not a replacement for the clinician's cost. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Products exist (monitoring dashboards, AI-assisted vital-sign interpretation) but operate narrowly within established monitor outputs and require anesthesiologist/nurse anesthetist validation. No production system replaces real-time clinical monitoring and decision-making; deployed tools augment rather than substitute. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Deployed patient monitors and alarm systems provide continuous vital sign display and threshold alerts, but the interpretive, integrative clinical monitoring and response by the anesthetist itself is not performed by any autonomous product today. |
Develop anesthesia care plans.
13CI 0–25 · exposure 13 · augmentation 50 · importance 4.7/5 · click for rater detail
Develop anesthesia care plans.
13| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Healthcare remains a laggard sector for autonomous AI automation due to high liability exposure, stringent regulation, and patient safety criticality. Production adoption of AI-authored anesthesia plans is minimal; clinical settings are far more conservative than finance or tech sectors. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Healthcare, especially high-acuity clinical decision-making like anesthesia, is a slow-adopting sector for autonomous AI due to regulatory, safety, and liability constraints, though clinical AI pilots exist elsewhere in the field. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can provide useful augmentation by suggesting protocol options, flagging drug interactions, or summarizing patient comorbidities, thereby speeding plan development and reducing cognitive load. However, the human anesthetist must critically review and finalize the plan, limiting the transformative impact. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can assist by aggregating patient data, flagging drug interactions, and suggesting risk factors to inform the plan, but the nurse anesthetist still performs the core judgment and planning. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can assist with literature review and protocol suggestions, developing anesthesia care plans requires integrating complex patient physiology, medical history, and real-time clinical judgment that current systems cannot reliably do end-to-end. The task demands personalized decision-making under uncertainty that falls well short of the 50% time-saving threshold. |
| Task automatability | claude-sonnet-5 | 1/5 | Developing an anesthesia care plan requires integrating patient history, comorbidities, physical exam findings, and real-time judgment about drug interactions and risk—no current AI system can perform this end-to-end at equal quality. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Strong legal and regulatory barriers exist: anesthesia care planning is a licensed clinical function performed by credentialed nurse anesthetists, often subject to supervision requirements and malpractice liability. Regulatory bodies (state boards, The Joint Commission) and standard of care require a qualified human clinician to author and take responsibility for the anesthesia plan. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Anesthesia care plans must be developed and legally signed off by licensed, credentialed anesthesia professionals under strict medical liability and regulatory frameworks, making this a hard-barrier task. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | The cost of deploying AI for plan generation, including validation oversight, integration into EHRs, and liability management, likely exceeds the marginal benefit compared to a skilled nurse anesthetist developing plans directly. Human expertise remains cheaper for this task given current AI reliability gaps. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | Given the absence of a viable automated substitute and the high liability/oversight cost required for any AI-assisted output, AI is not cheaper than the trained nurse anesthetist performing this task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed clinical product reliably generates anesthesia care plans independently in production settings. AI tools exist for protocol support and decision-aiding, but healthcare providers do not trust autonomous AI systems to author initial anesthesia plans without expert human authorship and oversight. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed clinical product autonomously generates anesthesia care plans for patient use; existing clinical decision-support tools only offer reference information or flag risks, not full plan authorship. |
Select, prepare, or use equipment, monitors, supplies, or drugs for the administration of anesthetics.
9CI 0–18 · exposure 13 · augmentation 50 · importance 4.7/5 · click for rater detail
Select, prepare, or use equipment, monitors, supplies, or drugs for the administration of anesthetics.
9| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Healthcare remains a laggard sector for workflow automation in anesthesia due to regulatory oversight, malpractice risk aversion, and the critical safety implications of any delegation error. Actual displacement of anesthetist tasks through automation is minimal in practice. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Healthcare, especially perioperative and anesthesia care, is a highly regulated, low-digitization-for-physical-tasks sector with minimal AI-driven displacement of hands-on procedures. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI-assisted monitoring dashboards, drug-interaction alerts, and evidence-based dosing recommendations do improve a nurse anesthetist's situational awareness and decision quality. However, augmentation is limited to advisory functions rather than handling selection, preparation, or administration directly. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-enabled monitors and clinical decision-support systems can assist with dosing calculations, alerts, and physiological monitoring, improving safety and efficiency while the anesthetist remains fully in control. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI could assist in drug selection and monitor interpretation, the physical preparation and real-time administration of anesthetics requires embodied action, sensory judgment, and immediate responsiveness to patient conditions that current systems cannot reliably perform end-to-end. Equipment handling, IV placement, and dynamic dose adjustment remain fundamentally human-dependent tasks. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a hands-on physical and clinical decision-making task involving direct patient contact, physical manipulation of equipment and drug administration, which current AI cannot perform end-to-end.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Strong legal and regulatory barriers exist: nurse anesthetists are licensed clinical professionals whose scope includes drug administration and airway management; liability for anesthetic complications falls directly on the credentialed provider. Automation of anesthetic administration is explicitly prohibited without a human's legal responsibility and sign-off. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Administering anesthesia is tightly regulated and legally restricted to licensed CRNAs or anesthesiologists, with severe liability for errors, making this one of the most protected clinical tasks. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | The cost of AI-assisted monitoring and decision support does not offset the loaded wage of a nurse anesthetist, and any system must operate alongside human expertise rather than replace it. Integration and validation costs are substantial relative to the price point for this specialized clinical role. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute performing this physical clinical task, so cost comparison favors the human by default; any AI-assisted monitoring tools add cost rather than replace the anesthetist. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed AI system today reliably manages anesthetic equipment selection, preparation, and administration autonomously. Decision-support products exist for drug dosing, but they operate as advisory tools only; humans remain the executive agent for all physical and clinical decisions in real production settings. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product autonomously selects, prepares, and administers anesthesia; this remains firmly in the domain of licensed clinicians with physical presence required. |
Disassemble and clean anesthesia equipment.
7CI 0–14 · exposure 8 · augmentation 13 · importance 3.6/5 · click for rater detail
Disassemble and clean anesthesia equipment.
7| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Healthcare has been slow to adopt automation for equipment maintenance due to regulatory complexity, high liability costs of equipment failure, and lack of proven automated solutions. Few hospitals have invested in this domain. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Physical equipment maintenance tasks in clinical/surgical settings show minimal AI or robotic adoption, as this domain remains low-digitization and highly manual. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI-guided visual inspection or checklists could assist staff in verifying equipment condition, but the manual disassembly and cleaning work itself offers limited augmentation opportunities; most gains require human hands-on labor rather than AI partnership. |
| Augmentation potential | claude-sonnet-5 | 1/5 | Current AI offers no meaningful assistance for the physical acts of disassembling and cleaning equipment, though checklists or reminder systems might tangentially support scheduling, not the task itself. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Disassembling equipment has some automatable elements (part identification, sequencing), but reassembly, contamination verification, and sterilization validation require dexterous manipulation and quality assurance that current robotics cannot reliably perform at clinical safety standards. Less than 50% time savings is achievable with today's systems. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a physical, manual task requiring hands to disassemble equipment components, clean them per sterilization protocols, and reassemble correctly; no AI system today performs physical manipulation of medical equipment. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Federal regulations (FDA, OSHA) mandate specific cleaning and sterilization protocols for anesthesia equipment; compliance requires documented verification by trained personnel with legal accountability. Automated systems cannot yet satisfy regulatory audit trails and human sign-off requirements. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Sterilization and infection-control protocols in clinical settings typically require trained personnel to follow validated cleaning procedures, and equipment safety/liability concerns create strong barriers to automation absent specialized robotics. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Specialized medical cleaning robots are extremely expensive to acquire and maintain, while a nurse anesthetist performing this task is already salaried and on-site. The capital and integration costs far exceed the marginal labor cost of manual cleaning. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | Without any viable AI or robotic solution for this physical task, there is no AI cost basis to compare; the human remains the only cost-effective option. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed medical robotics system reliably disassembles and cleans anesthesia equipment at scale in production clinical settings. Research prototypes exist, but clinical validation and regulatory approval for autonomous cleaning of critical anesthesia gear remain absent. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product disassembles, cleans, or reassembles physical anesthesia equipment; this remains entirely a manual clinical/technical task performed by humans. |
Select, order, or administer anesthetics, adjuvant drugs, accessory drugs, fluids or blood products as necessary.
6CI 0–11 · exposure 5 · augmentation 50 · importance 4.8/5 · click for rater detail
Select, order, or administer anesthetics, adjuvant drugs, accessory drugs, fluids or blood products as necessary.
6| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | While healthcare is digitizing, anesthesia delivery remains tightly regulated and human-dependent; adoption of AI in anesthesia is limited to narrow decision-support roles in research or pilot settings, not production-level automation of drug selection and administration. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Perioperative anesthesia care is a highly regulated, physically-present clinical function with minimal AI displacement; adoption is limited to adjunct monitoring/decision-support tools. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can provide real-time clinical alerts, drug interaction checking, and dosing recommendations during anesthesia management, improving a practitioner's safety and efficiency, but the human anesthetist remains the primary decision-maker and actor. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-based monitoring systems, dosing calculators, and clinical decision support can assist anesthetists in tracking vitals and suggesting drug interactions, improving safety and efficiency without replacing hands-on judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | While AI could assist in drug selection via clinical decision support, the actual administration of anesthetics, monitoring of patient response, and real-time adjustment of dosing require continuous human judgment and physical intervention that cannot be fully automated end-to-end today. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires real-time physiological monitoring, hands-on drug administration, and split-second clinical judgment during surgery that current AI cannot perform end-to-end without a human physically present. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Administering anesthetics is a controlled, licensed task requiring a Certified Registered Nurse Anesthetist (CRNA) or anesthesiologist; regulatory and legal frameworks mandate human licensure, clinical judgment, and accountability—strong hard barriers prevent substitution by AI alone. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Administering anesthesia is a licensed medical act with strict legal, regulatory, and liability requirements mandating a qualified anesthesia provider be present and accountable. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Current AI systems (decision support, EHR integration) reduce some preparation time but do not eliminate the need for a certified nurse anesthetist; the cost of the AI plus required human oversight does not yet approach an order-of-magnitude saving compared to human performance alone. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI substitute performing this task, so no meaningful cost comparison exists—human labor (with liability coverage) is the only viable option. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Clinical decision support systems exist to recommend drug selections, but no deployed product can independently select, order, and administer anesthetics—the task requires licensed practitioner oversight, physical administration, and immediate response to patient physiology that remains human-centered in all production settings. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product independently selects and administers anesthesia; this remains firmly in the domain of licensed human clinicians with only decision-support tools in limited use. |
Evaluate patients' post-surgical or post-anesthesia responses, taking appropriate corrective actions or requesting consultation if complications occur.
6CI 0–11 · exposure 5 · augmentation 50 · importance 4.6/5 · click for rater detail
Evaluate patients' post-surgical or post-anesthesia responses, taking appropriate corrective actions or requesting consultation if complications occur.
6| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Despite digitization in healthcare, adoption of AI for autonomous post-operative assessment remains very limited. Organizations use monitoring support tools but retain human clinicians in the direct decision-making loop due to patient safety, legal, and liability constraints. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Perioperative/anesthesia care is a highly regulated, physically-grounded clinical environment with minimal autonomous AI adoption for direct patient response tasks. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI-assisted vital sign monitoring, trend alerts, and decision support dashboards can meaningfully assist nurse anesthetists in detecting anomalies and prioritizing interventions, but the human remains accountable for clinical judgment and action. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-enabled monitoring systems, predictive alerts (e.g., early warning scores) and decision-support tools can help flag complications, aiding but not replacing the anesthetist's judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Post-operative assessment requires real-time physiological monitoring, clinical judgment under uncertainty, and immediate intervention decisions that depend on patient-specific context. Current AI cannot reliably detect all complications, make triage decisions, or execute corrective actions without human oversight and decision-making. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires real-time physical patient assessment, hands-on intervention, and clinical judgment under uncertainty that cannot be delegated to current AI systems end-to-end.It is a live, safety-critical monitoring and response task. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Only licensed nurse anesthetists or physicians can legally evaluate post-operative patients and make clinical decisions about complications and corrective actions. Strong regulatory requirements, liability exposure, and the legal mandate for human clinical judgment create near-absolute barriers to autonomous AI substitution. |
| Adoption barriers | claude-sonnet-5 | 5/5 | This is a licensed clinical task with direct legal and patient-safety accountability requiring a credentialed anesthesia provider; regulation and liability make substitution essentially impossible. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | The cost of AI monitoring systems, integration into clinical workflows, and required human oversight and validation would likely exceed or approach the cost of a nurse anesthetist performing the task, especially given low error tolerance in post-operative care. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | AI cannot substitute for the licensed clinician performing this task, so there is no viable cost comparison—human presence is mandatory. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While monitoring algorithms exist for vital signs interpretation, no deployed system independently evaluates post-anesthesia responses and makes clinical decisions without human oversight. Products in this space remain largely supportive dashboards rather than autonomous decision-makers in high-stakes clinical settings. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product autonomously evaluates post-anesthesia patients and takes corrective clinical action; AI is at most a monitoring/alerting aid, not a decision-and-action system. |
Perform pre-anesthetic screenings, including physical evaluations and patient interviews, and document results.
4CI 0–7 · exposure 5 · augmentation 50 · importance 4.6/5 · click for rater detail
Perform pre-anesthetic screenings, including physical evaluations and patient interviews, and document results.
4| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Adoption of AI automation for pre-anesthetic screening in clinical practice is negligible. Anesthesia remains a high-liability domain where regulatory and institutional friction heavily favors human evaluation and documentation. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Healthcare, especially perioperative clinical assessment, has been slow to adopt AI for hands-on clinical tasks despite faster adoption in documentation and administrative support. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist by auto-populating screening templates from EHR data, flagging risk factors, or summarizing prior records, raising documentation efficiency. However, augmentation is limited because the core clinical judgment and patient interaction cannot be meaningfully automated. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can assist with pulling patient history, flagging risk factors from records, and drafting documentation, meaningfully speeding parts of the task while the clinician still performs the exam and interview. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Pre-anesthetic screening requires nuanced physical examination, real-time patient interaction, clinical judgment about risk stratification, and synthesis of complex medical history. Current AI cannot conduct physical exams or build rapport to elicit critical patient-reported symptoms reliably. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires hands-on physical examination, direct patient interaction, and clinical judgment integrating subtle physical findings; no current AI system can perform the physical evaluation component at all. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Anesthesia regulation requires a licensed CRNA or physician anesthetist to perform and sign off on pre-anesthetic evaluations. Legal liability, patient safety standards, and accreditation bodies mandate human clinical judgment and accountability for this gatekeeping task. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Pre-anesthetic assessment is a licensed clinical act with direct patient safety and legal liability implications, requiring a qualified anesthesia provider by law and standard of care. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | AI infrastructure for this task would require integrated computer vision, natural language processing, and clinical decision support, along with significant human oversight. The all-in cost far exceeds a nurse anesthetist's wage for the marginal output. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | Since AI cannot perform the physical evaluation or interview independently, there is no viable cost comparison for full task substitution—a licensed clinician is required. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed AI system performs full pre-anesthetic screening end-to-end. While AI can assist with documentation and some data extraction, the core tasks of physical examination and clinical risk assessment remain human-dependent in all production anesthesia settings. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs pre-anesthetic physical screenings; AI tools exist only for note documentation, not the clinical evaluation itself. |
Obtain informed consent from patients for anesthesia procedures.
3CI 0–6 · exposure 0 · augmentation 38 · importance 4.7/5 · click for rater detail
Obtain informed consent from patients for anesthesia procedures.
3| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Healthcare consent practices are heavily regulated and risk-averse; adoption of AI to replace human consent-obtaining is negligible because legal and professional standards require human accountability. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Healthcare broadly adopts AI slowly for high-stakes clinical/legal interactions, and consent processes remain almost entirely human-driven due to regulatory and liability constraints. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI might assist with generating educational materials or checklists to support the consent conversation, but the core task of verifying patient understanding and obtaining valid legal consent remains primarily human-driven with minimal productivity transformation. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help draft patient-friendly consent materials, translate documents, or answer general procedural questions beforehand, aiding but not replacing the clinician's role in the consent conversation. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Obtaining informed consent requires genuine two-way communication, patient comprehension verification, and legally valid documented agreement that cannot be automated. AI cannot ethically or legally substitute for the human professional's direct interaction and accountability. |
| Task automatability | claude-sonnet-5 | 1/5 | Informed consent requires a licensed clinician to interactively assess patient understanding, answer specific medical questions, and document a legally binding interaction; AI cannot perform this end-to-end today. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Informed consent is a legal and ethical requirement that typically mandates direct interaction with a licensed healthcare provider; liability, regulatory requirements (state medical boards, Joint Commission standards), and malpractice exposure create hard barriers to automation. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Informed consent is a legally and ethically mandated clinician responsibility with strict liability implications, making this one of the most protected tasks in healthcare. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The human anesthetist must conduct this task regardless; AI cannot reduce cost per valid consent obtained because a licensed professional's time and legal liability remain non-negotiable components. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Since a licensed provider must still conduct and document consent, AI cannot substitute the core labor cost; any AI use is supplementary rather than cost-replacing. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product reliably performs informed consent collection end-to-end; this requires human presence, legal accountability, and professional judgment that remains entirely human-dependent in practice. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed clinical product independently obtains legally valid informed consent for anesthesia; at most AI helps generate consent forms or educational material, not the consent act itself. |
Select, order, or administer pre-anesthetic medications.
3CI 0–5 · exposure 5 · augmentation 50 · importance 4.4/5 · click for rater detail
Select, order, or administer pre-anesthetic medications.
3| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Adoption of autonomous medication administration is virtually non-existent in practice; healthcare remains heavily regulated and conservative, with strong resistance to removing human clinical judgment and direct care from drug delivery. Legal and safety requirements have prevented any measurable displacement. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Direct medication administration in anesthesia care shows essentially no AI adoption; healthcare's high-stakes, heavily regulated, physical-contact nature makes this a laggard area for autonomous AI action. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI-based clinical decision support can assist anesthetists by recommending pre-medication regimens based on patient history and vital signs, reducing cognitive load and improving protocol adherence, but the human anesthetist remains responsible for final selection and administration. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI decision-support tools can help flag drug interactions, dosing guidelines, or allergies to inform the anesthetist's medication choices, but the core decision and act remain human-driven. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | While AI could theoretically assist in drug selection via clinical decision support, the actual administration of pre-anesthetic medications involves direct patient contact, real-time vital sign assessment, and clinical judgment in response to patient-specific factors that cannot be automated end-to-end today. The task requires a licensed anesthetist to perform the injection and monitor immediate responses. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires physical administration of controlled substances, real-time patient assessment, and clinical judgment about drug interactions and physiological status—far beyond current AI capabilities to perform end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Strict legal and regulatory barriers mandate that a licensed anesthesia provider (nurse anesthetist or anesthesiologist) must directly administer medication and maintain liability. Scope-of-practice laws, DEA controlled-substance regulations, and malpractice liability create hard legal requirements preventing automation. |
| Adoption barriers | claude-sonnet-5 | 5/5 | This is a tightly regulated, licensed medical act involving controlled substances where only certified anesthesia providers may legally order and administer medications, with severe liability for errors. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The cost of AI-driven decision support plus the required human oversight, verification, and drug administration remains higher than or comparable to the cost of a nurse anesthetist performing the task directly, given liability and the lack of autonomous injection capability. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | AI cannot perform the physical administration or bear clinical liability, so there is no viable cost comparison—human labor is currently the only option. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed product autonomously administers medications to patients. Clinical decision support systems exist for drug selection, but they operate only as assistants to human clinicians and cannot execute the injection or make independent final decisions in real-time clinical settings. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product selects, orders, or administers anesthetic medications autonomously; this remains squarely a human clinical function performed by licensed practitioners. |
Manage patients' airway or pulmonary status, using techniques such as endotracheal intubation, mechanical ventilation, pharmacological support, respiratory therapy, and extubation.
1CI 0–3 · exposure 0 · augmentation 38 · importance 5.0/5 · click for rater detail
Manage patients' airway or pulmonary status, using techniques such as endotracheal intubation, mechanical ventilation, pharmacological support, respiratory therapy, and extubation.
1| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Healthcare organizations show minimal adoption of AI for autonomous airway management, and regulatory frameworks actively prohibit it. Anesthesia remains one of the most conservative and safety-focused domains. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Healthcare adopts AI unevenly and cautiously for high-risk clinical procedures, with anesthesia practice remaining largely human-performed despite growth in monitoring analytics. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI can provide some assistance through patient monitoring alerts and pharmacological reminders, but these represent marginal enhancements to a task where the human clinician retains full decision and execution authority. Most airway management remains unaugmented. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-enabled monitors, predictive algorithms for hemodynamic/respiratory instability, and smart ventilators can assist decision-making and alert clinicians, improving situational awareness during the task. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Airway and pulmonary management requires real-time clinical judgment, manual dexterity, and rapid response to patient physiology changes. Current AI systems cannot perform endotracheal intubation, adjust mechanical ventilation parameters based on live patient feedback, or handle emergent airway crises autonomously. |
| Task automatability | claude-sonnet-5 | 1/5 | This is hands-on, high-stakes physical intervention requiring real-time tactile and clinical judgment during anesthesia; no AI system can physically intubate, adjust ventilators in situ, or manage acute airway crises end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Hard legal and professional barriers exist: only licensed anesthesia professionals can manage airways in most jurisdictions, and patient safety liability is extreme. Medical boards and regulations explicitly require licensed human oversight of airway management. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Airway management is a licensed, safety-critical medical procedure requiring certified anesthesia providers, with direct legal and patient-safety liability precluding non-human execution. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The cost of developing and validating an autonomous airway management system would vastly exceed the wage of a nurse anesthetist. Liability, regulatory approval, and the high stakes of failure make this economically infeasible today. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI substitute performing this task, so cost comparison favors the human entirely; any AI role would only add cost as a monitoring aid. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI system can reliably manage a patient's airway or pulmonary status end-to-end. While monitoring and alert systems exist, the core manual and decision-making components remain entirely human-dependent in clinical practice. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs airway management or extubation; clinical decision-support tools exist for monitoring but not for executing the physical/pharmacological management task itself. |
Administer post-anesthesia medications or fluids to support patients' cardiovascular systems.
1CI 0–3 · exposure 0 · augmentation 50 · importance 4.5/5 · click for rater detail
Administer post-anesthesia medications or fluids to support patients' cardiovascular systems.
1| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Healthcare settings have seen minimal displacement of medication administration by AI; clinical care remains human-centric due to regulatory requirements, patient safety imperatives, and the complexity of individualized titration. Adoption velocity for autonomous medication administration remains negligible. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Healthcare bedside clinical intervention is a slow-adopting sector for autonomous AI due to safety regulation and physical hands-on requirements, though decision-support tools are slowly emerging. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can provide useful assistance through hemodynamic monitoring, drug interaction alerts, dose calculations, and decision-support tools that help anesthetists choose and administer medications more safely. However, the human anesthetist must remain in the loop for all critical decisions and physical acts. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-enabled monitoring systems and predictive analytics can help alert nurse anesthetists to hemodynamic changes, aiding decision-making even though the human must administer treatment. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Administering post-anesthesia medications and fluids requires real-time clinical judgment, direct patient contact, physical drug delivery, and immediate response to changing vital signs—functions that current AI systems cannot perform end-to-end autonomously. No AI system today can independently assess hemodynamic status, select appropriate agents, establish IV access, and titrate infusions without human oversight. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires real-time physical assessment, hands-on drug administration, and continuous clinical judgment in a dynamic physiological state that current AI cannot perform end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Administering medications is a scope-of-practice requirement for licensed nurses and anesthetists; federal and state regulations mandate human licensure, direct assessment, and accountability. Liability for adverse outcomes, patient safety standards, and the legal requirement for a licensed professional to verify, prepare, and administer drugs create hard legal barriers. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Administering medications and fluids to manage a patient's cardiovascular status is a licensed medical act with direct liability, requiring a credentialed anesthesia provider by law and hospital policy. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The task demands physical intervention, liability assumption, and real-time clinical expertise that no AI system can replace cost-effectively. Oversight and integration costs would far exceed the benefit of any partial automation. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI substitute performing this physical task, so cost comparison favors the human entirely; any AI-adjacent monitoring tools add cost rather than replace the clinician. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | While AI can assist with monitoring and decision support in clinical settings, no deployed product reliably administers medications or fluids to real patients without a licensed human performing the actual clinical act. This remains a task requiring direct human agency and accountability. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product autonomously administers post-anesthesia cardiovascular medications; this remains firmly in the domain of licensed clinicians performing physical interventions. |
Select and prescribe post-anesthesia medications or treatments to patients.
1CI 0–3 · exposure 0 · augmentation 50 · importance 4.5/5 · click for rater detail
Select and prescribe post-anesthesia medications or treatments to patients.
1| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Healthcare adoption of autonomous AI for prescriptive decisions remains minimal due to regulatory and liability constraints. Deployment in this domain is negligible; clinical decision support has not displaced prescriber decision-making. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Healthcare, especially acute perioperative care, adopts autonomous AI decision-making slowly due to regulatory and safety constraints, though clinical decision-support pilots exist. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI-powered clinical decision support (e.g., drug interaction checkers, dosing calculators, protocol reminders) can meaningfully assist anesthetists in reviewing options and avoiding errors, though the final clinical selection and prescription remain the human's responsibility. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can assist by suggesting dosing ranges, flagging drug interactions, or summarizing patient history, improving efficiency while the anesthetist retains prescribing authority and final judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Prescribing post-anesthesia medications requires real-time clinical judgment, patient assessment, and legal authorization to prescribe—functions that current AI systems cannot perform end-to-end. While AI can assist in suggesting treatment options, the actual selection and prescription remain the licensed provider's legal and clinical responsibility. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires real-time clinical judgment based on patient vitals, comorbidities, and surgical context; no off-the-shelf AI system can safely select/prescribe post-anesthesia treatments end-to-end today. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Prescribing is a regulated medical act that requires a licensed practitioner's credential and signature. Legal, liability, and regulatory frameworks explicitly mandate human professional judgment and accountability for medication selection and administration. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Prescribing is a legally restricted act requiring licensure (CRNA/physician), with strict liability, regulatory, and DEA/scope-of-practice constraints preventing AI from independently performing it. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The task requires a licensed nurse anesthetist's expertise and legal authority; there is no cost-effective AI substitute. Any AI support would be additive to, not replacing, the anesthetist's labor and liability. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | Because AI cannot independently perform this prescribing task, there is no standalone AI cost basis to compare; a licensed clinician remains mandatory, making AI substitution cost irrelevant/non-competitive. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI product reliably prescribes medications independently in clinical practice. Clinical decision support systems exist but do not make final prescriptive decisions; a credentialed provider must evaluate the patient and issue the prescription. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed clinical product autonomously prescribes post-anesthesia medications; decision-support tools exist only as adjuncts to licensed provider judgment. |
Insert peripheral or central intravenous catheters.
1CI 0–3 · exposure 0 · augmentation 38 · importance 4.3/5 · click for rater detail
Insert peripheral or central intravenous catheters.
1| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Despite decades of robotics research, hospitals continue to rely entirely on human clinicians for IV insertion; adoption remains negligible because the task requires licensed personnel and there are no mature competing solutions in production. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Healthcare procedural tasks involving direct physical patient contact see slow AI adoption; even assistive imaging/vein-visualization tools have limited penetration in anesthesia practice. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI offers minimal assistance for this highly manual, judgment-rich task; ultrasound guidance and vein-finding devices help but are not AI-driven, and the clinician must still execute the insertion with full manual control and responsibility. |
| Augmentation potential | claude-sonnet-5 | 3/5 | Vein-visualization devices and ultrasound guidance (some enhanced by AI-based image processing) can assist clinicians in locating vessels, improving first-attempt success rates, though the human still performs the procedure. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Inserting IV catheters requires precise manual dexterity, real-time tactile feedback, and immediate adjustment to anatomical variation and patient response—capabilities current AI systems cannot replicate. No end-to-end automation exists; this remains a manual clinical procedure. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a hands-on invasive procedure requiring physical dexterity, tactile feedback, and real-time anatomical judgment; no current AI system can perform physical catheter insertion. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Federal and state regulations require a licensed healthcare provider (MD, DO, NP, PA, or certified nurse anesthetist) to perform or directly supervise IV catheter insertion; liability and patient safety standards create hard legal barriers to substitution. |
| Adoption barriers | claude-sonnet-5 | 5/5 | This is an invasive medical procedure legally restricted to licensed clinicians, with high liability for errors like infection or vascular injury, making autonomous automation legally and ethically barred. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The hardware and oversight costs for any robotic or AI-assisted IV insertion system far exceed the loaded wage of a nurse anesthetist performing the task directly and quickly. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI-only substitute for the physical act, so any comparison favors human cost since AI cannot perform the task at all. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product reliably performs IV catheter insertion autonomously in clinical settings. While robotic research exists, no production system demonstrates safe, independent IV insertion at scale in real hospitals. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs physical IV catheter placement autonomously; robotic vein-finding assist devices exist only in limited research/early-commercial contexts, not as autonomous insertion tools. |
Respond to emergency situations by providing airway management, administering emergency fluids or drugs, or using basic or advanced cardiac life support techniques.
0CI 0–0 · exposure 0 · augmentation 25 · importance 4.9/5 · click for rater detail
Respond to emergency situations by providing airway management, administering emergency fluids or drugs, or using basic or advanced cardiac life support techniques.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Emergency anesthesia response is a core clinical duty that cannot be delegated to AI; adoption velocity for automation is effectively zero because the task involves irreplaceable human licensing, judgment, and physical capability requirements. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Acute care and emergency medicine involve physical, high-stakes intervention with essentially no AI displacement in production; adoption in this specific task domain is negligible. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could potentially assist with decision support (e.g., suggesting drug doses or alerting to patient parameters), but the task statement focuses on active execution of airway management and emergency interventions where a human clinician remains fully engaged and responsible. Augmentation is limited to peripheral support. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can support monitoring, decision-support algorithms, or documentation during emergencies, but offers minimal direct assistance during the acute hands-on intervention itself. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires real-time clinical judgment, physical intervention (airway management, drug administration, cardiac life support), and immediate adaptation to patient physiology. Current AI cannot perform any of these manual interventions or make the split-second clinical decisions autonomously in emergency contexts. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires real-time physical intervention (airway management, drug administration, hands-on CPR) in life-threatening situations, which current AI cannot physically perform or safely direct end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | This task is legally restricted to licensed nurse anesthetists (CRNAs) or anesthesiologists who must personally perform or directly oversee airway management and emergency interventions. Regulatory and legal requirements explicitly mandate human licensure and accountability for these life-critical functions. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Emergency clinical interventions require licensed, credentialed practitioners under strict medical, legal, and liability frameworks; only authorized humans may legally perform these acts. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | AI has no role in the core execution of this emergency response task, and any AI support would require human oversight and intervention. The cost of AI infrastructure adds to rather than reduces the cost of delivering emergency anesthesia care. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute performing this task, so no meaningful cost comparison exists; the human clinician remains the only option. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI system can independently manage airways, administer drugs, or perform cardiac life support in emergency settings. These tasks demand licensed human practitioners and direct physical engagement with patients that AI cannot currently execute. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs emergency airway management or administers emergency drugs autonomously; this remains entirely in the domain of human clinicians. |
Perform or manage regional anesthetic techniques, such as local, spinal, epidural, caudal, nerve blocks and intravenous blocks.
0CI 0–0 · exposure 0 · augmentation 38 · importance 4.7/5 · click for rater detail
Perform or manage regional anesthetic techniques, such as local, spinal, epidural, caudal, nerve blocks and intravenous blocks.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Healthcare delivery, especially in operating rooms, remains heavily regulated and resistant to full automation of clinical procedures; adoption of AI in anesthesia has been limited to non-clinical tasks like documentation and scheduling. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Anesthesia delivery is a highly regulated, physically embodied clinical task with essentially no autonomous AI deployment or displacement occurring in practice. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could provide limited assistance (e.g., ultrasound guidance visualization, landmark identification, pre-procedure planning) but the core task of needle placement, anatomical navigation, and patient management remains human-dependent with minimal augmentation benefit today. |
| Augmentation potential | claude-sonnet-5 | 3/5 | Ultrasound-guided imaging software and AI-assisted needle-tracking tools can help clinicians visualize anatomy more precisely during nerve blocks, offering moderate procedural assistance. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Regional anesthesia requires real-time physical manipulation, anatomical assessment, patient interaction, and immediate response to complications. Current AI cannot perform needle placement, palpate anatomical landmarks, or manage dynamic patient responses during the procedure. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a hands-on invasive medical procedure requiring physical dexterity, tactile feedback, and real-time judgment; no current AI system can physically place a spinal or epidural needle or manage a nerve block. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Regulatory frameworks require a licensed, credentialed healthcare provider (nurse anesthetist or anesthesiologist) to perform and take legal responsibility for regional anesthesia; liability, patient safety, and scope-of-practice laws create hard legal barriers to substitution. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Administering regional anesthesia is a licensed medical procedure with strict scope-of-practice laws, malpractice liability, and mandatory human clinical judgment and physical presence. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | AI systems capable of assisting with this task (if they existed) would require significant infrastructure, real-time imaging integration, and oversight, making them more expensive than employing a trained nurse anesthetist. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI substitute performing this physical task, so cost comparison favors the human entirely; any AI-assisted imaging tool adds cost on top of the clinician, not instead of them. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI system performs regional anesthetic techniques independently or manages the full clinical workflow. This task requires licensed clinical judgment, physical intervention, and real-time patient monitoring that no current product handles end-to-end. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs or manages regional anesthesia; AI in this space is limited to research-stage image guidance aids, not autonomous execution. |
Prepare prescribed solutions and administer local, intravenous, spinal, or other anesthetics, following specified methods and procedures.
0CI 0–0 · exposure 0 · augmentation 38 · importance 4.6/5 · click for rater detail
Prepare prescribed solutions and administer local, intravenous, spinal, or other anesthetics, following specified methods and procedures.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Healthcare remains a laggard in automation due to regulatory barriers, malpractice liability, and the requirement for continuous human judgment and presence during invasive procedures. No meaningful displacement of anesthetic administration is occurring. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Healthcare's hands-on clinical procedures, especially anesthesia delivery, show minimal AI adoption due to safety-critical physical execution and heavy regulation. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could assist with pre-operative protocol review, patient history analysis, or drug interaction checking, but the core task—preparing and administering anesthetics—offers limited room for augmentation since the human must perform it end-to-end with direct accountability. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-assisted monitoring, dosage calculators, and clinical decision support can help nurse anesthetists track vitals and optimize dosing, but the core preparation and administration remains manual. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires direct patient contact, real-time clinical judgment, and hands-on administration of controlled substances. No AI system can physically administer anesthetics or adapt dosing in real time based on live patient response, which is the core of the task. |
| Task automatability | claude-sonnet-5 | 1/5 | Physically preparing and administering anesthetics requires hands-on manipulation of patients, needles, IV lines, and real-time physiological monitoring that current AI cannot perform end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Administering anesthetics is a legally protected function requiring a licensed anesthetist (CRNA or physician) to be present and responsible. Regulatory frameworks (FDA, state licensing boards) and liability law mandate human oversight and sign-off. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Administering anesthesia is a tightly regulated, licensed medical procedure with severe liability exposure and legal requirements for a certified human provider (CRNA or anesthesiologist) to perform or directly supervise it. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The loaded cost of a nurse anesthetist is high, but the liability, regulatory oversight, and need for a human to be physically and legally present during anesthetic administration mean AI cannot substitute on cost grounds alone. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute performing this physical clinical act, 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 | While AI might assist in dose calculation or protocol review, no deployed product performs anesthetic administration end-to-end. The task inherently requires a licensed human operator managing a controlled, high-stakes intervention. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product administers anesthesia autonomously; this remains firmly in the domain of licensed human clinicians with physical presence and dexterity. |
Calibrate and test anesthesia equipment.
0CI 0–0 · exposure 0 · augmentation 25 · importance 4.6/5 · click for rater detail
Calibrate and test anesthesia equipment.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Healthcare delivery remains heavily regulated and risk-averse; anesthesia equipment maintenance is handled by specialized biomedical staff or manufacturer service engineers. Adoption of autonomous AI for safety-critical medical device calibration is essentially zero. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Healthcare equipment verification is a highly regulated, physical, low-digitization task with essentially no AI/automation penetration in current practice. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could potentially assist by providing checklist reminders or monitoring logs, but the core task—hands-on equipment calibration and functional testing—requires human judgment and physical presence. Augmentation opportunities are limited and marginal. |
| Augmentation potential | claude-sonnet-5 | 2/5 | Smart anesthesia machines may provide digital self-diagnostic readouts and alerts that assist the clinician, but this is limited built-in equipment feedback rather than a general AI augmentation tool. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Calibrating and testing anesthesia equipment requires physical manipulation, sensory inspection (listening for leaks, visual checks), and real-time troubleshooting of complex medical machinery. Current AI cannot perform these hands-on mechanical tasks or ensure safety-critical equipment certification. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a physical, hands-on safety-critical procedure requiring manual inspection, connection checks, and functional testing of gas and delivery systems; no AI system can perform this end-to-end today. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Anesthesia equipment calibration is governed by FDA regulations, hospital accreditation standards (JCAHO), and manufacturer specifications that typically require credentialed human technicians or anesthetists to sign off on safety certification. Liability and regulatory requirements create hard barriers to automation. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Anesthesia equipment checks are mandated by patient safety protocols and require a licensed anesthesia provider to verify machine function before use, with high liability for equipment failure. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The cost of developing, deploying, and maintaining AI-robotic systems to handle anesthesia equipment calibration would far exceed the loaded wage of a nurse anesthetist performing this task periodically as part of their role. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI substitute for the physical checks involved, so any AI cost comparison is moot—human performance remains the only viable and thus cheaper option currently. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI systems currently perform independent calibration and testing of anesthesia equipment in clinical settings. This task demands physical interaction with hardware and regulatory compliance verification that exists only in manual or semi-automated legacy systems, not AI-driven solutions. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs physical calibration and testing of anesthesia machines; this remains a manual clinical task performed by trained personnel. |
Insert arterial catheters or perform arterial punctures to obtain arterial blood samples.
0CI 0–0 · exposure 0 · augmentation 25 · importance 4.3/5 · click for rater detail
Insert arterial catheters or perform arterial punctures to obtain arterial blood samples.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Healthcare remains conservative in delegating invasive procedures to automation; adoption of fully autonomous arterial access is not occurring in practice, and regulatory and liability structures actively discourage it. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Physical invasive procedures in anesthesia care are among the least digitized/automatable clinical tasks, with essentially no production-level AI or robotic adoption for this specific action. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could assist through ultrasound image analysis to identify vessel location and provide guidance, but the core manual insertion remains human-performed. Augmentation is limited to pre-procedure localization rather than transforming the task itself. |
| Augmentation potential | claude-sonnet-5 | 2/5 | Ultrasound guidance systems and vein/artery visualization tools (some AI-enhanced) can assist in locating vessels, offering modest procedural support, but the core manual task remains unaided by AI. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires precise manual dexterity, real-time visual-tactile feedback, and anatomical judgment to locate and puncture an artery safely. Current AI systems cannot perform physical manipulation or adapt to real-time physiological feedback; this remains entirely dependent on human motor control. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a hands-on invasive medical procedure requiring physical dexterity, tactile feedback, and real-time adaptation to patient anatomy; no current AI system can perform arterial cannulation end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | This is a licensure-protected clinical procedure that requires a licensed practitioner (nurse anesthetist, physician, or credentialed clinician) to perform or directly supervise. Legal and regulatory frameworks mandate human accountability and professional judgment. |
| Adoption barriers | claude-sonnet-5 | 5/5 | This is an invasive clinical procedure legally restricted to licensed, credentialed practitioners, with high liability and direct patient-contact requirements. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The infrastructure required for robotic arterial access (specialized equipment, setup, calibration, human oversight) far exceeds the cost of a skilled nurse anesthetist performing the procedure directly. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI-driven substitute for this physical procedure, so a per-task cost comparison favors the human clinician entirely. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product can autonomously perform arterial catheterization or puncture. Surgical robotics exist but require continuous human control and oversight; no system operates end-to-end without an operator at the controls. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs arterial catheter insertion or puncture; robotic-assisted vascular access remains research-stage with very limited clinical deployment. |
Discharge patients from post-anesthesia care.
0CI 0–0 · exposure 0 · augmentation 25 · importance 4.2/5 · click for rater detail
Discharge patients from post-anesthesia care.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Healthcare adoption of AI for clinical discharge decisions remains minimal; institutional and regulatory conservatism, patient safety liability concerns, and the requirement for human licensure create structural resistance to displacement in this critical care setting. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Clinical anesthesia care is a highly regulated, low-digitization-for-decision-making sector where AI adoption for discharge authority is essentially nonexistent. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | While AI-powered vital sign monitoring and trend alerts could assist anesthetists in tracking patient recovery metrics, the core discharge decision—integrating multiple clinical signals and patient-specific factors—remains fundamentally a human clinical judgment task with limited augmentation potential. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can support monitoring data aggregation and vital sign trend analysis to inform the decision, but it plays a minor assistive role rather than transforming the task. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Discharging patients from post-anesthesia care requires continuous clinical judgment, assessment of patient responsiveness, pain levels, nausea, vital sign stability, and individual risk factors—decisions that depend on real-time human observation and patient interaction that current AI cannot perform end-to-end without human oversight. |
| Task automatability | claude-sonnet-5 | 1/5 | Discharging a patient from post-anesthesia care requires real-time clinical judgment, physical assessment, and legal accountability that no current AI system can perform end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Strong legal and regulatory barriers exist: only licensed nurse anesthetists or physicians can authorize patient discharge from post-anesthesia care, making this a task with hard professional licensing requirements and liability constraints that prevent substitution. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Discharge decisions require a licensed anesthesia provider's sign-off with direct legal and patient-safety liability, making this a hard-barrier clinical task. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | AI systems cannot reduce the cost of this task because a qualified human must legally perform or directly supervise discharge assessment; the labor cannot be substantially displaced, making any AI deployment purely supplementary to the required human cost. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI alternative performing this 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 | 1/5 | No deployed AI system independently manages post-anesthesia discharge decisions in production; this task demands licensed clinical expertise, legal accountability, and real-time patient monitoring that only trained anesthesia personnel can reliably perform. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs autonomous discharge decisions from anesthesia recovery; this remains firmly a clinician task with no production AI substitute. |
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.