Critical Care Nurses

29-1141.03
Median wage $97,550/yr3,379,720 employed (US)Rank #836 of 923 scored · top 91% by substitution

Provide specialized nursing care for patients in critical or coronary care units.

Sub-scores

0–100 · band = confidence interval from rater disagreement

Substitution13
Exposure13
Augmentation50

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

29 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%12

panel mean rating 1.5/5 → substitution pressure 12/100

Technical feasibility todayw 20%15

panel mean rating 1.6/5 → substitution pressure 15/100

Cost vs. human wagew 15%11

panel mean rating 1.4/5 → substitution pressure 11/100

Adoption barriersw 20%inverted — strong barriers lower the score11

panel mean rating 4.5/5 (barrier strength) → substitution pressure 11/100

Sector adoption velocityw 10%19

panel mean rating 1.7/5 → substitution pressure 19/100

Task breakdown (29 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.

Participate in professional organizations and continuing education to improve practice knowledge and skills.

43

CI 3–84 · exposure 45 · augmentation 75 · importance 3.7/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510014/5Healthcare organizations are rapidly adopting AI-powered learning management systems, competency tracking, and personalized education recommendation platforms as part of workforce development infrastructure. This reflects broader digital transformation in healthcare and professional services.
Sector adoption velocityclaude-sonnet-52/5Healthcare adoption of AI is growing but this specific task—personal professional engagement—is not an area of AI substitution or displacement.
Augmentation potentialclaude-haiku-4-5-202510015/5AI significantly augments nurse participation by recommending tailored courses based on specialty and gaps, tracking CEU progress automatically, organizing schedule conflicts, and surfacing relevant journal articles or webinars—all while the nurse retains full control over learning decisions and implementation.
Augmentation potentialclaude-sonnet-53/5AI can help nurses find relevant courses, summarize research literature, and prepare for exams or continuing education modules, aiding but not replacing the human engagement.
Task automatabilityclaude-haiku-4-5-202510015/5Identifying relevant professional organizations, curating continuing education opportunities, organizing learning schedules, and summarizing educational content are readily automatable tasks that current AI systems can handle end-to-end with significant time savings. AI can monitor professional updates, aggregate educational resources, and create personalized learning plans more efficiently than manual processes.
Task automatabilityclaude-sonnet-51/5This involves personal professional development, networking, and skill-building through human engagement with organizations and courses; AI cannot participate in or fulfill this activity on a nurse's behalf.
Adoption barriersclaude-haiku-4-5-202510012/5No regulatory requirement mandates that a human personally select or organize their own continuing education; many organizations already use automated learning management systems. The primary barrier is organizational preference to involve nurses in choice and self-direction, not legal or liability constraints.
Adoption barriersclaude-sonnet-55/5Licensing boards and professional organizations require personal participation, attendance, and completion by the individual nurse to maintain certification and licensure.
Cost vs. human wageclaude-haiku-4-5-202510014/5AI-driven learning platforms, content curation systems, and automated scheduling tools cost substantially less than compensating a nurse for research time and organizational overhead. The per-task cost is likely one-tenth that of manual time investment by a nurse.
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute performing this task, so no meaningful cost comparison exists; the human must personally engage.
Technical feasibility todayclaude-haiku-4-5-202510014/5Deployed products (learning management systems, content recommendation engines, professional networking platforms with AI features) reliably perform components of this task, including identifying courses, tracking certifications, and organizing educational materials. However, actual participation and knowledge synthesis still require human engagement, limiting full end-to-end deployment.
Technical feasibility todayclaude-sonnet-51/5No deployed product performs professional membership participation or continuing education requirements for a nurse; this is an inherently human, credential-bound activity.

Document patients' medical histories and assessment findings.

35

CI 25–45 · exposure 38 · augmentation 63 · importance 4.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5While health systems deploy EHR and voice-to-text tools broadly, autonomous AI documentation in critical care remains rare; adoption is limited to incremental efficiency gains, not substitution, because clinicians must verify every clinical claim for legal and patient-safety reasons.
Sector adoption velocityclaude-sonnet-53/5Healthcare has moderate AI adoption for documentation (ambient scribes, EHR-integrated tools) but critical care settings lag due to complexity, regulatory scrutiny, and safety-critical nature of the work.
Augmentation potentialclaude-haiku-4-5-202510013/5Speech recognition, templating, and AI-suggested documentation elements moderately assist nurses by reducing transcription burden and prompting completeness, though the high-acuity context and accountability requirements mean the nurse remains the primary decision-maker on what to document.
Augmentation potentialclaude-sonnet-54/5AI documentation assistants can significantly speed up drafting and organizing notes from verbal input or structured data, letting nurses focus more time on direct patient care while reviewing AI-generated drafts.
Task automatabilityclaude-haiku-4-5-202510012/5While EHR systems can capture structured vital signs and some templated fields, critical care patient documentation requires interpreting complex clinical findings, integrating real-time physiologic changes, and exercising clinical judgment about what matters—tasks that current AI cannot reliably perform end-to-end without substantial human oversight.
Task automatabilityclaude-sonnet-53/5AI can draft documentation from dictation or structured inputs, but capturing accurate assessment findings from a critical care patient still requires clinical judgment, verification, and hands-on evaluation that current systems cannot fully replace.
Adoption barriersclaude-haiku-4-5-202510014/5Medical record documentation is heavily regulated (HIPAA, state nursing boards, malpractice liability); clinicians remain legally accountable for the accuracy and completeness of the medical record, and institutional risk management typically requires a licensed nurse to author and sign critical care assessments.
Adoption barriersclaude-sonnet-54/5Medical record accuracy and legal liability mean a licensed nurse must verify and often personally attest to documentation accuracy, especially in critical care where errors have high stakes.
Cost vs. human wageclaude-haiku-4-5-202510012/5While transcription and template tools reduce typing time marginally, the cost of AI infrastructure, integration, and the extensive clinician review required to ensure accuracy and liability compliance approaches or exceeds the time saved relative to nursing wages.
Cost vs. human wageclaude-sonnet-53/5AI scribe/documentation tools cost a fraction of nurse time but still require licensed nurse review, editing, and sign-off, so net cost savings are moderate rather than order-of-magnitude given oversight requirements.
Technical feasibility todayclaude-haiku-4-5-202510012/5Voice-to-text and EHR templates exist, but AI systems struggle with the contextual medical accuracy, completeness, and legal defensibility required for critical care notes; deployed products function as rough drafts requiring significant clinician review rather than replacing the documentation task.
Technical feasibility todayclaude-sonnet-53/5Ambient clinical documentation tools (e.g., AI scribes) are deployed in some hospitals and reliably draft notes from clinician-patient interactions, but critical care documentation requires integrating complex vitals, device data, and rapid changes, which remains narrower in scope and error-prone.

Document patients' treatment plans, interventions, outcomes, or plan revisions.

33

CI 25–41 · exposure 38 · augmentation 75 · importance 4.1/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510013/5Healthcare institutions are piloting AI documentation tools, but adoption remains cautious and incomplete; production deployments are increasing in large health systems but remain far from ubiquitous, particularly in critical care where stakes are highest.
Sector adoption velocityclaude-sonnet-52/5Healthcare, especially acute/critical care, has historically been a slower adopter of AI tools due to regulatory, safety, and workflow integration barriers, though scribe pilots are growing.
Augmentation potentialclaude-haiku-4-5-202510014/5AI can meaningfully assist critical care nurses by auto-populating structured fields, suggesting documentation templates, and flagging missing data, which can reduce manual typing time while the nurse retains full clinical responsibility and oversight of the final record.
Augmentation potentialclaude-sonnet-54/5AI-assisted documentation and summarization tools can meaningfully speed up charting and reduce administrative burden while the nurse retains responsibility for accuracy and clinical decisions.
Task automatabilityclaude-haiku-4-5-202510012/5Documentation of treatment plans and outcomes requires understanding complex clinical context, patient history, and care decisions that AI can partially assist with (structured data entry, template completion), but cannot reliably generate end-to-end without substantial human review and revision due to clinical accuracy and liability concerns.
Task automatabilityclaude-sonnet-53/5AI can draft documentation from clinical notes or ambient dictation, but nurses must verify accuracy, add clinical judgment, and finalize legally binding records, limiting full end-to-end automation.olgen
Adoption barriersclaude-haiku-4-5-202510014/5Medical documentation has strong regulatory and liability requirements (HIPAA, CMS, state nursing boards); nursing licensure requirements mean a licensed nurse must review and sign off on clinical documentation, creating a hard legal barrier to full automation.
Adoption barriersclaude-sonnet-54/5Medical record documentation is legally required to be authored or verified by licensed clinical staff, with significant liability exposure for inaccuracies in critical care settings.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI-assisted documentation tools reduce some clerical burden but still require clinician verification and editing time; the per-task cost reduction is modest compared to the loaded wage of a critical care nurse, and integration overhead remains high.
Cost vs. human wageclaude-sonnet-52/5AI documentation tools reduce time but still require licensed nurse review and correction, plus integration with EHR systems, so net savings are moderate rather than order-of-magnitude cheaper.
Technical feasibility todayclaude-haiku-4-5-202510012/5While EHR systems have some auto-documentation and templating features, current AI deployments in critical care are limited to drafting assistance rather than autonomous documentation; production use requires extensive clinical oversight and editing, not reliable standalone performance.
Technical feasibility todayclaude-sonnet-53/5Ambient documentation and clinical scribe products exist and are used in some hospital settings, but critical care's complexity and rapid changes mean adoption is narrower and error-checking remains essential.

Plan, provide, or evaluate educational programs for nursing staff, interdisciplinary health care team members, or community members.

28

CI 25–30 · exposure 25 · augmentation 63 · importance 3.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Healthcare organizations adopt AI-assisted content tools slowly; most educational program planning remains human-led due to accountability requirements, learner variability, and cultural preference for expert educators in clinical domains.
Sector adoption velocityclaude-sonnet-52/5Healthcare training and education functions are adopting AI tools slowly, with pilots for content creation but limited integration in nursing education programs at scale.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can meaningfully assist nurse educators by generating draft curricula, analyzing learning data, creating assessment tools, and providing evidence-based content suggestions, supporting their productivity while educators retain judgment on clinical competency and evaluation standards.
Augmentation potentialclaude-sonnet-54/5AI can significantly assist by drafting training materials, generating quiz questions, summarizing best practices, and creating educational content, saving nurse educators substantial preparation time.
Task automatabilityclaude-haiku-4-5-202510012/5AI can assist with content generation, curriculum design templates, and learning material creation, but planning and evaluating educational programs requires understanding of learner needs, organizational context, regulatory requirements, and adaptive feedback loops that current AI cannot fully handle end-to-end with 50% time savings at equal quality.
Task automatabilityclaude-sonnet-52/5AI can help draft curricula or content but planning, delivering, and evaluating education for clinical staff requires contextual judgment, live facilitation, and assessment of learner competency that current systems cannot fully replace.
Adoption barriersclaude-haiku-4-5-202510014/5Healthcare education programs are subject to accreditation standards (ACPE, CCNE), state nursing board requirements, and institutional compliance obligations that typically require licensed professionals to design and certify competency outcomes, creating regulatory barriers to full automation.
Adoption barriersclaude-sonnet-53/5No strict licensing requirement for AI use in curriculum design, but clinical education often requires accredited nurse educators and institutional sign-off, creating moderate friction.
Cost vs. human wageclaude-haiku-4-5-202510012/5While AI can reduce some content creation costs, the setup, customization, oversight, and quality assurance for clinical education programs remain labor-intensive, making total cost comparable to or higher than a human educator's loaded cost.
Cost vs. human wageclaude-sonnet-52/5AI can cheaply generate draft materials, but the overall task including live teaching, competency evaluation, and interdisciplinary coordination still requires expensive human oversight and delivery, keeping costs comparable to human-led efforts.
Technical feasibility todayclaude-haiku-4-5-202510012/5Products exist for generating training materials and simple quizzes, but no deployed system reliably handles the full scope of planning, delivery, and evaluation of nurse-specific educational programs in clinical settings with the rigor required for healthcare competency assurance.
Technical feasibility todayclaude-sonnet-52/5Some LMS platforms use AI for content generation and quiz creation, but no deployed product autonomously plans and evaluates full educational programs for healthcare teams in production.

Monitor patients' fluid intake and output to detect emerging problems, such as fluid and electrolyte imbalances.

26

CI 25–28 · exposure 25 · augmentation 63 · importance 4.8/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Despite digitization of healthcare, ICU monitoring remains highly manual and nurse-centered; AI-assisted dashboards exist in some facilities but deep automation of clinical decision-making around fluid/electrolyte management is rare, with adoption limited to pilot projects rather than standard practice.
Sector adoption velocityclaude-sonnet-53/5Hospitals are adopting predictive analytics and EWS/deterioration algorithms at a moderate pace, with some large health systems in production, but broad ICU-wide adoption is still uneven.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist by automatically aggregating and visualizing fluid balance trends, alerting to statistical outliers, and flagging labs needing review—useful for productivity—but the nurse must still interpret context, correlate with clinical status, and make decisions about intervention.
Augmentation potentialclaude-sonnet-54/5AI-driven alerts and trend analysis on fluid balance and lab values can meaningfully help nurses catch emerging imbalances earlier, augmenting vigilance without removing the human from the loop.
Task automatabilityclaude-haiku-4-5-202510012/5Current AI can aggregate and analyze numerical fluid intake/output data and flag statistical anomalies, but the task requires real-time patient assessment, clinical judgment about emerging imbalances, and integration with physical examination findings that AI cannot reliably perform end-to-end without substantial human oversight and interpretation.
Task automatabilityclaude-sonnet-52/5AI can flag abnormal trends from intake/output data feeding into EHRs, but hands-on measurement (catheter output, IV checks, physical assessment) and clinical judgment for intervention remain human tasks, so full end-to-end automation is not achievable today.
Adoption barriersclaude-haiku-4-5-202510014/5Nurses are legally responsible for patient monitoring in most jurisdictions; liability for missed complications (acute kidney injury, hyperkalemia, hypovolemic shock) falls on the institution and clinician, creating strong regulatory and medicolegal barriers to full automation without human sign-off.
Adoption barriersclaude-sonnet-54/5Critical care nursing is a licensed function with legal accountability for patient monitoring and intervention; regulations and hospital liability structures require a licensed nurse to be responsible for this surveillance and response.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI inference cost is low, but integration into ICU workflows, validation oversight, and liability exposure make the all-in cost comparable to or exceed the cost of a nurse performing this task, especially in high-acuity settings where errors carry severe consequences.
Cost vs. human wageclaude-sonnet-52/5Deploying integrated sensors, EHR analytics, and alert systems requires significant infrastructure and ongoing oversight, and does not replace the nurse's bedside monitoring and response, so cost savings versus a nurse's wage are limited.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed clinical system independently monitors fluid balance and detects electrolyte problems; existing EHR tools require manual data entry and human interpretation of trends, while diagnostic certainty demands validation against bloodwork and physical signs that only clinicians assess in real time.
Technical feasibility todayclaude-sonnet-52/5Clinical decision support and EWS tools exist in some ICUs that alert on fluid/electrolyte trends, but they are adjuncts to nursing surveillance rather than autonomous monitoring systems performing the task itself.

Compile and analyze data obtained from monitoring or diagnostic tests.

25

CI 25–25 · exposure 25 · augmentation 63 · importance 4.7/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Although health IT adoption is advancing, critical care remains highly conservative; most AI-assisted monitoring in ICUs is still pilot-stage, with clinicians requiring full manual review and validation, limiting actual displacement of nursing analysis work.
Sector adoption velocityclaude-sonnet-52/5Healthcare, especially acute/critical care, is a slower-adopting sector for full AI-driven decision-making due to regulatory, safety, and integration hurdles, though monitoring analytics pilots are increasing.
Augmentation potentialclaude-haiku-4-5-202510013/5AI dashboards and automated alerts assist nurses by surfacing trends and flagging abnormal values, reducing manual scan time; however, the core interpretive and decision-making work remains firmly human-driven, offering moderate productivity gain rather than transformation.
Augmentation potentialclaude-sonnet-54/5AI-based monitoring systems, predictive alerts, and data visualization tools meaningfully help nurses spot trends and prioritize attention, augmenting their analytic capacity substantially.
Task automatabilityclaude-haiku-4-5-202510012/5Current AI can extract and organize numeric data from monitoring systems and generate summaries, but clinical interpretation requires real-time integration of multiple data streams, patient context, and critical judgment that AI cannot reliably perform end-to-end without substantial human oversight, preventing the ≥50% time-saving threshold.
Task automatabilityclaude-sonnet-52/5AI can help aggregate and flag trends in monitoring data, but true clinical analysis integrating patient context, judgment, and real-time decision-making in critical care remains largely human-driven.direct end-to-end automation at equal quality is not yet achievable.
Adoption barriersclaude-haiku-4-5-202510014/5Liability law requires that a licensed nurse or physician sign off on clinical interpretations; regulatory bodies (FDA, hospital accreditation) mandate human accountability for patient safety decisions; and malpractice exposure creates strong legal barriers to full automation of diagnostic analysis in critical care.
Adoption barriersclaude-sonnet-54/5Clinical interpretation of diagnostic data in critical care is subject to licensure, liability, and hospital protocols requiring a qualified nurse or physician to interpret and act on results.
Cost vs. human wageclaude-haiku-4-5-202510012/5Integration of AI monitoring systems into ICU infrastructure, ongoing validation, and necessary human oversight mean total cost-per-analysis remains comparable to or higher than the labor of experienced critical care nurses who command high wages.
Cost vs. human wageclaude-sonnet-52/5Deploying and validating clinical AI analytics systems in critical care requires significant integration, compliance, and oversight costs that largely offset savings versus a nurse's judgment-based analysis.
Technical feasibility todayclaude-haiku-4-5-202510012/5While AI products exist for data aggregation and basic pattern flagging in ICU settings, they operate only within narrow scopes (e.g., single parameter trending) and require significant validation by clinicians; no mature system performs reliable end-to-end compilation and clinical analysis at scale in production without human intervention.
Technical feasibility todayclaude-sonnet-52/5Some clinical decision support and monitoring analytics tools exist in ICUs (e.g., early warning scores, sepsis alerts), but they are narrow, error-prone, and require human verification rather than fully reliable autonomous analysis.

Identify malfunctioning equipment or devices.

25

CI 25–25 · exposure 25 · augmentation 50 · importance 4.1/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Most hospitals and ICUs rely on manufacturers' built-in alarms and nurse vigilance rather than AI-augmented detection systems in production. Adoption remains in pilot phases; widespread deployment of AI-driven equipment monitoring in critical care is not yet standard practice.
Sector adoption velocityclaude-sonnet-52/5Healthcare, especially bedside critical care equipment monitoring, adopts AI more slowly than digital-native sectors due to safety validation and certification requirements.
Augmentation potentialclaude-haiku-4-5-202510013/5Automated alerts and dashboards that surface equipment telemetry and historical failure patterns can meaningfully assist nurses in prioritizing checks and catching subtle drift, but human judgment and physical inspection remain essential for definitive malfunction diagnosis.
Augmentation potentialclaude-sonnet-53/5AI-enabled alarms, predictive maintenance alerts, and anomaly detection in monitoring systems can help nurses flag potential equipment issues faster, though human judgment remains central.
Task automatabilityclaude-haiku-4-5-202510012/5AI could assist in monitoring sensor data and logs to flag equipment anomalies, but ICU equipment failures often require tactile inspection, contextual judgment about severity, and integration with real-time clinical data. Current systems lack the embodied and contextual reasoning needed for end-to-end autonomous detection meeting the 50% time-saving threshold.
Task automatabilityclaude-sonnet-52/5Detecting malfunctioning medical equipment relies on hands-on sensory checks, physical inspection, and integrating patient context that current AI cannot fully replicate end-to-end in real clinical settings.
Adoption barriersclaude-haiku-4-5-202510014/5Critical care equipment malfunctions directly impact patient safety and liability; regulatory bodies (FDA, hospital accreditation) mandate human verification and sign-off on equipment status. Licensing and patient-safety regulations create strong friction against fully autonomous automation without clinician oversight.
Adoption barriersclaude-sonnet-54/5Patient safety regulations, liability, and hospital protocols generally require a licensed clinician to confirm and respond to equipment malfunctions before altering patient care.
Cost vs. human wageclaude-haiku-4-5-202510012/5Equipment monitoring systems and AI-driven alerting still require substantial integration, calibration, and ongoing maintenance. The cost per malfunction detected often exceeds the value of a nurse's time spent on manual spot-checks, especially given the critical nature of false negatives.
Cost vs. human wageclaude-sonnet-52/5Embedded diagnostic sensors add cost to devices, and any AI-based monitoring still requires human verification and intervention, so savings versus a trained nurse's vigilance are limited.
Technical feasibility todayclaude-haiku-4-5-202510012/5While some hospital IT systems and equipment manufacturers offer basic monitoring alerts, most deployed solutions have significant gaps in reliability and false-positive rates. Production-grade autonomous malfunction detection across diverse ICU equipment is not yet mature in real clinical deployments.
Technical feasibility todayclaude-sonnet-52/5Some smart medical devices have built-in self-diagnostics and alarms, but no deployed AI product independently identifies malfunctioning equipment across the range of ICU devices reliably.

Identify patients at risk of complications due to nutritional status.

25

CI 25–25 · exposure 25 · augmentation 63 · importance 3.9/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Healthcare, while digitizing, remains slow to adopt AI-driven clinical decision-making, especially in acute care where liability concerns and regulatory uncertainty are high; most ICUs still rely on manual screening and clinician judgment rather than AI systems.
Sector adoption velocityclaude-sonnet-52/5Healthcare, especially critical care, is a slower-adopting sector for autonomous AI decision tools due to regulatory caution, though clinical decision-support pilots are increasing.
Augmentation potentialclaude-haiku-4-5-202510013/5AI-assisted dashboards that surface relevant nutritional labs, flag patients meeting NUTRIC criteria, or suggest intervention protocols could meaningfully support a nurse's assessment process, reducing cognitive load and missed cases while the nurse retains judgment.
Augmentation potentialclaude-sonnet-54/5AI-driven risk scoring and predictive analytics can meaningfully assist nurses by highlighting at-risk patients and relevant lab trends, improving triage efficiency while the nurse retains final judgment.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can analyze lab values and nutritional markers to flag risk patterns, the task requires synthesis of clinical context (medications, organ function, trajectory) and human judgment about individual patient presentation—factors that current systems handle inconsistently. Partial automation of data review is feasible, but end-to-end replacement with 50% time savings at equal quality is not yet demonstrated.
Task automatabilityclaude-sonnet-52/5AI can flag risk factors from structured EHR data, but integrating clinical judgment, physical assessment, and nuanced patient context limits full end-to-end automation today.
Adoption barriersclaude-haiku-4-5-202510014/5Regulatory and liability barriers are substantial: nurses are legally responsible for patient assessment, clinical decisions based on AI flags still require human sign-off, and errors in nutritional assessment can cause serious patient harm, creating asymmetric risk that deters full automation.
Adoption barriersclaude-sonnet-54/5Clinical risk identification in critical care is tightly bound to licensed nursing/physician judgment and liability, with regulatory and safety oversight limiting autonomous AI decision-making.
Cost vs. human wageclaude-haiku-4-5-202510012/5The loaded cost of a critical care nurse is very high (specialized labor, 24/7 availability), and current AI solutions for nutritional risk assessment require significant clinical validation, integration, and oversight labor—approaching parity rather than cost advantage.
Cost vs. human wageclaude-sonnet-52/5Deploying and validating clinical AI tools requires significant integration, oversight, and liability management, making costs comparable to or only modestly cheaper than nurse time for this specific subtask.
Technical feasibility todayclaude-haiku-4-5-202510012/5Clinical decision-support tools exist to flag nutritional risk using standardized screening tools (MUST, NUTRIC), but these are typically rule-based rather than AI-driven, and their integration into real ICU workflows remains incomplete. Standalone AI products that reliably perform this task in production ICU settings have not been widely deployed.
Technical feasibility todayclaude-sonnet-52/5Some clinical decision-support tools exist for nutritional risk scoring, but they are narrow, often rule-based, and not widely deployed as reliable standalone products in ICUs.

Evaluate patients' vital signs or laboratory data to determine emergency intervention needs.

23

CI 20–25 · exposure 30 · augmentation 75 · importance 5.0/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5While healthcare is digitizing, adoption of AI for autonomous critical-care decision-making remains slow and conservative. Most deployments are advisory (alerts to human nurses) rather than autonomous, and many hospitals remain in pilot phases. Deep production substitution is rare.
Sector adoption velocityclaude-sonnet-52/5Healthcare, especially acute/critical care, adopts AI decision-support tools cautiously and slowly due to regulatory scrutiny, liability concerns, and integration complexity with clinical workflows.
Augmentation potentialclaude-haiku-4-5-202510014/5AI-powered clinical decision support demonstrably assists nurses by surfacing anomalies, suggesting protocols, and aggregating data—raising their speed and confidence in triage. This is a core use case for medical AI, even if replacement remains limited.
Augmentation potentialclaude-sonnet-54/5AI-driven monitoring systems and predictive analytics (e.g., early warning scores, sepsis alerts) meaningfully help nurses prioritize attention and catch subtle deteriorations faster, augmenting their assessment without replacing their judgment.
Task automatabilityclaude-haiku-4-5-202510012/5AI can flag abnormal vital signs and lab values quickly, but evaluating *which* interventions are needed requires integrating multiple clinical contexts (medications, comorbidities, trends, patient history) and real-time judgment that current systems handle unreliably. End-to-end automation with 50% time savings at equal quality is not demonstrated in production.
Task automatabilityclaude-sonnet-52/5AI can flag abnormal vitals or labs and suggest urgency levels, but the actual clinical judgment integrating context, trends, and patient-specific factors to decide on emergency intervention remains a licensed nursing function that current systems cannot fully replace.
Adoption barriersclaude-haiku-4-5-202510015/5Nurses perform this task as a core clinical responsibility; malpractice liability, regulatory requirements (standard of care, scope of practice), and institutional credentialing rules legally and professionally mandate human clinical judgment. Automation of intervention decisions in critical care faces high regulatory and legal barriers.
Adoption barriersclaude-sonnet-55/5Emergency intervention decisions in critical care require licensed nursing/physician judgment and carry direct life-safety and liability implications, making regulatory and legal barriers to full automation very high.
Cost vs. human wageclaude-haiku-4-5-202510012/5Implementing clinical AI systems requires substantial infrastructure, model maintenance, hospital IT integration, and clinician oversight. The all-in cost per evaluation is comparable to or exceeds the cost of a nurse's time for the same decision, especially given required human validation.
Cost vs. human wageclaude-sonnet-52/5Deploying and maintaining monitoring/alert systems integrated with EHRs involves substantial infrastructure and oversight costs, and does not eliminate the need for a nurse's continuous presence and judgment, so savings versus a nurse's time are limited.
Technical feasibility todayclaude-haiku-4-5-202510013/5Clinical decision-support systems exist and hospitals deploy them (e.g., sepsis alerts, arrhythmia detection), but they still produce material false-positive and false-negative rates. Clinicians must verify and override frequently, so the AI performs a screening or alerting role, not independent reliable evaluation.
Technical feasibility todayclaude-sonnet-52/5Clinical decision support and early-warning score systems (e.g., sepsis alerts, deterioration indices) are deployed in some hospitals but have known false-positive/negative rates and are used only as adjuncts, not as autonomous decision-makers.

Participate in the development, review, or evaluation of nursing practice protocols.

21

CI 16–25 · exposure 17 · augmentation 63 · importance 3.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Healthcare adoption of AI in protocol development is slow and cautious; institutions favor evidence-based but human-led processes. Pilots of AI-assisted evidence synthesis exist, but production-scale replacement or autonomous protocol design remains rare in critical care settings.
Sector adoption velocityclaude-sonnet-52/5Healthcare, especially clinical protocol development, adopts AI cautiously due to regulatory and safety concerns, with pilots more common than production use in this specific task.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can usefully assist by synthesizing literature, flagging gaps, and generating initial drafts of protocol sections, which nurses then review, modify, and validate. This augmentation improves productivity on evidence-gathering and documentation phases, though the strategic and clinical judgment work remains human-driven.
Augmentation potentialclaude-sonnet-54/5AI can meaningfully assist by synthesizing literature, benchmarking against best practices, and drafting protocol language for nurse review, improving efficiency while humans retain final judgment.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires deep clinical judgment, institutional knowledge, and collaborative decision-making with multidisciplinary teams to shape evidence-based practice standards. Current AI systems lack the autonomy and accountability to author or meaningfully drive protocol development end-to-end.
Task automatabilityclaude-sonnet-52/5AI can draft or summarize protocol language and pull relevant evidence, but the actual clinical judgment, consensus-building, and evaluation against unit-specific practice require human expertise and cannot be fully automated today.
Adoption barriersclaude-haiku-4-5-202510014/5Nursing practice protocols must be developed by licensed clinical professionals with accountability for patient safety; regulatory bodies and accreditation standards (e.g., The Joint Commission) typically require human clinical leadership and sign-off. Liability and quality assurance create strong requirements for human clinician ownership.
Adoption barriersclaude-sonnet-54/5Nursing practice protocols typically require sign-off by licensed nursing leadership and compliance with institutional/regulatory standards, creating strong authorization and liability barriers.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI tools for evidence synthesis and drafting support have modest costs, but this task demands senior clinical expertise and institutional coordination that remain expensive. The all-in cost of AI assistance plus required human oversight does not achieve significant cost advantage over experienced nurses leading the process.
Cost vs. human wageclaude-sonnet-52/5AI could cut research/drafting time, but human nurse experts must still review, validate, and approve protocols, so overall cost savings are modest relative to the specialized labor involved.
Technical feasibility todayclaude-haiku-4-5-202510012/5While AI can assist with literature searches and summarize evidence, no deployed system reliably performs independent protocol development or evaluation at scale. Products exist for evidence synthesis but require substantial human oversight and clinical judgment to translate into institutional standards.
Technical feasibility todayclaude-sonnet-52/5Some clinical decision-support and literature-synthesis tools exist, but no deployed product reliably drafts or evaluates nursing protocols in production without heavy clinician oversight.

Assess patients' pain levels or sedation requirements.

14

CI 3–25 · exposure 13 · augmentation 50 · importance 4.6/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Healthcare sectors remain slow adopters of autonomous AI decision-making, particularly in ICUs where liability and patient safety concerns are highest; current adoption is limited to assistive tools (vital-sign dashboards) rather than replacement of the assessment function itself.
Sector adoption velocityclaude-sonnet-52/5Healthcare, especially bedside critical care nursing, has been slow to adopt autonomous AI systems for direct patient assessment despite broader digitization in hospitals.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can meaningfully assist critical care nurses by highlighting vital-sign trends, flagging abnormalities, or summarizing behavioral observations, thereby speeding data review and supporting human clinical judgment without replacing it.
Augmentation potentialclaude-sonnet-53/5AI-enabled monitoring tools (e.g., sedation scoring algorithms, vital sign pattern analysis) can support nurses' assessments, but the core judgment remains human-driven.
Task automatabilityclaude-haiku-4-5-202510012/5AI cannot reliably assess pain or sedation without direct multimodal patient interaction (facial expression, vital signs, verbal communication, behavioral cues). While AI can assist in analyzing vital signs or interpreting patient responses, the full task requires clinical judgment and real-time patient presence that current systems cannot perform autonomously at equal quality.
Task automatabilityclaude-sonnet-51/5Assessing pain or sedation requires hands-on physical exam, patient interaction, and interpretation of nonverbal cues in real time, none of which current AI can perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Strong legal and regulatory barriers exist: a licensed critical care nurse must independently assess pain and sedation per clinical standards and liability frameworks; medication dosing (especially sedatives) depends on validated human clinical judgment; malpractice exposure makes substitution of AI judgment legally and ethically risky.
Adoption barriersclaude-sonnet-55/5This is a licensed nursing responsibility tied to patient safety and legal scope-of-practice requirements, with no allowance for full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI inference for vital-sign analysis is cheap, but the end-to-end task requires integration with clinical workflows, real-time monitoring infrastructure, and human validation, making total cost comparable to or exceeding the loaded wage of a critical care nurse for this assessment function.
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute performing this task, so cost comparison favors the human nurse who must be present regardless.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed AI product reliably performs independent pain/sedation assessment in production ICU settings. Research systems exist for vital-sign analysis and facial coding, but they achieve only partial accuracy and all clinical deployments require substantial human oversight and validation.
Technical feasibility todayclaude-sonnet-51/5No deployed product independently assesses ICU patients' pain or sedation levels; this remains a bedside clinical judgment task performed by nurses.

Assess patients' psychosocial status and needs, including areas such as sleep patterns, anxiety, grief, anger, and support systems.

14

CI 3–25 · exposure 13 · augmentation 50 · importance 3.9/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Healthcare sectors are slow to adopt AI for subjective clinical judgment; psychosocial assessment remains one of the least automated nursing competencies in production systems, with most implementations still in pilot or documentation-support phases.
Sector adoption velocityclaude-sonnet-52/5Healthcare, especially bedside critical care, has slow AI adoption for direct patient interaction tasks despite faster adoption in administrative or diagnostic support areas.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can usefully assist by flagging documented symptom patterns, suggesting assessment frameworks, or auto-generating draft notes, but the human nurse must conduct the actual interaction and validate findings, so augmentation is moderate rather than transformative.
Augmentation potentialclaude-sonnet-53/5AI can help by flagging risk factors from EHR data, suggesting standardized psychosocial screening questions, or summarizing patient history, aiding but not replacing the nurse's assessment.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can extract clinical data and flag patterns in documented symptoms, the core task requires real-time observation, rapport-building, and nuanced interpretation of non-verbal cues and emotional context that current systems cannot reliably perform end-to-end at equal quality to a trained nurse.
Task automatabilityclaude-sonnet-51/5Direct psychosocial assessment requires in-person observation, rapport-building, and clinical judgment about a critically ill patient's emotional state, which current AI cannot perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Psychosocial assessment is embedded in licensure requirements for nurses, involves significant liability around missed psychological risk (suicide, trauma), and relies on therapeutic relationship and direct patient contact that regulatory frameworks expect to be human-delivered.
Adoption barriersclaude-sonnet-55/5Licensed nursing judgment, patient safety, and legal/clinical accountability strictly require a human professional to perform this assessment in critical care settings.
Cost vs. human wageclaude-haiku-4-5-202510012/5Integration costs for AI-assisted psychosocial screening remain significant relative to the labor savings, and oversight by a nurse is still required, making the all-in cost comparable to or higher than direct human assessment.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing this task, so cost comparison favors the human nurse entirely for now.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed clinical product reliably performs comprehensive psychosocial assessment independently; AI can assist with symptom screening or documentation review, but real-world deployment still requires substantial human judgment and patient interaction to validate findings.
Technical feasibility todayclaude-sonnet-51/5No deployed products conduct autonomous psychosocial assessment of ICU patients; at best, AI tools support documentation or screening but do not perform the assessment itself.

Set up and monitor medical equipment and devices such as cardiac monitors, mechanical ventilators and alarms, oxygen delivery devices, transducers, or pressure lines.

13

CI 5–21 · exposure 17 · augmentation 63 · importance 4.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Hospitals have adopted isolated monitoring dashboards and alert systems, but end-to-end automation of device setup and configuration remains rare in production. Adoption is cautious due to regulatory, liability, and safety-critical nature of the task; pilot projects exist but widespread deployment is limited.
Sector adoption velocityclaude-sonnet-51/5Bedside critical care nursing is a highly physical, low-digitization environment where equipment setup remains entirely manual; adoption of automation for this specific task is negligible.
Augmentation potentialclaude-haiku-4-5-202510014/5AI significantly augments nurses through real-time alerting systems, automated trend analysis, predictive alarms, and integrated monitoring dashboards that reduce cognitive load and allow faster response to device changes. Nurses remain central to setup and clinical decisions, but AI-assisted monitoring substantially raises productivity and situational awareness.
Augmentation potentialclaude-sonnet-53/5AI-enabled monitors can provide smarter alerting, trend analysis, and predictive alarms that help nurses prioritize attention, offering moderate assistance even though the physical setup remains manual.
Task automatabilityclaude-haiku-4-5-202510012/5While AI systems can assist with monitoring and alerting on equipment parameters through sensor data, the physical setup of medical devices (inserting lines, securing monitors, troubleshooting connections) requires human dexterity and contextual judgment. Current AI lacks the embodied capability and real-time clinical reasoning to set up equipment end-to-end or achieve 50% time savings at equal safety.
Task automatabilityclaude-sonnet-51/5This requires hands-on physical setup of invasive lines, calibration of devices on a patient's body, and continuous clinical judgment about changing physiological status—no AI system can physically place or adjust these devices.
Adoption barriersclaude-haiku-4-5-202510014/5Strong barriers exist: federal regulations (FDA, Joint Commission, CMS) require licensed nursing personnel to set up and bear responsibility for critical medical equipment; patient safety liability and malpractice risk are asymmetric if automation fails; hospitals have significant regulatory and institutional requirements for human sign-off on device configuration.
Adoption barriersclaude-sonnet-55/5Physical patient contact, invasive device placement, and licensure requirements mean only credentialed critical care nurses can legally perform this task, creating hard regulatory and safety barriers.
Cost vs. human wageclaude-haiku-4-5-202510011/5Current AI-based monitoring and alerting systems still require substantial infrastructure, integration, and clinical oversight, making all-in costs comparable to or higher than a nurse's labor for this task. The physical setup component necessitates human presence regardless.
Cost vs. human wageclaude-sonnet-51/5AI cannot substitute for the physical labor and hands-on skill involved, so there is no viable AI cost comparison for the core task; a human must perform it regardless of cost.
Technical feasibility todayclaude-haiku-4-5-202510012/5Monitoring components are partially deployed in hospital systems (automated alerts, dashboards), but fully autonomous setup and troubleshooting of complex medical equipment is not reliably performed by any production system. Clinical decision-support tools exist but require human operators; no mature AI product independently manages this task.
Technical feasibility todayclaude-sonnet-52/5Smart monitors and alarm systems exist that display and flag data, but no deployed product autonomously sets up, calibrates, or manages ventilators, transducers, or pressure lines without a nurse physically performing the task.

Conduct pulmonary assessments to identify abnormal respiratory patterns or breathing sounds that indicate problems.

13

CI 0–25 · exposure 13 · augmentation 50 · importance 4.7/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Despite digital health adoption in hospitals, automated pulmonary assessment tools are not in meaningful production use in ICU workflows. Adoption remains in pilot phase or research settings; clinical inertia and liability concerns slow real-world deployment.
Sector adoption velocityclaude-sonnet-52/5Healthcare bedside clinical tasks generally show slow AI adoption due to physical, regulatory, and safety constraints, despite faster uptake in administrative or diagnostic-support areas.
Augmentation potentialclaude-haiku-4-5-202510013/5AI-assisted analysis of recorded breath sounds or continuous respiratory monitoring data could help nurses flag subtle patterns or provide differential suggestions, improving sensitivity. However, the nurse must remain the primary decision-maker, and augmentation is limited to support rather than transformation of the core assessment.
Augmentation potentialclaude-sonnet-53/5AI-enabled digital stethoscopes and decision-support tools can help flag abnormal sounds or patterns, assisting the nurse's interpretation, though the core physical assessment remains human-led.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can analyze digital recordings of breath sounds and respiratory patterns from monitoring data, conducting a pulmonary assessment requires real-time auscultation, physical examination, and contextual integration that current systems cannot reliably perform end-to-end. AI falls well short of the 50% time-saving-at-equal-quality threshold for the full task.
Task automatabilityclaude-sonnet-51/5This requires hands-on physical examination, auscultation with a stethoscope, and real-time clinical judgment integrated with patient context; no AI system can perform the physical assessment itself.
Adoption barriersclaude-haiku-4-5-202510014/5Legal and regulatory barriers are substantial: assessment and diagnosis of respiratory abnormalities are within the scope of licensed nursing practice, and liability for missed or misidentified pathology rests with the clinician. ICU settings also require human judgment and physical presence for patient safety.
Adoption barriersclaude-sonnet-55/5Direct physical patient assessment in critical care requires licensed nursing presence and clinical judgment, with significant liability and regulatory requirements mandating human hands-on care.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI-based auscultation analysis, where it exists, requires specialized hardware, integration into clinical workflows, and human oversight. The total cost per assessment is likely comparable to or higher than the incremental cost of a nurse performing the task, given current deployment maturity.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing the full physical assessment, so cost comparison favors the human nurse who must be present regardless.
Technical feasibility todayclaude-haiku-4-5-202510012/5Research prototypes exist for algorithmic analysis of lung sounds from recordings, but no deployed clinical product reliably performs a complete pulmonary assessment in production ICU settings. Error rates remain material, and systems lack the ability to integrate findings with vital signs, patient history, and exam context as a licensed practitioner must do.
Technical feasibility todayclaude-sonnet-51/5No deployed product independently conducts pulmonary physical assessments in critical care; digital stethoscope AI aids exist only as research-stage or narrow adjuncts, not autonomous assessment tools.

Monitor patients for changes in status and indications of conditions such as sepsis or shock and institute appropriate interventions.

7

CI 7–7 · exposure 9 · augmentation 63 · importance 4.8/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5ICUs are digitized and early adopters of monitoring technology, but actual AI-driven intervention automation remains rare; most deployments are pilot decision-support aids rather than autonomous action, reflecting both technical immaturity and regulatory/liability caution.
Sector adoption velocityclaude-sonnet-52/5Hospitals are adopting AI-based early warning systems and sepsis alerts gradually, but ICU care remains a high-touch, physically-mediated environment with slow, cautious rollout of automation.
Augmentation potentialclaude-haiku-4-5-202510013/5AI-assisted monitoring (real-time vital-sign dashboards, predictive alerts for deterioration) can help nurses prioritize which patients to check, reducing cognitive load. However, the scope is limited because human judgment, physical assessment, and clinical reasoning remain central to the task.
Augmentation potentialclaude-sonnet-54/5AI-driven predictive analytics and sepsis/shock alert algorithms meaningfully augment nurses by flagging deteriorating patients earlier and prioritizing attention, while the nurse remains responsible for interpretation and action.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires continuous real-time judgment integrating multiple vital signs, clinical context, and rapid decision-making about life-threatening conditions. While AI can flag abnormal values, current systems cannot reliably synthesize the full clinical picture or make the nuanced treatment decisions that sepsis/shock require, and they cannot physically institute interventions.
Task automatabilityclaude-sonnet-51/5This task requires continuous bedside physical assessment, hands-on interventions, and real-time clinical judgment integrating multiple data streams that current AI cannot execute end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Critical care nursing is a licensed clinical role; interventions for sepsis and shock (medications, fluids, escalation) are legally restricted to licensed providers and often require physician orders. Liability and regulatory requirements (standards of care, Joint Commission) make autonomous intervention infeasible.
Adoption barriersclaude-sonnet-55/5Nursing licensure, scope-of-practice law, and patient safety liability require a human nurse to perform physical assessment and administer interventions.
Cost vs. human wageclaude-haiku-4-5-202510011/5AI monitoring systems still require substantial infrastructure, integration, validation, and clinician oversight; the all-in cost remains comparable to or higher than the marginal cost of a nurse's time spent on passive vital-sign observation.
Cost vs. human wageclaude-sonnet-51/5A licensed critical care nurse must be physically present to intervene; AI monitoring tools add cost on top of, rather than replacing, nursing staff.
Technical feasibility todayclaude-haiku-4-5-202510012/5Some monitoring dashboards and alerting systems use AI to flag vital-sign anomalies in production ICUs, but they function as narrow alerts requiring clinician interpretation, not autonomous diagnosis or intervention. No deployed system reliably detects sepsis or shock onset and recommends interventions independently at clinical-grade accuracy.
Technical feasibility todayclaude-sonnet-52/5Some deployed clinical decision support and early-warning-score systems (e.g., sepsis alert algorithms) flag risk in production, but they only support monitoring rather than performing patient assessment or physically instituting interventions.

Ensure that equipment or devices are properly stored after use.

6

CI 0–13 · exposure 0 · augmentation 25 · importance 3.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Healthcare remains a laggard sector for physical automation tasks, with very limited deployment of robots in actual ICU environments due to contamination risks, regulatory burden, and organizational resistance to replacing hands-on clinical staff.
Sector adoption velocityclaude-sonnet-52/5Healthcare physical/logistics tasks in critical care settings see slow robotic/AI adoption compared to administrative or diagnostic AI uses in the same sector.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could theoretically assist via equipment location tracking or inventory management systems, but most of this task is physical labor requiring human judgment and accountability, limiting meaningful augmentation potential.
Augmentation potentialclaude-sonnet-52/5AI could assist with inventory tracking, checklists, or reminders about storage protocols, but offers minimal help with the core physical act of storing equipment.
Task automatabilityclaude-haiku-4-5-202510011/5Equipment storage after clinical use requires physical manipulation in a complex medical environment with strict protocols, inventory tracking, and context-specific decisions about sterilization and placement that current AI cannot perform end-to-end without human intervention.
Task automatabilityclaude-sonnet-51/5This is a physical task requiring manipulation, cleaning verification, and storage of medical equipment in a physical hospital environment, which current AI cannot perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Critical care equipment handling is tightly regulated by hospital protocols, medical device manufacturer guidelines, and healthcare compliance standards; human nurses are required by practice standards and liability frameworks to ensure proper sterilization and secure storage.
Adoption barriersclaude-sonnet-53/5While not licensure-specific, infection control protocols, equipment safety standards, and accountability for proper storage create institutional and regulatory friction against non-human handling.
Cost vs. human wageclaude-haiku-4-5-202510011/5The cost of robotics capable of safely handling sensitive medical equipment in a sterile ICU environment vastly exceeds the hourly wage of nursing staff, making automation economically unfeasible at scale.
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute performing this physical task, so cost comparison favors the human by default since AI cannot execute it.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI system can reliably navigate a critical care unit to physically locate, retrieve, identify, and properly store diverse medical equipment with the safety and regulatory compliance required in production environments.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product physically stores or handles medical equipment; this remains a manual nursing/technician responsibility.

Assess family adaptation levels and coping skills to determine whether intervention is needed.

6

CI 0–11 · exposure 0 · augmentation 38 · importance 4.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Critical care nursing remains in high-touch, heavily regulated settings with strong professional oversight. Family assessment is relational and context-dependent; adoption of full automation is minimal, and organizations strongly prefer human nurses for these functions due to liability and care quality concerns.
Sector adoption velocityclaude-sonnet-52/5Healthcare, especially bedside critical care nursing, has been slow to adopt AI for interpersonal and psychosocial assessment tasks compared to administrative or diagnostic support areas.
Augmentation potentialclaude-haiku-4-5-202510013/5AI could assist nurses by flagging risk indicators in patient/family documentation, suggesting evidence-based intervention frameworks, or prompting structured assessment questions; however, the nurse must still perform the core relational and clinical work of understanding family dynamics and making clinical judgment about need for intervention.
Augmentation potentialclaude-sonnet-52/5AI could help nurses by summarizing patient/family history, flagging risk factors from documentation, or suggesting coping-assessment frameworks, but it plays a minor supportive role in the actual interpersonal assessment.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires nuanced clinical judgment about family psychological and emotional states, adaptive responses, and readiness for intervention—domains where current AI cannot reliably detect the subtle behavioral, verbal, and contextual cues that determine coping adequacy. No current AI system can conduct the holistic psychosocial assessment needed to replace a trained nurse's evaluation.
Task automatabilityclaude-sonnet-51/5This requires in-person emotional and psychosocial assessment via observation, tone, rapport, and clinical judgment in a high-stakes hospital setting—far beyond what current AI can perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Assessing family adaptation is a core clinical nursing function; most jurisdictions require a licensed nurse to conduct psychosocial assessments, document findings, and make intervention decisions. Liability for missed family crisis, regulatory standards for nursing practice, and the requirement for a licensed professional to provide care coordination create hard legal and professional barriers.
Adoption barriersclaude-sonnet-54/5Nursing licensure, clinical accountability, and the inherently relational/human-contact nature of family psychosocial assessment create strong professional and liability barriers to automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5An AI system would require significant setup (integration with EHR, custom training on institutional protocols, ongoing oversight) and would still need a nurse to validate findings, making the total cost comparable to or higher than the nurse's direct time on the task.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing this task, so cost comparison favors the human nurse entirely; any AI attempt would require extensive human oversight negating savings.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product reliably performs comprehensive family coping and adaptation assessment in clinical settings. While chatbots might screen for general stress, they cannot replicate the observational, relational, and clinical judgment skills required to assess family dynamics and determine intervention necessity in a critical care context.
Technical feasibility todayclaude-sonnet-51/5No deployed clinical product autonomously assesses family coping and adaptation to determine intervention needs; this remains a human nursing judgment task.

Collaborate with other health care professionals to develop and revise treatment plans, based on identified needs and assessment data.

5

CI 3–7 · exposure 5 · augmentation 63 · importance 4.6/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Healthcare adoption of clinical decision-support AI remains cautious and incremental; regulatory uncertainty, liability concerns, and clinician skepticism slow deployment. AI-assisted documentation and data triage are common, but autonomous or semi-autonomous plan generation is still rare in production critical care settings.
Sector adoption velocityclaude-sonnet-52/5Healthcare, especially acute/critical care clinical decision-making, has been slower than sectors like finance or professional services to adopt AI for core judgment tasks, though administrative AI use is growing.
Augmentation potentialclaude-haiku-4-5-202510014/5AI excels at rapidly synthesizing patient data, suggesting evidence-based interventions, flagging clinical alerts, and drafting plan language—capabilities that can substantially accelerate a nurse's ability to formulate comprehensive plans while the clinician retains judgment and final authority. This is a natural assistive fit.
Augmentation potentialclaude-sonnet-53/5AI can help synthesize patient data, flag risks, and support literature review or documentation, aiding clinicians during care planning discussions, though it doesn't replace the collaborative decision process itself.
Task automatabilityclaude-haiku-4-5-202510011/5Developing and revising treatment plans requires nuanced clinical judgment, integration of patient-specific medical history, multidisciplinary coordination, and real-time responsiveness to changing patient conditions—activities that demand human expertise and accountability. Current AI cannot independently perform this end-to-end task at the 50% time-saving threshold while maintaining clinical safety standards.
Task automatabilityclaude-sonnet-51/5This requires real-time clinical judgment, physical patient assessment, interpersonal collaboration, and accountability that current AI cannot perform end-to-end; at best it can assist with summarization or data aggregation.
Adoption barriersclaude-haiku-4-5-202510015/5Treatment plan development is legally and ethically anchored to licensed healthcare professionals (nurses, physicians) who bear liability for care decisions; regulatory bodies (state nursing boards, CMS, accrediting bodies) and malpractice frameworks require human clinical judgment and accountability. Substitution is not legally permissible.
Adoption barriersclaude-sonnet-55/5Treatment planning in critical care legally and ethically requires licensed clinicians (nurses, physicians) to exercise judgment and sign off, with strong liability and regulatory constraints preventing AI substitution.
Cost vs. human wageclaude-haiku-4-5-202510011/5The oversight, validation, and liability costs of AI-assisted plan development, combined with the need for human clinician review and modification, make the total cost comparable to or higher than direct human planning. The value lies in augmentation, not cost reduction.
Cost vs. human wageclaude-sonnet-51/5Given the high liability, need for licensed oversight, and limited AI capability in this domain, AI cannot substitute for the human role, making cost comparison favor human labor despite AI's low per-query cost.
Technical feasibility todayclaude-haiku-4-5-202510012/5While AI can assist with data synthesis and draft documentation, no deployed clinical product reliably generates treatment plans independently; healthcare systems use AI for support (risk flagging, evidence retrieval) but humans retain full decision authority and legal responsibility. Production systems remain in a supportive, not autonomous, role.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously collaborates with clinical teams to develop or revise critical care treatment plans; decision-support tools exist but are advisory only and narrowly scoped.

Prioritize nursing care for assigned critically ill patients, based on assessment data or identified needs.

4

CI 0–7 · exposure 5 · augmentation 50 · importance 4.7/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Healthcare, particularly critical care, remains a laggard sector for autonomous AI automation due to regulatory constraints, malpractice risk, and the requirement for human licensure and accountability in life-or-death decisions. Adoption is limited to supportive tools, not replacement.
Sector adoption velocityclaude-sonnet-52/5Healthcare, especially acute/critical care nursing, has historically slow AI adoption for direct patient care decisions due to regulatory, safety, and workflow integration constraints, though monitoring/alert tools are spreading.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can meaningfully assist nurses by aggregating and highlighting vital sign trends, flagging potential deterioration, and surfacing relevant patient data, thereby supporting the nurse's prioritization process. However, the nurse remains the decision-maker and retains responsibility for final prioritization.
Augmentation potentialclaude-sonnet-53/5AI-driven early warning scores, predictive analytics, and monitoring alerts can help nurses flag deteriorating patients and support prioritization, meaningfully aiding judgment without replacing it.
Task automatabilityclaude-haiku-4-5-202510011/5Prioritizing nursing care for critically ill patients requires real-time clinical judgment that integrates complex physiological data, patient history, and evolving conditions. Current AI systems cannot reliably perform this integrative assessment and prioritization end-to-end with the nuance and accountability required in critical care.
Task automatabilityclaude-sonnet-51/5This requires real-time bedside clinical judgment integrating physiological monitoring, physical assessment, and rapid triage of critically ill patients—no AI system performs this end-to-end task autonomously today.
Adoption barriersclaude-haiku-4-5-202510015/5Legal, regulatory, and ethical barriers are substantial: critical care prioritization decisions must be made by licensed nurses who are held accountable for patient outcomes. Medical liability, regulatory requirements under nursing practice acts, and the legal necessity of human clinical judgment create hard barriers to automation.
Adoption barriersclaude-sonnet-55/5Nursing licensure, scope-of-practice laws, and direct patient-safety liability require a licensed critical care nurse to make and be accountable for these prioritization decisions.
Cost vs. human wageclaude-haiku-4-5-202510011/5The cost of developing, validating, and maintaining AI systems for clinical prioritization in ICU settings, combined with required human oversight and liability costs, far exceeds the cost of a trained critical care nurse performing this judgment directly.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing this task alone, so cost comparison favors the human nurse who must be present regardless; any AI tool would be additive cost, not a replacement.
Technical feasibility todayclaude-haiku-4-5-202510012/5While AI tools can assist with data analysis and flagging abnormalities, no deployed product reliably prioritizes care for critically ill patients independently. Existing clinical decision support systems are narrow, require human validation, and do not operate at the full scope and reliability needed for this task in production critical care settings.
Technical feasibility todayclaude-sonnet-51/5No deployed product independently prioritizes and executes nursing care decisions for ICU patients; clinical decision-support tools exist but remain advisory, not autonomous prioritization systems.

Administer blood and blood products, monitoring patients for signs and symptoms related to transfusion reactions.

4

CI 0–7 · exposure 5 · augmentation 38 · importance 4.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Healthcare adoption of AI for autonomous clinical procedures remains minimal, with strong regulatory and professional barriers. Hospitals continue to rely on human nursing for direct patient care tasks, particularly those involving medication and blood product administration.
Sector adoption velocityclaude-sonnet-52/5Critical care nursing is a highly physical, regulated healthcare setting where AI adoption for hands-on clinical procedures remains slow and limited to decision-support tools.
Augmentation potentialclaude-haiku-4-5-202510012/5AI can assist by providing alerting on vital sign trends or flagging pre-transfusion lab results, but it offers limited augmentation for the core competency of real-time bedside monitoring and rapid recognition of transfusion reaction symptoms. The nurse's judgment and sensory assessment remain irreplaceable.
Augmentation potentialclaude-sonnet-53/5AI-enabled monitoring systems and EHR alerts can help flag early signs of transfusion reactions or track vitals, assisting nurses without replacing direct observation and intervention.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires continuous bedside patient monitoring, physical IV administration, and real-time clinical judgment to detect subtle signs of transfusion reactions. Current AI systems cannot physically administer blood products or perform the nuanced, moment-to-moment assessment of patient status that defines this critical care function.
Task automatabilityclaude-sonnet-51/5Physical administration of blood products and hands-on monitoring for transfusion reactions requires direct patient contact, physical assessment, and rapid clinical judgment that current AI cannot perform end-to-end.'
Adoption barriersclaude-haiku-4-5-202510015/5Blood product administration and transfusion reaction monitoring are legally and clinically mandated to be performed or directly supervised by a licensed nurse. Regulatory requirements, liability frameworks, and clinical standards explicitly require human licensure and accountability for this high-risk procedure.
Adoption barriersclaude-sonnet-55/5Administering blood products is a licensed nursing scope-of-practice task with strict clinical protocols, legal liability, and mandatory human oversight for patient safety.
Cost vs. human wageclaude-haiku-4-5-202510011/5The cost of AI systems, oversight infrastructure, and liability management to attempt this task would far exceed the wage of a nurse, who performs it as part of broader ICU duties. The liability exposure for transfusion errors alone makes automation economically unfeasible.
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute performing this physical task, so no cost comparison favors AI; the human nurse remains the only viable option.
Technical feasibility todayclaude-haiku-4-5-202510012/5While AI can assist in flagging laboratory abnormalities or documented vital signs, no deployed product reliably performs the core task of administering blood products or conducting the continuous, multisensory monitoring required to detect transfusion reactions in real time. Existing systems lack the embodied capability and clinical validation for this function.
Technical feasibility todayclaude-sonnet-51/5No deployed product administers blood transfusions or performs the physical monitoring required; this remains entirely a hands-on nursing task.

Identify patients' age-specific needs and alter care plans as necessary to meet those needs.

4

CI 0–7 · exposure 5 · augmentation 50 · importance 3.8/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Critical care remains a high-touch, heavily regulated domain with strong cultural emphasis on human clinical expertise and legal accountability. Adoption of autonomous AI for core care planning is minimal; AI is used only as a supporting tool under human direction.
Sector adoption velocityclaude-sonnet-52/5Healthcare, especially bedside critical care, adopts AI slowly due to safety, regulatory, and workflow integration constraints compared to information-sector adoption rates.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can usefully assist nurses by flagging age-specific clinical parameters, suggesting guideline-aligned considerations, and surfacing relevant patient data, thereby supporting faster and more comprehensive care planning decisions while the nurse retains full authority.
Augmentation potentialclaude-sonnet-53/5AI decision-support tools and EHR-integrated alerts can help flag age-specific risk factors or protocols, aiding nurses' judgment without replacing their assessment.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires nuanced clinical judgment integrating patient history, developmental stage, comorbidities, and complex care protocols—decisions that demand human expertise. Current AI systems cannot reliably assess the full spectrum of age-specific physiological and psychological needs or independently modify critical care plans without human oversight.
Task automatabilityclaude-sonnet-51/5This requires direct clinical assessment, physical examination, and real-time judgment about a specific patient's condition that AI cannot independently perform end-to-end today.entiment
Adoption barriersclaude-haiku-4-5-202510015/5Strong regulatory and legal barriers protect this task: licensed nurses are required by law to assess and modify care plans, and malpractice liability is high. Medical boards and healthcare law require a qualified human to take responsibility for care plan decisions in critical care.
Adoption barriersclaude-sonnet-55/5Clinical care planning and patient assessment in critical care require licensed nursing judgment and carry high liability, with strict regulatory and scope-of-practice requirements.
Cost vs. human wageclaude-haiku-4-5-202510011/5The computational and integration costs of AI systems to support this task, combined with necessary nursing oversight and validation, exceed the cost-savings over a critical care nurse's judgment, especially given liability and error tolerances.
Cost vs. human wageclaude-sonnet-51/5AI cannot substitute for the licensed nurse performing this task, so there is no viable cost comparison—human labor remains mandatory.
Technical feasibility todayclaude-haiku-4-5-202510012/5While AI can flag age-related clinical parameters and suggest protocol adjustments based on guidelines, no deployed product reliably performs the full task of independently identifying all relevant age-specific needs and altering care plans in a critical care setting. Products exist for clinical decision support but not autonomous plan modification.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously identifies age-specific patient needs and modifies bedside critical care plans; this remains firmly in the clinician's domain.

Provide post-mortem care.

3

CI 0–5 · exposure 0 · augmentation 13 · importance 3.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Healthcare sectors have no incentive to automate this task because it remains intrinsically tied to nursing licensure, dignity protocols, and human-contact requirements; adoption of AI here is not occurring.
Sector adoption velocityclaude-sonnet-51/5Healthcare's hands-on bedside and end-of-life care tasks are among the least digitized and least likely to see AI/robotic adoption in the near term.
Augmentation potentialclaude-haiku-4-5-202510012/5AI might assist with documentation or family communication templates, but the core caregiving and physical tasks remain entirely human-dependent; assistance potential is marginal.
Augmentation potentialclaude-sonnet-51/5AI offers essentially no meaningful assistance for the physical and emotional labor of preparing a deceased patient's body and supporting grieving families.
Task automatabilityclaude-haiku-4-5-202510011/5Post-mortem care involves direct physical contact with deceased bodies (positioning, cleaning, dressing), cultural/religious protocols, documentation, and family coordination—tasks requiring human embodiment, tactile judgment, and compassionate presence that AI systems cannot provide.
Task automatabilityclaude-sonnet-51/5Post-mortem care involves physical handling of a deceased patient's body, cleaning, preparing for family viewing, and emotionally sensitive interaction with family—none of which current AI systems can perform physically or emotionally.
Adoption barriersclaude-haiku-4-5-202510015/5Post-mortem care is legally and professionally bound to licensed nurses or designated healthcare staff; cultural sensitivity, legal chain-of-custody for the body, and family presence requirements create hard legal and ethical barriers to any substitution.
Adoption barriersclaude-sonnet-54/5While not strictly licensed in the way a diagnosis is, post-mortem care carries strong cultural, ethical, and organizational expectations that a human nurse perform it with dignity and compassion, effectively barring automation.
Cost vs. human wageclaude-haiku-4-5-202510011/5This task has no meaningful cost comparison since AI cannot perform the core work; a human nurse must be present regardless.
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute performing this physical task, so cost comparison favors the human by default since AI cannot execute it at all.
Technical feasibility todayclaude-haiku-4-5-202510011/5No AI system can perform physical post-mortem care tasks; this is a domain where human nurses remain fully responsible and no deployed products attempt automation.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product performs physical bodily care tasks; this remains entirely in the domain of physical human labor and compassionate presence.

Perform approved therapeutic or diagnostic procedures, based upon patients' clinical status.

1

CI 0–3 · exposure 0 · augmentation 50 · importance 4.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Healthcare remains a laggard sector for task automation due to regulatory constraints, patient safety liability, and the irreplaceable need for licensed human practitioners to perform and sign off on therapeutic procedures.
Sector adoption velocityclaude-sonnet-52/5Healthcare bedside/physical care lags in AI adoption due to safety-critical, hands-on nature, though administrative and diagnostic support tools are spreading faster elsewhere in the field.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist via clinical decision support, real-time monitoring alerts, and procedure protocol recommendations, which may improve the nurse's efficiency and safety margin; however, the nurse remains the active executor of the procedure.
Augmentation potentialclaude-sonnet-53/5AI can assist with clinical decision support, monitoring alerts, and documentation around the procedure, improving situational awareness even though it cannot perform the procedure itself.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires real-time clinical judgment, hands-on patient contact, and adaptive responses to dynamic physiological states. Current AI systems cannot perform invasive or therapeutic procedures, interpret live vital signs in context, or execute the fine motor control needed for critical care interventions.
Task automatabilityclaude-sonnet-51/5This requires hands-on physical procedures (e.g., intubation assistance, line management, defibrillation) performed on a patient's body, which current AI cannot physically execute.
Adoption barriersclaude-haiku-4-5-202510015/5This task is legally and professionally gated: nursing licensure, scope-of-practice laws, medical liability, patient safety regulations, and the requirement for human accountability create hard barriers to any form of automation or delegation to non-licensed agents.
Adoption barriersclaude-sonnet-55/5Performing therapeutic/diagnostic procedures on patients legally requires a licensed nurse or clinician, with strict scope-of-practice, liability, and regulatory requirements.
Cost vs. human wageclaude-haiku-4-5-202510011/5The cost of AI infrastructure, oversight, liability, and regulatory compliance to attempt procedure automation far exceeds the loaded wage of a critical care nurse, and the human must remain present regardless.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing physical clinical procedures, so cost comparison favors the human by default since AI cannot deliver the output at all.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI product performs approved therapeutic or diagnostic procedures on patients. This remains entirely within human professional practice; AI tools may assist with decision support, but cannot substitute for the nurse executing the procedure.
Technical feasibility todayclaude-sonnet-51/5No deployed product performs hands-on therapeutic or diagnostic procedures on ICU patients; robotic surgery/nursing systems remain narrow, human-supervised, and not applicable to this scope.

Supervise and monitor unit nursing staff.

1

CI 0–3 · exposure 0 · augmentation 38 · importance 4.1/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Healthcare organizations have not adopted AI for supervisory functions over nursing staff; this remains entirely human-staffed and is unlikely to change given regulatory requirements and the critical governance role supervision plays in patient safety accountability.
Sector adoption velocityclaude-sonnet-52/5Healthcare, especially clinical supervisory functions, adopts AI slowly due to regulatory, safety, and licensure constraints, with pilots focused on documentation rather than supervision.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could provide modest assistance via alerts on staffing gaps, shift pattern analytics, or flagging unusual incident clusters, but human supervisors remain the sole decision-maker. The task itself is primarily relational and evaluative, offering limited scope for meaningful augmentation.
Augmentation potentialclaude-sonnet-53/5AI can assist with staffing analytics, scheduling optimization, and monitoring compliance metrics, helping supervisors make better-informed staffing decisions, but does not replace the human oversight role.
Task automatabilityclaude-haiku-4-5-202510011/5Supervising and monitoring unit nursing staff requires complex human judgment about staff performance, interpersonal dynamics, workload distribution, and real-time responsiveness to changing unit conditions. Current AI systems cannot independently assess staff competence, make staffing decisions, or handle the nuanced management situations that arise daily in critical care units.
Task automatabilityclaude-sonnet-51/5Supervising and monitoring nursing staff in a critical care unit requires real-time human judgment, leadership, accountability, and interpersonal management that AI cannot perform end-to-end today.
Adoption barriersclaude-haiku-4-5-202510015/5Nursing supervision is legally and professionally mandated to be performed by a licensed registered nurse manager; regulatory bodies (state boards, Joint Commission) require human oversight and accountability for unit operations, staff conduct, and patient safety.
Adoption barriersclaude-sonnet-55/5Supervisory nursing roles require licensure, clinical accountability, and legal responsibility for patient safety and staff oversight, making this a hard barrier to automation.
Cost vs. human wageclaude-haiku-4-5-202510011/5The human nursing supervisor performing this task earns a professional salary (typically $70k–$120k annually with overhead) for full-shift presence and decision-making. No AI system can provide equivalent supervision at lower cost, nor does one exist that could reduce this cost burden.
Cost vs. human wageclaude-sonnet-51/5There is no AI system replacing this supervisory function, so no meaningful cost comparison exists; the human cost is the only viable option.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI product performs nursing supervision and staff monitoring autonomously in production settings. This task fundamentally requires a licensed nurse manager to observe behavior, make performance judgments, and take corrective actions that remain the legal responsibility of human leadership.
Technical feasibility todayclaude-sonnet-51/5No deployed product performs staff supervision or performance monitoring in clinical units; this remains a human managerial function with no automated substitute in production.

Coordinate patient care conferences.

1

CI 0–3 · exposure 0 · augmentation 38 · importance 3.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Healthcare remains a highly regulated, risk-averse sector with slow AI adoption in high-stakes clinical coordination tasks; critical care environments prioritize human presence and accountability over automation.
Sector adoption velocityclaude-sonnet-52/5Healthcare, especially critical care nursing, has been slower to adopt AI-driven coordination tools compared to information/finance sectors, with adoption largely limited to documentation and decision-support pilots.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could offer modest assistance (e.g., scheduling suggestions, pre-meeting documentation summaries, reminder systems), but the core task of facilitating multidisciplinary dialogue and ensuring consensus around patient care decisions remains human-centric and offers limited room for meaningful augmentation.
Augmentation potentialclaude-sonnet-53/5AI can assist by summarizing patient records, drafting agendas, or flagging key clinical data ahead of conferences, improving efficiency, but the coordination and facilitation itself remains human-led.
Task automatabilityclaude-haiku-4-5-202510011/5Coordinating patient care conferences requires nuanced interpersonal judgment, negotiation among multidisciplinary teams, and real-time responsiveness to conflicting clinical priorities and human concerns. Current AI systems cannot autonomously schedule, facilitate, synthesize diverse expert opinions, or manage the complex social dynamics inherent to these meetings.
Task automatabilityclaude-sonnet-51/5Coordinating a care conference requires synthesizing complex clinical judgment, scheduling with multiple stakeholders, facilitating real-time interpersonal discussion, and making care decisions—AI cannot perform this end-to-end today with equal quality.
Adoption barriersclaude-haiku-4-5-202510015/5Coordinating care conferences involves legally and professionally mandated human accountability; nurses must personally ensure all stakeholders are informed, present, and heard. Clinical and family trust, liability, and regulatory standards (e.g., informed consent documentation) create hard barriers to AI substitution.
Adoption barriersclaude-sonnet-55/5Care coordination in critical settings involves licensed clinical judgment, legal accountability, and interpersonal trust with patients/families, making this an area with strong regulatory and liability barriers to automation.
Cost vs. human wageclaude-haiku-4-5-202510011/5AI systems cannot yet reduce the labor cost of conference coordination because the task remains fundamentally dependent on human decision-makers, stakeholders, and attendees; any AI tooling would be supplementary to, not substitutive for, nursing coordination.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing this task, so cost comparison favors the human nurse entirely; any AI tools only add marginal support cost, not replacement value.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product reliably coordinates actual patient care conferences end-to-end; this task demands live interaction with physicians, specialists, families, and care teams where human judgment and authority are essential and non-delegable.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously coordinates and runs multidisciplinary patient care conferences in ICU settings; this remains firmly a human clinical leadership function.

Administer medications intravenously, by injection, orally, through gastric tubes, or by other methods.

0

CI 0–0 · exposure 0 · augmentation 38 · importance 4.8/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Despite decades of healthcare automation, medication administration remains a core nursing function performed by humans everywhere. No sector trend toward autonomous AI-driven medication administration exists in deployed clinical practice.
Sector adoption velocityclaude-sonnet-51/5Direct physical medication administration in critical care shows essentially no AI displacement or robotic automation in production settings.
Augmentation potentialclaude-haiku-4-5-202510012/5AI can assist with dose checking, drug-interaction alerts, and patient record lookup, but the physical act and clinical judgment of medication delivery itself remains the nurse's responsibility; augmentation is limited to peripheral decision support rather than transformative productivity gain.
Augmentation potentialclaude-sonnet-53/5AI can assist with dosage calculations, drug interaction checks, and infusion pump programming alerts, improving safety and efficiency while the nurse still performs the physical administration.
Task automatabilityclaude-haiku-4-5-202510011/5Administering medications requires direct physical manipulation of needles, IV lines, and patient bodies in real-time, plus immediate clinical judgment about patient response. Current AI cannot perform these embodied medical interventions without human execution, regardless of time savings.
Task automatabilityclaude-sonnet-51/5Physically administering medications via IV, injection, or gastric tube requires hands-on manipulation of patients in critical condition, which current AI systems cannot perform without robotic embodiment far beyond deployed capability.
Adoption barriersclaude-haiku-4-5-202510015/5Medication administration is legally restricted to licensed nurses or physicians in virtually all jurisdictions. Regulatory bodies (FDA, state nursing boards) explicitly require human licensure and accountability for this task, creating insurmountable legal barriers to automation.
Adoption barriersclaude-sonnet-55/5Medication administration is tightly regulated, requires licensed nursing credentials, and carries high liability for errors, making it legally restricted to authorized human personnel.
Cost vs. human wageclaude-haiku-4-5-202510011/5Even hypothetically, robotic drug administration systems would cost orders of magnitude more per task than the fully-loaded wage of a nurse administering medications, especially when including safety certification, maintenance, and liability insurance.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing this physical task, so AI cost is effectively infinite relative to human labor for this specific action.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI system can independently administer medications via any route. This remains purely human-performed work in all clinical settings; research on automated drug delivery exists but is not in production use by hospitals.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product administers medications to ICU patients; this remains a physical nursing act performed exclusively by licensed humans.

Collect specimens for laboratory tests.

0

CI 0–0 · exposure 0 · augmentation 13 · importance 4.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Healthcare lags in automation of direct patient-contact tasks due to regulatory, safety, and liability constraints. No measurable displacement of specimen collection to automated systems is occurring in critical care settings.
Sector adoption velocityclaude-sonnet-51/5Physical bedside clinical tasks in hospitals show minimal AI/robotic adoption due to safety, regulatory, and infrastructure constraints.
Augmentation potentialclaude-haiku-4-5-202510011/5AI offers minimal assistance for the physical act of specimen collection itself, though data management systems may streamline downstream labeling and tracking. The core task—sterile puncture and collection—remains entirely nurse-executed.
Augmentation potentialclaude-sonnet-52/5AI can help with labeling, tracking, and documentation of specimens but offers little assistance with the physical act of collection itself.
Task automatabilityclaude-haiku-4-5-202510011/5Specimen collection (blood draws, swabs, etc.) requires direct physical contact with patients, sterile technique, and real-time adaptation to anatomical variability. Current AI systems have no dexterous embodiment to perform this task end-to-end in clinical settings.
Task automatabilityclaude-sonnet-51/5Specimen collection (blood draws, catheter samples, line access) requires physical manipulation of a patient's body in a sterile, safety-critical manner that current AI systems cannot perform end-to-end.'
Adoption barriersclaude-haiku-4-5-202510015/5Specimen collection involves direct patient contact, requires sterile technique compliance, and is typically embedded in nursing licensure and clinical protocols. Regulatory and liability frameworks mandate that a qualified healthcare professional perform or directly supervise this invasive procedure.
Adoption barriersclaude-sonnet-55/5Specimen collection in critical care requires licensed clinical staff performing invasive procedures on patients, with strict regulatory, safety, and liability requirements precluding automation.
Cost vs. human wageclaude-haiku-4-5-202510011/5Any conceivable AI system capable of specimen collection (specialized robotic hardware, integration, maintenance, oversight) would cost far more than the loaded wage of a nurse performing the task directly.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI-based substitute performing this physical task, so cost comparison favors the human nurse entirely.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product reliably performs patient specimen collection independently. While robotic arms exist in research, they lack the dexterity, safety verification, and clinical integration needed for production use in critical care.
Technical feasibility todayclaude-sonnet-51/5No deployed clinical product autonomously collects specimens from ICU patients; robotic phlebotomy devices remain experimental and not used in critical care settings.

Advocate for patients' and families' needs, or provide emotional support for patients and their families.

0

CI 0–0 · exposure 0 · augmentation 25 · importance 4.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Healthcare, particularly critical care, lags in AI adoption for patient-facing relational work; trust, regulatory oversight, and the irreducible human element mean this task is rarely targeted for automation even in digitally advanced settings.
Sector adoption velocityclaude-sonnet-51/5Healthcare bedside care, especially critical care nursing, is a slow-adopting, highly regulated, physically embodied sector with minimal AI penetration into direct emotional/advocacy work.
Augmentation potentialclaude-haiku-4-5-202510012/5AI might assist nurses with identifying which families need support or flagging emotional distress signals from medical data, but the core act of advocacy and emotional presence cannot be meaningfully augmented by current AI; the human relationship is central, not peripheral.
Augmentation potentialclaude-sonnet-52/5AI can help with documentation, translation, or surfacing relevant patient/family information, but offers minimal direct assistance to the core emotional and advocacy interaction itself.
Task automatabilityclaude-haiku-4-5-202510011/5This task fundamentally requires empathetic human presence, trust-building, and nuanced understanding of individual emotional states and family dynamics. Current AI systems cannot authentically advocate or provide emotional support in ways that substitute for human care.
Task automatabilityclaude-sonnet-51/5Emotional support and patient advocacy require genuine human presence, empathy, and relational trust that current AI cannot replicate or perform end-to-end in a clinical setting.
Adoption barriersclaude-haiku-4-5-202510015/5Strong legal and ethical barriers exist: nursing licensure requires human judgment, patient trust and consent favor human contact, and liability for failed emotional support or missed advocacy needs falls on the organization, effectively requiring licensed nurses to remain responsible.
Adoption barriersclaude-sonnet-55/5Licensed nursing staff are required for patient care and advocacy in critical care settings, with strong regulatory, ethical, and liability requirements for human presence and accountability.
Cost vs. human wageclaude-haiku-4-5-202510011/5Even if narrow automation were possible, the cost of oversight, liability management, and ensuring appropriate human handoff would exceed the cost of direct human care, making the economic case unfavorable.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute delivering this output, so cost comparison favors the human nurse entirely; any AI attempt would add cost without replacing the function.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product reliably performs genuine patient advocacy or emotional support; chatbots exist but fall far short of the relational and clinical judgment required in critical care settings where stakes are high and families are in distress.
Technical feasibility todayclaude-sonnet-51/5No deployed product performs bedside emotional support or family advocacy in ICU settings; this remains firmly outside current product capabilities.

Assist physicians with procedures such as bronchoscopy, endoscopy, endotracheal intubation, or elective cardioversion.

0

CI 0–0 · exposure 0 · augmentation 25 · importance 4.1/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Critical care settings are high-touch, procedure-centric environments with slow adoption of automation; nurses remain tightly integrated into patient safety protocols and procedure workflows with little digitization of the bedside assistance role itself.
Sector adoption velocityclaude-sonnet-51/5Critical care nursing involves physical, high-stakes bedside procedures in a sector with low digitization of hands-on tasks; AI adoption for physical procedural assistance is essentially nonexistent.
Augmentation potentialclaude-haiku-4-5-202510012/5While AI-assisted vital sign monitoring dashboards or procedure-step documentation exist, they offer only marginal assistance to the nurse's core procedural role. Real-time AI guidance during active procedures remains minimal and experimental.
Augmentation potentialclaude-sonnet-52/5AI can help with pre-procedure documentation, checklist reminders, or monitoring data interpretation, but offers minimal assistance to the actual physical act of assisting with these procedures.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires real-time physical assistance during invasive medical procedures—holding instruments, positioning patients, monitoring vital signs in situ, and responding to dynamic clinical changes. Current AI systems cannot perform these embodied, procedurally-dependent actions.
Task automatabilityclaude-sonnet-51/5This requires hands-on physical assistance, sterile technique, and real-time manual coordination with a physician during invasive procedures—no AI system can perform physical clinical assistance today.
Adoption barriersclaude-haiku-4-5-202510015/5Procedural assistance is fundamentally tied to licensed nursing practice and direct patient safety responsibilities. Regulatory frameworks, hospital credentialing, malpractice liability, and legal requirements for a qualified human in the room create hard barriers to substitution.
Adoption barriersclaude-sonnet-55/5Licensed nursing credentials, hands-on patient safety requirements, and strict regulatory/liability frameworks around invasive procedures make this task legally and physically restricted to authorized human clinicians.
Cost vs. human wageclaude-haiku-4-5-202510011/5The specialized physical and cognitive labor of a critical care nurse during procedures is relatively low-cost compared to physician time; any autonomous system capable of procedural assistance would require expensive robotics, integration, and liability infrastructure far exceeding the wage replacement value.
Cost vs. human wageclaude-sonnet-51/5There is no AI product performing this physical task, so cost comparison favors the human by default; any hypothetical robotic assistance would be far more costly than a nurse's wage today.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI system can autonomously assist physicians during live bronchoscopy, endoscopy, intubation, or cardioversion. These procedures demand precise physical coordination, real-time decision-making, and direct patient contact that only robotic systems (still nascent and tele-operated, not autonomous AI) approach.
Technical feasibility todayclaude-sonnet-51/5No deployed product provides physical procedural assistance; this remains purely a human physical-presence task with no robotic or AI substitute in production.

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.