Acute Care Nurses

29-1141.01
Median wage $97,550/yr3,379,720 employed (US)Rank #814 of 923 scored · top 88% by substitution

Provide advanced nursing care for patients with acute conditions such as heart attacks, respiratory distress syndrome, or shock. May care for pre- and post-operative patients or perform advanced, invasive diagnostic or therapeutic procedures.

Sub-scores

0–100 · band = confidence interval from rater disagreement

Substitution14
Exposure14
Augmentation56

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

27 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%16

panel mean rating 1.7/5 → substitution pressure 16/100

Cost vs. human wagew 15%14

panel mean rating 1.6/5 → substitution pressure 14/100

Adoption barriersw 20%inverted — strong barriers lower the score12

panel mean rating 4.5/5 (barrier strength) → substitution pressure 12/100

Sector adoption velocityw 10%20

panel mean rating 1.8/5 → substitution pressure 20/100

Task breakdown (27 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.

Document data related to patients' care, including assessment results, interventions, medications, patient responses, or treatment changes.

36

CI 30–41 · exposure 42 · augmentation 75 · importance 4.6/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510013/5Hospitals and health systems are piloting AI documentation assistants (e.g., ambient voice capture), but adoption remains uneven; deployment is growing but most organizations still rely heavily on manual documentation, with production-scale AI-only documentation rare.
Sector adoption velocityclaude-sonnet-52/5Healthcare, especially acute inpatient nursing, adopts AI more slowly than white-collar sectors due to regulatory complexity, EHR fragmentation, and safety-critical workflows, though pilots are growing.
Augmentation potentialclaude-haiku-4-5-202510014/5AI-powered speech-to-text, template suggestions, and auto-population of structured fields substantially reduce documentation burden and time for nurses who remain in the loop, meaningfully raising their documentation productivity—this is one of the clearer augmentation wins in acute care.
Augmentation potentialclaude-sonnet-54/5AI scribes, templated note generation, and structured data extraction meaningfully speed up documentation while the nurse remains responsible for verification and clinical judgment.
Task automatabilityclaude-haiku-4-5-202510012/5While structured data entry (vitals, medications) can be partially automated through EHR integration, the task requires clinical judgment to document assessment results, interventions, and patient responses—nuanced narrative elements that current AI cannot reliably capture without human nurses making the key clinical decisions and initial entries.
Task automatabilityclaude-sonnet-53/5AI can draft clinical notes from ambient dictation or structured inputs, but nurses must verify accuracy, capture nuanced observations, and ensure legal/clinical completeness, limiting full end-to-end automation.
Adoption barriersclaude-haiku-4-5-202510014/5Strong legal and regulatory barriers exist: nurses must personally verify and sign documentation for liability and compliance (HIPAA, state board standards), and many organizations require human nurses to review and authenticate all entries—the nurse's legal signature is often non-delegable.
Adoption barriersclaude-sonnet-54/5Medical documentation is a legally required, auditable record tied to licensure and liability; a qualified nurse must review and attest to accuracy, creating strong regulatory and liability barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI documentation tools are relatively inexpensive to license, but the required nurse review and correction time means the all-in cost remains comparable to or higher than the time savings realized, given that nurses still perform most of the cognitive work.
Cost vs. human wageclaude-sonnet-52/5AI scribe tools reduce documentation time but still require licensing fees, integration with EHRs, and mandatory human review/edit, so savings are real but not order-of-magnitude versus nurse wages.
Technical feasibility todayclaude-haiku-4-5-202510013/5Speech-to-text and template-based EHR systems exist and are deployed, but they require significant human oversight and correction; AI-assisted documentation tools show promise but still struggle with clinical accuracy and produce material error rates when left unreviewed.
Technical feasibility todayclaude-sonnet-53/5Ambient documentation and voice-to-note products (e.g., Nuance DAX, various EHR-integrated tools) are deployed in some hospitals, but adoption for nursing-specific charting is narrower and less mature than for physician documentation.

Interpret information obtained from electrocardiograms (EKGs) or radiographs (x-rays).

33

CI 30–36 · exposure 34 · augmentation 75 · importance 4.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Healthcare remains a laggard sector for autonomous AI deployment despite digitization. Adoption is concentrated in pilots and narrow decision-support use cases; production replacement of nurse or physician interpretation is rare and typically confined to triage or screening contexts rather than primary diagnosis.
Sector adoption velocityclaude-sonnet-53/5Healthcare is adopting AI diagnostic aids at a moderate pace with growing pilot and some production deployments (e.g., stroke/PE detection tools), but overall clinical workflow integration remains slower than in pure information-sector industries.
Augmentation potentialclaude-haiku-4-5-202510014/5AI augmentation is strong: real-time alerting on EKG abnormalities or incidental radiograph findings can substantially boost nurse efficiency and catch sensitivity. Current decision-support tools meaningfully assist nurses in flagging findings for escalation, reducing cognitive load and improving clinical safety.
Augmentation potentialclaude-sonnet-54/5AI-based EKG and radiograph analysis tools meaningfully speed up screening and flag urgent findings, allowing nurses to prioritize and verify results faster while remaining the final decision-maker.
Task automatabilityclaude-haiku-4-5-202510012/5While AI excels at pattern recognition in EKGs and radiographs in controlled settings, clinical interpretation requires contextualizing findings with patient history, symptoms, and clinical judgment. Current AI can flag abnormalities but cannot reliably perform the full interpretive task end-to-end with quality matching experienced nurses at ≥50% time savings.
Task automatabilityclaude-sonnet-52/5AI can flag abnormalities in EKGs and x-rays with reasonable accuracy but clinical interpretation integrating patient context, history, and decision-making still requires licensed human judgment, so full end-to-end automation at equal quality is not yet achieved.
Adoption barriersclaude-haiku-4-5-202510014/5Regulatory barriers are substantial: radiologists and physicians are legally required to interpret diagnostic images in most jurisdictions; nurses support but do not independently sign interpretations. Liability, malpractice risk, and established clinical governance create hard friction against autonomous AI substitution.
Adoption barriersclaude-sonnet-54/5Clinical interpretation of diagnostic tests typically requires sign-off by licensed clinicians (nurses, physicians, or radiologists) due to liability and regulatory requirements, creating a substantial barrier to full substitution.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI inference cost is low, but integration into clinical workflows, regulatory compliance, human oversight, and liability management add significant overhead. The all-in cost remains comparable to or exceeds that of a nurse performing the task, especially accounting for mandatory secondary review.
Cost vs. human wageclaude-sonnet-53/5AI-assisted interpretation tools have licensing and integration costs that are meaningful relative to marginal nurse time saved, making cost roughly comparable rather than dramatically cheaper, especially given required human oversight.
Technical feasibility todayclaude-haiku-4-5-202510013/5FDA-cleared AI tools exist for EKG and radiograph analysis (e.g., EKG abnormality detection, chest X-ray findings), but they operate as decision-support aids with material error rates and typically narrow scope (specific pathologies). Production deployment remains under radiologist/clinician oversight rather than autonomous.
Technical feasibility todayclaude-sonnet-53/5FDA-cleared products exist for ECG arrhythmia detection and radiograph triage (e.g., fracture, pneumothorax detection) deployed in some hospitals, but they are narrow-scope decision-support tools rather than autonomous interpreters used at scale across all acute care settings.

Perform administrative duties that facilitate admission, transfer, or discharge of patients.

28

CI 25–30 · exposure 25 · augmentation 63 · importance 4.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Healthcare sectors have adopted AI slowly for administrative tasks due to regulatory caution, integration complexity with legacy EHRs, and organizational resistance. Pilots are common but production adoption of end-to-end administrative automation remains limited.
Sector adoption velocityclaude-sonnet-52/5Healthcare administrative AI adoption is growing but slower than in finance or professional services due to compliance, interoperability, and EHR fragmentation issues.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can meaningfully assist nurses by auto-populating forms, flagging missing documents, and suggesting discharge protocols, reducing manual data entry and cognitive load. However, the human nurse must still review, authorize, and troubleshoot complex cases.
Augmentation potentialclaude-sonnet-54/5AI scribes and documentation assistants meaningfully speed up drafting of admission/discharge notes and administrative paperwork, letting nurses focus more on patient care while reviewing AI-generated content.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist with form population and documentation from structured data, the task requires coordination across multiple systems, patient-specific exceptions, insurance verification, and legal compliance that demand human judgment and real-time problem-solving. Current systems cannot reliably handle the full workflow end-to-end.
Task automatabilityclaude-sonnet-52/5Some sub-components like generating discharge paperwork or transfer summaries can be drafted by AI, but coordinating across systems, verifying clinical status, and physically executing admission/transfer/discharge workflows require human presence and judgment.
Adoption barriersclaude-haiku-4-5-202510014/5Healthcare automation faces strong regulatory barriers (HIPAA, state licensing laws), liability exposure if errors occur in discharge paperwork, and legal requirements that a licensed nurse sign off on transfers and discharges. Institutional workflows are tightly governed.
Adoption barriersclaude-sonnet-53/5No strict licensing requirement for administrative paperwork itself, but liability, patient safety documentation standards, and institutional policy create meaningful friction against full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI tools for healthcare administration are expensive (licensing, EHR integration, compliance oversight), and nursing staff must still verify and correct outputs, meaning labor savings are modest compared to total deployment cost.
Cost vs. human wageclaude-sonnet-52/5AI drafting tools can reduce documentation time cheaply, but the overall administrative task still requires nurse verification, system interaction, and compliance checks, keeping total cost comparable to human-only performance.
Technical feasibility todayclaude-haiku-4-5-202510012/5Some hospitals have deployed AI for limited administrative tasks (appointment scheduling, basic documentation), but these are narrow use cases with high error rates when exceptions arise. No production system reliably performs the complete admission/transfer/discharge bundle independently.
Technical feasibility todayclaude-sonnet-52/5EHR-integrated tools and ambient documentation assistants exist and are being piloted for discharge summaries, but reliable end-to-end automation of admission/transfer/discharge administrative workflows in production is limited and error-prone.

Read current literature, talk with colleagues, and participate in professional organizations or conferences to keep abreast of developments in acute care.

28

CI 16–39 · exposure 17 · augmentation 75 · importance 4.1/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5While healthcare organizations use AI for literature retrieval aids, the core professional development requirement—nurse judgment and participation—remains unmechanized. Adoption of AI as a supplementary tool is growing slowly; substitution for the human practice is not occurring.
Sector adoption velocityclaude-sonnet-52/5Healthcare has been slower than information/finance sectors to deploy AI tools for continuing education, though literature summarization tools are gaining some traction among individual clinicians.
Augmentation potentialclaude-haiku-4-5-202510014/5AI can effectively summarize recent literature, flag relevant clinical updates, and organize conference materials, meaningfully reducing the nurse's time to digest new knowledge. LLM-based tools and literature aggregation services demonstrably assist nurses in staying current faster and more comprehensively.
Augmentation potentialclaude-sonnet-54/5AI literature summarization, alerting on new studies, and conference content curation can meaningfully speed up how acute care nurses stay current, even though it doesn't replace the human networking element.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires subjective judgment about relevance, synthesis across diverse sources, and meaningful peer discussion—activities that depend on human expertise, context, and professional discretion. Current AI cannot autonomously maintain professional knowledge currency at the standard required of practicing nurses.
Task automatabilityclaude-sonnet-52/5AI can summarize literature and surface relevant articles, but the task inherently includes networking, discussion, and conference participation that require human presence and judgment, limiting full automation.
Adoption barriersclaude-haiku-4-5-202510014/5Professional licensure and scope-of-practice regulations require that nurses themselves maintain competency and stay current with their field. Regulatory bodies and employers mandate individual professional development, and this cannot be legally delegated to an AI system.
Adoption barriersclaude-sonnet-52/5No licensing requirement blocks using AI tools for literature review, though professional certification/CE requirements may mandate documented human participation in accredited activities.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI literature retrieval and summarization are cheap, but the human nurse must still read, interpret, and participate in discussions themselves. The AI offsets only a fraction of the human time cost; the nurse's engagement remains the dominant cost.
Cost vs. human wageclaude-sonnet-53/5AI-assisted literature review is cheap relative to a nurse's time spent reading, but the colleague discussion and conference attendance portions have no AI cost-saving equivalent, balancing overall ratio.
Technical feasibility todayclaude-haiku-4-5-202510012/5While AI can retrieve and summarize literature, it cannot meaningfully participate in collegial discussions or conferences, evaluate nuanced clinical developments, or make professional judgment calls about what matters. No deployed product reliably performs the full task of keeping a nurse abreast of developments in the human-judgment sense.
Technical feasibility todayclaude-sonnet-52/5Products like AI literature summarizers and research digests exist, but no deployed system replaces the professional networking and conference engagement components of this task.

Provide formal and informal education to other staff members.

28

CI 25–30 · exposure 25 · augmentation 75 · importance 4.1/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5While hospitals increasingly use learning management systems and AI-assisted content platforms, the actual replacement of nurse educators with autonomous AI is rare. Adoption remains limited to supplementary tools rather than end-to-end automation of the educational role.
Sector adoption velocityclaude-sonnet-52/5Healthcare, especially bedside clinical education, is a historically slow-adopting sector for AI-driven training tools compared to information/finance industries.
Augmentation potentialclaude-haiku-4-5-202510014/5AI can meaningfully assist nurse educators by generating presentation slides, summarizing clinical evidence, designing quiz items, and personalizing learning paths, substantially raising productivity while the nurse remains the primary educator and decision-maker.
Augmentation potentialclaude-sonnet-54/5AI can meaningfully help nurses draft training materials, in-services, competency checklists, and educational content, boosting efficiency while the nurse still delivers the instruction.
Task automatabilityclaude-haiku-4-5-202510012/5AI can help generate training materials and draft educational content, but delivering contextual, interactive education to staff requires live adaptation, relationship-building, and real-time feedback that current systems cannot reliably provide end-to-end. The interpersonal and judgment-heavy aspects of teaching are beyond current AI capabilities.
Task automatabilityclaude-sonnet-52/5Peer education to nursing staff relies on clinical credibility, adaptive in-person demonstration, and answering unpredictable clinical questions, which current AI cannot fully replicate end-to-end even though it can help prepare materials.
Adoption barriersclaude-haiku-4-5-202510014/5Healthcare organizations rely on credentialed, accountable staff to train others on clinical protocols and patient safety—there is strong institutional expectation and often regulatory/accreditation requirements that licensed clinicians lead formal education. Liability and patient safety concerns create high friction against full automation.
Adoption barriersclaude-sonnet-53/5No licensing barrier prevents AI-assisted content, but organizational norms, trust, and the value of experienced-nurse mentorship create moderate friction against full substitution.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI-generated training materials and learning platforms have low per-unit cost, but when accounting for the customization, validation, and human oversight required for clinical education, the total integrated cost remains comparable to or exceeds the cost of a nurse educator delivering training directly.
Cost vs. human wageclaude-sonnet-52/5AI can cheaply produce supporting materials, but the actual teaching, mentoring, and informal knowledge transfer still requires paid nurse time, so overall cost savings are limited.
Technical feasibility todayclaude-haiku-4-5-202510012/5While AI tools can draft educational documents and identify knowledge gaps via chatbots, no deployed product reliably replaces a human educator conducting formal or informal staff training in a clinical setting. Existing systems handle content generation at best, not the full educational interaction.
Technical feasibility todayclaude-sonnet-52/5AI tools exist for generating training content, quizzes, and slide decks, but no deployed product actually delivers formal/informal peer education to nursing staff autonomously in clinical settings.

Assist patients in organizing their health care system activities.

28

CI 25–30 · exposure 25 · augmentation 75 · importance 3.6/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Healthcare adoption of AI in direct patient care activities remains cautious and heavily regulated; while scheduling tools exist, meaningful adoption of AI for core nursing coordination tasks is still in pilot phase rather than production at scale.
Sector adoption velocityclaude-sonnet-52/5Healthcare is a historically slow-adopting sector for AI-driven patient coordination tools, with pilots more common than widespread production deployment for this specific task.
Augmentation potentialclaude-haiku-4-5-202510014/5AI can meaningfully assist nurses by automating appointment scheduling, flagging medication interactions, organizing records, and summarizing patient information, thereby raising nurse productivity while the nurse maintains clinical judgment and patient communication.
Augmentation potentialclaude-sonnet-54/5AI can meaningfully assist nurses by organizing patient records, generating reminders, summarizing care plans, and flagging scheduling conflicts, improving efficiency while the nurse retains responsibility for judgment and patient interaction.
Task automatabilityclaude-haiku-4-5-202510012/5Organizing health care activities involves understanding individual patient context, preferences, and complex medical needs. While AI could help with scheduling or information structuring, the task requires personalized judgment, empathy, and real-time adaptation to patient circumstances that current systems cannot reliably automate end-to-end.
Task automatabilityclaude-sonnet-52/5This task requires ongoing personalized coordination, judgment about clinical priorities, and interpersonal trust-building that current AI cannot fully replicate end-to-end.wait
Adoption barriersclaude-haiku-4-5-202510014/5Nursing is a licensed profession with legal requirements for human accountability in patient care. Direct patient contact, regulatory requirements around care coordination, and organizational liability for health care decisions create substantial barriers to full automation.
Adoption barriersclaude-sonnet-53/5No strict licensing requirement mandates a nurse specifically organize these activities, but liability concerns, patient trust, and the need for clinical judgment in coordinating care create moderate friction against full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5Implementing and maintaining AI systems for personalized patient health care organization requires substantial integration and human oversight costs, making the all-in expense comparable to or exceeding a nurse's time for this task.
Cost vs. human wageclaude-sonnet-52/5While AI tools can cheaply handle reminders and scheduling, the human oversight, clinical judgment, and relationship management needed keep overall costs comparable to or only modestly cheaper than a nurse performing this task.
Technical feasibility todayclaude-haiku-4-5-202510012/5Some tools exist for appointment scheduling and basic information management, but no deployed system reliably performs the full task of patient health care organization with the interpersonal nuance and clinical context required; most solutions remain narrow or research-stage.
Technical feasibility todayclaude-sonnet-52/5Some care-coordination and patient-navigation software exists, but deployed products mostly assist with scheduling or reminders rather than comprehensively organizing a patient's healthcare activities reliably.

Collaborate with patients to plan for future health care needs or to coordinate transitions and referrals.

25

CI 25–25 · exposure 25 · augmentation 63 · importance 4.1/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Despite healthcare's digital investment, autonomous AI for patient care coordination has seen minimal production adoption; most tools remain decision-support only. Regulatory constraints, liability concerns, and the requirement for licensed human sign-off keep velocity low compared to information-sector automation.
Sector adoption velocityclaude-sonnet-52/5Healthcare, especially bedside acute care, has been slower than sectors like finance or tech to deploy AI agents directly into patient-facing coordination workflows.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist nurses by suggesting relevant referral options, summarizing past transitions, and drafting communication templates, improving workflow efficiency. However, the core collaborative and clinical judgment elements remain nurse-led, so augmentation is bounded to administrative and information-synthesis tasks rather than transformative.
Augmentation potentialclaude-sonnet-54/5AI can meaningfully assist by summarizing patient records, drafting referral letters, and flagging care gaps, improving nurse efficiency while the nurse remains central to patient interaction.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist in organizing referral information and drafting transition plans, this task fundamentally requires human judgment about patient values, shared decision-making, and individualized care coordination that current AI cannot reliably replicate. The collaborative, trust-based nature of planning future care needs with vulnerable patients remains largely out of reach for autonomous systems.
Task automatabilityclaude-sonnet-52/5This task requires relational trust-building, negotiation of patient values, and judgment about complex clinical transitions that current AI cannot fully replicate end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Strong regulatory and licensing barriers exist: only licensed nurses (and sometimes social workers) can legally perform care coordination and discharge planning in many jurisdictions. Liability for wrong transitions, patient trust requirements, and clinical accountability create high legal and organizational friction against substitution.
Adoption barriersclaude-sonnet-54/5Care planning and referral coordination typically require licensed nurse involvement for legal, clinical safety, and continuity-of-care reasons, creating strong regulatory and liability barriers.
Cost vs. human wageclaude-haiku-4-5-202510012/5The domain-specific oversight, clinical validation, and human-in-the-loop requirements needed to safely deploy AI in this task mean end-to-end costs remain high relative to a nurse's labour, with no clear cost advantage for the automation that exists today.
Cost vs. human wageclaude-sonnet-52/5While AI can cut some documentation time, the human relational and coordination work still dominates cost, so overall savings versus nurse wages are modest.
Technical feasibility todayclaude-haiku-4-5-202510012/5No production system reliably performs autonomous patient care planning or transition coordination. While AI tools exist for scheduling referrals or surfacing care guidelines, deployed systems do not independently collaborate with patients to assess their needs and coordinate complex multi-provider transitions at scale.
Technical feasibility todayclaude-sonnet-52/5AI tools exist for care-coordination documentation and referral drafting, but no deployed product autonomously conducts patient collaboration and discharge planning at production scale.

Participate in the development of practice protocols.

24

CI 23–25 · exposure 25 · augmentation 63 · importance 3.8/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Healthcare adoption of AI for independent protocol development remains minimal; most health systems retain tight human control over clinical standards and governance. Organizations are highly risk-averse in delegating protocol authorship to systems lacking clear accountability.
Sector adoption velocityclaude-sonnet-52/5Healthcare, especially clinical protocol development, adopts AI slowly due to regulatory caution, liability concerns, and the specialized, high-stakes nature of the work.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist nurses in protocol development by summarizing evidence, generating draft language, and organizing research—useful productivity gains when nurses remain primary decision-makers. However, the augmentation is narrow and depends on human validation of clinical content.
Augmentation potentialclaude-sonnet-54/5AI tools can meaningfully assist by summarizing literature, drafting protocol language, and flagging evidence, improving efficiency while nurses retain final judgment and approval.
Task automatabilityclaude-haiku-4-5-202510012/5Developing practice protocols requires domain expertise, stakeholder input, and judgment about clinical best practices that AI cannot fully execute autonomously. While AI could draft sections or synthesize evidence, the task fundamentally requires human clinical judgment and organizational context that AI tools cannot replicate at scale.
Task automatabilityclaude-sonnet-52/5Drafting protocol language can be assisted by AI, but developing clinical practice protocols requires synthesizing evidence, institutional judgment, and multidisciplinary consensus that current AI cannot perform end-to-end reliably.
Adoption barriersclaude-haiku-4-5-202510014/5Protocol development typically requires sign-off by clinical leadership, institutional governance, and often regulatory bodies. Legal and professional liability for protocol content creates strong barriers to full automation, and institutional policy often mandates human accountability for clinical standards.
Adoption barriersclaude-sonnet-54/5Practice protocols typically require sign-off by licensed clinical staff and often institutional/regulatory review, creating strong governance and liability barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5Current AI tools require significant human oversight and iteration to produce usable protocol components, making the all-in cost competitive with or exceeding a nurse's time investment in the task. Loaded costs for validation and liability remain high.
Cost vs. human wageclaude-sonnet-52/5AI can cheaply generate draft text or evidence summaries, but the overall task still requires costly expert clinician time for review, validation, and consensus-building, keeping total cost comparable to human-only process.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed products reliably perform independent protocol development for clinical settings. AI can assist with literature synthesis and draft generation, but current systems lack the reliability and accountability needed for healthcare protocol authorship in production environments.
Technical feasibility todayclaude-sonnet-52/5Some clinical decision-support and literature-summarization tools exist, but no deployed product autonomously develops or finalizes clinical practice protocols in production hospital settings.

Distinguish between normal and abnormal developmental and age-related physiological and behavioral changes in acute, critical, and chronic illness.

23

CI 20–25 · exposure 25 · augmentation 50 · importance 3.9/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5While hospitals deploy vital-sign monitoring and alerting systems, adoption of AI for real-time developmental and age-specific physiological judgment remains slow; most use is still in pilots or narrow applications (sepsis screening) rather than widespread production replacement of this core nursing task.
Sector adoption velocityclaude-sonnet-52/5Healthcare, especially acute/critical care bedside nursing, has historically slow AI adoption due to regulatory, safety, and workflow integration challenges compared to information-sector professions.
Augmentation potentialclaude-haiku-4-5-202510013/5Clinical decision-support dashboards and alert systems help nurses by surfacing abnormal patterns and providing reference data on age-appropriate norms, moderately raising their efficiency in screening and documentation without removing their judgment role.
Augmentation potentialclaude-sonnet-53/5AI-based monitoring systems and clinical decision support can flag deviations from expected physiological trends, helping nurses prioritize attention, though behavioral/developmental judgment remains largely human-driven.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can help analyze vital signs and flag statistical deviations from norms, the task requires integrating multiple subtle physiological and behavioral cues in context—fever patterns, mental status shifts, pain presentation across age groups—that still demand real-time clinical judgment. Current AI cannot reliably perform this end-to-end with 50% time savings at equal safety.
Task automatabilityclaude-sonnet-52/5This requires real-time clinical assessment integrating patient history, physical exam, and behavioral observation at bedside, which current AI cannot perform end-to-end without a human clinician physically present and interpreting cues.asa AI can support pattern recognition from structured data but cannot replace the holistic bedside judgment.rationale trimmed
Adoption barriersclaude-haiku-4-5-202510014/5Patient safety liability is severe; misclassifying normal development as pathology or vice versa can delay critical care or cause unnecessary intervention. Regulatory oversight (FDA for clinical decision support), nursing licensure requirements, and organizational risk management create strong friction against full automation.
Adoption barriersclaude-sonnet-55/5This is a core licensed nursing scope-of-practice task involving direct patient assessment and clinical judgment, with strict regulatory and liability requirements mandating human professional performance.
Cost vs. human wageclaude-haiku-4-5-202510012/5Monitoring and alerting tools cost thousands annually per patient or institution, while a nurse's wages for this cognitive work are spread across many patients. AI does not yet deliver order-of-magnitude savings, and integration overhead for reliable clinical AI remains high.
Cost vs. human wageclaude-sonnet-52/5Bedside clinical judgment integrating physiological and behavioral assessment still requires a licensed nurse; AI tools that assist add cost on top of, not instead of, the nurse's wage.
Technical feasibility todayclaude-haiku-4-5-202510012/5Decision-support tools exist to flag abnormal vital signs and lab values, but no deployed system reliably distinguishes normal age-related changes from pathological ones across the full range of acute, critical, and chronic presentations. Clinical validation and liability concerns keep most applications to narrow alerting rather than autonomous differentiation.
Technical feasibility todayclaude-sonnet-52/5Some clinical decision support tools flag abnormal vitals or lab trends, but no deployed product independently distinguishes normal vs abnormal developmental/behavioral changes across acute, critical, and chronic contexts reliably in production.

Analyze the indications, contraindications, risk complications, and cost-benefit tradeoffs of therapeutic interventions.

20

CI 20–20 · exposure 25 · augmentation 75 · importance 3.9/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Healthcare is slow-moving on AI autonomy; adoption has focused on narrow diagnostic support (imaging, labs) rather than therapeutic decision-making. Clinical staff remain skeptical of fully autonomous clinical reasoning, and institutional friction is high.
Sector adoption velocityclaude-sonnet-52/5Healthcare adoption of AI for direct clinical decision-making is cautious and slow due to regulatory, safety, and liability concerns, despite faster uptake in administrative and documentation tasks.
Augmentation potentialclaude-haiku-4-5-202510014/5AI can significantly assist by rapidly retrieving evidence, mapping contraindications against patient history, flagging drug interactions, and organizing cost-benefit summaries—tools that extend nurse capability in preparing and supporting therapeutic analysis while the nurse retains decision authority.
Augmentation potentialclaude-sonnet-54/5AI tools can meaningfully speed up gathering evidence on interventions, drug interactions, and risk data, helping nurses reason through tradeoffs faster while they retain final judgment.
Task automatabilityclaude-haiku-4-5-202510012/5This task requires integration of clinical evidence, patient-specific factors, and nuanced judgment about competing risks—areas where AI can assist with literature synthesis and guideline lookup but cannot independently weigh contraindications or patient context with the reliability required for clinical decisions. Current AI cannot autonomously perform this analysis end-to-end at the quality threshold for clinical practice.
Task automatabilityclaude-sonnet-52/5AI can support literature synthesis and risk data retrieval, but integrating patient-specific clinical judgment, uncertainty, and real-time bedside context into final therapeutic decisions remains beyond full automation today.
Adoption barriersclaude-haiku-4-5-202510015/5Healthcare regulation, standard of care requirements, and malpractice liability create a hard barrier: nurses and physicians retain legal responsibility for clinical decisions, and autonomous AI recommendations cannot satisfy the regulatory and ethical requirement that a qualified human clinician must evaluate and sign off on therapeutic choices.
Adoption barriersclaude-sonnet-55/5This is core licensed clinical judgment; regulations and standard of care require a credentialed nurse or physician to make and be accountable for these determinations.
Cost vs. human wageclaude-haiku-4-5-202510012/5The integration, validation, and compliance overhead for clinical decision support, combined with required human review and liability costs, means the total cost of AI-assisted analysis is comparable to or exceeds direct nurse labor for this cognitively demanding task.
Cost vs. human wageclaude-sonnet-52/5While AI querying is cheap, the human oversight, liability review, and integration into clinical workflow keep effective costs comparable to or only modestly below the cost of trained nursing judgment for this task.
Technical feasibility todayclaude-haiku-4-5-202510012/5While AI tools can retrieve guidelines and summarize evidence, no deployed product reliably performs independent analysis of therapeutic indications, contraindications, and risk-benefit tradeoffs in production clinical settings without substantial human oversight. Decision-support systems exist but require careful physician/nurse validation and are not autonomous.
Technical feasibility todayclaude-sonnet-52/5Clinical decision-support tools exist and are used for guideline lookups and drug interaction checks, but no deployed product autonomously performs full cost-benefit and contraindication analysis reliably at the point of care.

Order, perform, or interpret the results of diagnostic tests and screening procedures based on assessment results, differential diagnoses, and knowledge about age, gender and health status of clients.

18

CI 16–20 · exposure 25 · augmentation 75 · importance 4.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Healthcare lags in AI adoption due to regulatory caution, liability concerns, and data fragmentation. While EHR-embedded diagnostic decision support is growing, autonomous AI ordering and interpretation remain in pilot phases; production displacement is minimal.
Sector adoption velocityclaude-sonnet-52/5Healthcare, especially acute/inpatient care, adopts AI diagnostics cautiously due to regulatory approval (FDA), liability concerns, and integration into EHR workflows, resulting in slow, pilot-heavy adoption relative to other professional sectors.
Augmentation potentialclaude-haiku-4-5-202510014/5AI-powered diagnostic assistants meaningfully augment nurses by rapidly surfacing relevant tests, highlighting abnormal results, and flagging differential diagnoses for human review. These tools can increase speed and reduce cognitive load, keeping the nurse in the decision loop.
Augmentation potentialclaude-sonnet-54/5AI decision-support tools, risk-scoring algorithms, and diagnostic-imaging aids meaningfully help nurses interpret test results faster and flag anomalies, while the nurse retains responsibility for final clinical decisions.
Task automatabilityclaude-haiku-4-5-202510012/5AI can assist with test selection and interpretation of standardized results (e.g., analyzing lab values or imaging), but the task requires clinical judgment integrating patient assessment, differential diagnosis reasoning, and individualized health status—elements that demand human expertise. The assessment-to-order decision chain and liability for clinical accuracy prevent >50% autonomous time savings.
Task automatabilityclaude-sonnet-52/5AI can assist in interpreting some diagnostic images or lab patterns, but ordering tests requires clinical judgment integrating patient context, and final interpretation for acute care decisions still requires licensed clinician oversight, so it doesn't meet the 50% end-to-end automation bar today.
Adoption barriersclaude-haiku-4-5-202510015/5This task is embedded in healthcare licensure and scope-of-practice regulations. Nurses must possess clinical judgment credentials, and test ordering/interpretation carries legal liability that requires a licensed professional signature or direct supervision—hard regulatory and liability barriers prevent substitution.
Adoption barriersclaude-sonnet-55/5Ordering and interpreting diagnostic tests in an acute care setting is a licensed clinical act with direct patient safety and liability implications, requiring a credentialed nurse or provider by law and institutional policy.
Cost vs. human wageclaude-haiku-4-5-202510011/5AI diagnostic support requires integration costs, human clinician oversight, and malpractice risk management. The labor cost of a nurse performing this task is modest relative to integration and liability costs, making AI more expensive all-in.
Cost vs. human wageclaude-sonnet-52/5AI diagnostic tools have real per-use costs (imaging AI licenses, integration, compliance validation) and still require clinician review, so overall cost savings versus a nurse's judgment-integrated workflow are modest, not order-of-magnitude cheaper.
Technical feasibility todayclaude-haiku-4-5-202510012/5Clinical decision support tools exist for test interpretation, but deployed systems have significant limitations: they narrow scope (specific conditions/populations), generate false positives/negatives, and require clinician review/override. No production system reliably performs the full end-to-end task (assess → order → interpret) autonomously.
Technical feasibility todayclaude-sonnet-52/5Deployed products exist for narrow diagnostic aids (e.g., ECG interpretation, radiology triage flagging) but no product independently orders and interprets the full range of diagnostic tests for acute care patients in production without clinician oversight.

Assess the impact of illnesses or injuries on patients' health, function, growth, development, nutrition, sleep, rest, quality of life, or family, social and educational relationships.

18

CI 11–25 · exposure 20 · augmentation 50 · importance 3.8/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Despite health IT advancement, adoption of AI for autonomous clinical assessment in acute care remains limited to pilot/research settings; most deployments are narrow decision-support tools requiring nurse validation, and organizational inertia around credential and liability requirements slows replacement.
Sector adoption velocityclaude-sonnet-52/5Healthcare, especially bedside acute care nursing, has historically slow AI adoption for hands-on clinical assessment tasks despite documentation-support tools emerging.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can meaningfully assist by organizing and flagging relevant data (comorbidities, lab trends, functional decline patterns), generating preliminary summaries of documented impacts, and prompting assessment of domains nurses might overlook—but the nurse remains essential for direct patient interaction and final judgment synthesis.
Augmentation potentialclaude-sonnet-53/5AI can help synthesize patient history, suggest assessment considerations, and support documentation, but the core relational and physical assessment work remains human-driven.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can extract and summarize clinical data (vital signs, lab results, imaging), the core assessment requires holistic judgment about patient-specific impacts on function, development, nutrition, sleep, relationships, and psychosocial factors—all of which demand direct observation, patient communication, and contextual understanding that current systems cannot reliably perform end-to-end at 50% time savings with equal quality.
Task automatabilityclaude-sonnet-52/5This requires hands-on clinical examination, patient interaction, and holistic judgment integrating physical exam findings with psychosocial context that current AI cannot independently perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Strong regulatory and professional barriers exist: nursing assessment is legally within the scope of licensed RN practice, clinical liability for missed impacts (especially on vulnerable pediatric or development-dependent patients) is severe, and most acute care settings have institutional policies requiring nursing evaluation before major care decisions.
Adoption barriersclaude-sonnet-55/5Licensed nursing scope of practice, legal accountability for patient assessment, and mandatory human clinical judgment make this a hard-barrier task.
Cost vs. human wageclaude-haiku-4-5-202510012/5The specialized expertise of acute care nurses and the need for supervision/validation of AI output means total cost (AI + oversight + liability) remains comparable to or higher than direct human assessment, particularly in high-acuity settings where errors are costly.
Cost vs. human wageclaude-sonnet-51/5AI cannot independently deliver this task, so any comparison favors the human nurse who must be present regardless of AI tool costs.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed product reliably performs this multidimensional impact assessment independently; AI can assist with data synthesis and flagging abnormalities, but real-time clinical judgment about complex interplay of physical, developmental, nutritional, and social factors remains a human skill that deployed systems support rather than replace.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously performs comprehensive clinical impact assessments on patients; this remains a core in-person nursing function without production-scale substitutes.

Assess urgent and emergent health conditions, using both physiologically and technologically derived data.

16

CI 11–20 · exposure 17 · augmentation 75 · importance 4.8/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5While hospitals rapidly adopt monitoring and decision-support tools, actual displacement of nurse assessment in urgent/emergent triage remains minimal. Adoption is limited to augmentation (EHR integration, alerting) rather than replacement, reflecting both regulatory constraints and clinical risk aversion.
Sector adoption velocityclaude-sonnet-52/5Healthcare, especially acute/inpatient care, has historically been slow to adopt autonomous AI decision-making at the bedside due to regulatory, safety, and workflow integration challenges, though monitoring tools are spreading gradually.
Augmentation potentialclaude-haiku-4-5-202510014/5AI augmentation is strong and proven: real-time vital sign monitoring, automated alerts, integrated lab/imaging data synthesis, and predictive acuity scoring materially improve nurse productivity and decision quality. Many acute care nurses now rely on AI-enhanced monitoring and data aggregation as standard of care.
Augmentation potentialclaude-sonnet-54/5AI-driven monitoring systems, early warning scores, and data aggregation dashboards meaningfully help nurses synthesize physiological and technological data faster, improving situational awareness while the nurse retains clinical judgment and action.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist in triaging and analyzing technologically derived data (vital signs, lab results, imaging), clinical assessment of urgent conditions requires real-time integration of physical examination, patient history, and contextual judgment that current AI systems cannot reliably perform end-to-end. Human nurses must remain present for physical assessment and decision accountability.
Task automatabilityclaude-sonnet-51/5Real-time bedside assessment of acutely deteriorating patients requires hands-on physical examination, synthesis of ambiguous physiological signals, and immediate clinical judgment that current AI cannot perform end-to-end without a licensed clinician physically present and deciding.
Adoption barriersclaude-haiku-4-5-202510015/5Strong legal and regulatory barriers exist: nursing licensure is required by law, clinical assessment is a protected scope of practice, and liability for missed or incorrect urgent/emergent diagnosis falls on the responsible clinician. Hospitals have fiduciary and regulatory obligations that prevent substitution of human judgment in acute care.
Adoption barriersclaude-sonnet-55/5Acute condition assessment is a licensed clinical act with direct patient safety and legal liability implications, requiring a credentialed nurse or clinician to perform and be accountable for the assessment.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI infrastructure for continuous monitoring and decision support is expensive to deploy and maintain in hospital settings, and still requires nurse oversight and validation. The integrated cost per assessment remains higher than the nursing labor it might reduce, given liability and integration overhead.
Cost vs. human wageclaude-sonnet-52/5AI-based monitoring/alerting systems are cheap to run, but they only supplement rather than replace the nurse's assessment, so the full task still requires the human's wage cost, keeping the effective cost ratio unfavorable to full automation.
Technical feasibility todayclaude-haiku-4-5-202510012/5Clinical decision support systems exist and assist with data analysis, but no deployed product reliably performs independent urgent/emergent assessment at the quality and safety required for acute care. AI excels at pattern-matching in structured data but lacks the multimodal, embodied judgment and liability acceptance needed for front-line triage.
Technical feasibility todayclaude-sonnet-52/5Deployed clinical decision support tools (e.g., sepsis alerts, early warning scores) exist and are used in production, but they only flag risk from data streams and do not perform the actual hands-on assessment reliably or autonomously.

Refer patients for specialty consultations or treatments.

16

CI 6–25 · exposure 13 · augmentation 63 · importance 4.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Acute care remains a traditionally conservative sector with significant regulatory oversight and strong preference for human clinical judgment. Pilot implementations of decision-support tools exist, but production adoption of autonomous referral systems is minimal compared to information-sector automation.
Sector adoption velocityclaude-sonnet-52/5Healthcare, especially acute care nursing, has slow, cautious AI adoption due to regulatory and safety concerns, with pilots for documentation but not autonomous referral decisions.
Augmentation potentialclaude-haiku-4-5-202510014/5AI assistants that summarize patient data, flag relevant clinical indicators, or suggest specialty options based on guidelines can substantially raise nurse productivity in the referral intake and documentation phase while the nurse retains final clinical authority and sign-off.
Augmentation potentialclaude-sonnet-53/5AI can assist by flagging relevant specialists, summarizing patient history, or drafting referral paperwork, improving efficiency while the nurse retains decision authority.
Task automatabilityclaude-haiku-4-5-202510012/5AI could assist in identifying when referrals are indicated based on clinical guidelines and patient data, but the task fundamentally requires human clinical judgment to weigh individual patient context, severity, and complex medical histories. Current systems cannot reliably make autonomous triage and referral decisions with the safety margin required in acute care.
Task automatabilityclaude-sonnet-51/5Referring patients requires clinical judgment, legal accountability, and interpersonal coordination with patients and other providers that current AI cannot perform end-to-end.'
Adoption barriersclaude-haiku-4-5-202510014/5Strong legal and professional barriers exist: nurses operate under scope-of-practice regulations that expect human accountability, physicians typically must authorize specialty referrals, and liability asymmetries mean errors in referral triage can harm patients significantly. Healthcare licensing and duty-of-care standards create hard friction against full automation.
Adoption barriersclaude-sonnet-55/5Referrals require licensed clinical judgment and carry direct liability for patient outcomes, making this a hard legal/regulatory barrier to automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI-assisted decision support tools have meaningful implementation and licensing costs, and the time savings are partial—nurses still perform the core judgment and documentation. The all-in cost per referral decision remains comparable to or higher than the direct nurse time involved.
Cost vs. human wageclaude-sonnet-52/5While AI-assisted documentation could lower administrative time slightly, the referral decision still requires a licensed clinician's involvement, keeping overall cost comparable to human-only process.
Technical feasibility todayclaude-haiku-4-5-202510012/5While AI systems exist that can flag potential referral candidates or recommend specialty types based on symptom matching, no deployed product reliably performs the full referral decision independently in production acute-care settings. Clinical decision support tools exist but require substantial human oversight and final judgment.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously initiates specialty referrals; existing clinical decision support tools only suggest options for human review.

Collaborate with members of multidisciplinary health care teams to plan, manage, or assess patient treatments.

14

CI 3–25 · exposure 13 · augmentation 75 · importance 4.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Healthcare sectors remain cautious adopters of full care-team automation; most AI deployment in acute care focuses on narrow clinical support (alerting, data flagging) rather than collaborative team decision-making. Organizational resistance, regulatory conservatism, and liability concerns slow adoption of autonomous care coordination.
Sector adoption velocityclaude-sonnet-52/5Healthcare clinical workflows adopt AI slowly due to regulation, liability, and safety-critical nature, with most deployment limited to documentation and decision-support pilots rather than care-team substitution.
Augmentation potentialclaude-haiku-4-5-202510014/5AI demonstrates strong augmentation potential for this task: EHR-integrated clinical decision support, treatment protocol suggestions, and data synthesis can meaningfully boost nurse productivity and team coordination while keeping nurses in the loop for judgment and accountability.
Augmentation potentialclaude-sonnet-54/5AI tools (e.g., clinical decision support, summarization of patient data, predictive alerts) can meaningfully support nurses' contributions to team discussions and care planning, improving efficiency and information synthesis.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist with data aggregation and treatment plan documentation, the collaborative decision-making, interpersonal negotiation, and real-time clinical judgment required across multidisciplinary teams cannot be fully automated. AI lacks the contextual reasoning and human accountability essential to treatment planning.
Task automatabilityclaude-sonnet-51/5This requires real-time clinical judgment, physical presence, interpersonal negotiation, and accountability across a live care team, none of which current AI can perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Significant regulatory and liability barriers protect this task: clinical decision-making in acute care is governed by scope-of-practice laws, licensure requirements, and standard-of-care expectations that mandate qualified human accountability. Patient safety regulations and malpractice liability mean organizations cannot fully delegate treatment planning to AI.
Adoption barriersclaude-sonnet-55/5Licensure, scope-of-practice laws, clinical liability, and mandatory human sign-off on patient care decisions create hard legal and regulatory barriers to automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI infrastructure and integration costs for clinical collaboration systems are substantial, and they still require nursing staff to oversee and validate recommendations. The cost per task-equivalent remains comparable to or higher than direct human collaboration without displacement.
Cost vs. human wageclaude-sonnet-51/5Since AI cannot substitute for the collaborative clinical role itself, there is no viable AI-only cost comparison; any use requires the human nurse's full involvement plus additional tooling cost.
Technical feasibility todayclaude-haiku-4-5-202510012/5Current AI systems can support collaboration through documentation and protocol suggestions, but no deployed product reliably executes the full collaborative planning and assessment task independently. Clinical decision support tools exist but require human oversight and do not replace the team interaction itself.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product independently collaborates in multidisciplinary rounds or manages patient treatment plans; existing tools are decision-support aids used by humans, not autonomous participants.

Diagnose acute or chronic conditions that could result in rapid physiological deterioration or life-threatening instability.

13

CI 11–14 · exposure 9 · augmentation 75 · importance 4.9/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Healthcare sectors, especially acute care, are slow to adopt autonomous AI decision-making in clinical diagnosis due to liability, regulatory caution, and entrenched clinical workflows. While AI-assisted decision support pilots are increasing, autonomous diagnostic replacement in production acute care settings remains rare and typically narrow in scope.
Sector adoption velocityclaude-sonnet-53/5Hospitals are adopting predictive analytics and early-warning AI systems at a moderate pace, but full diagnostic autonomy remains rare due to regulatory and safety constraints in healthcare.
Augmentation potentialclaude-haiku-4-5-202510014/5AI augmentation of nursing diagnosis is already meaningful: real-time vital-sign monitoring dashboards, sepsis prediction algorithms, and clinical decision support tools assist nurses in pattern recognition and prioritization. These systems demonstrably help nurses recognize deterioration faster and raise diagnostic awareness, though the nurse retains diagnostic and action authority.
Augmentation potentialclaude-sonnet-54/5AI-based early-warning systems, risk scoring, and continuous monitoring significantly enhance a nurse's ability to detect deterioration early, improving speed and accuracy while the clinician remains the decision-maker.
Task automatabilityclaude-haiku-4-5-202510011/5Clinical diagnosis of acute or chronic conditions requiring rapid physiological assessment demands integration of physical examination, continuous monitoring, and real-time judgment in high-stakes environments. Current AI cannot perform the full diagnostic workflow—particularly bedside assessment, vital-sign interpretation in context, and the synthesis of subtle clinical cues—reliably enough to meet the 50% time-saving bar without substantial human oversight that negates time savings.
Task automatabilityclaude-sonnet-51/5Diagnosing acute, rapidly evolving physiological instability requires real-time physical assessment, integration of continuously changing vitals, and hands-on clinical judgment that current AI cannot perform end-to-end without a licensed clinician present.
Adoption barriersclaude-haiku-4-5-202510015/5Diagnosis and clinical judgment in acute settings are legally and professionally scoped to licensed clinicians (nurses, physicians, advanced practitioners); liability for missed or incorrect diagnosis rests with the responsible human, and regulatory frameworks (scope of practice, standards of care) require human clinical judgment at each decision point. No automation bypasses this licensing and liability structure.
Adoption barriersclaude-sonnet-55/5Diagnosing life-threatening conditions is legally restricted to licensed clinical professionals, with strict scope-of-practice, liability, and regulatory requirements preventing full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI diagnostic tools (licenses, infrastructure, integration, required clinical oversight) remain costly relative to the loaded wage of an experienced nurse who performs diagnosis as part of a broader clinical role. The overhead of validation and liability means the all-in cost per autonomous diagnostic episode is not yet substantially cheaper than the nurse's time.
Cost vs. human wageclaude-sonnet-52/5AI decision-support systems are relatively cheap to run, but since a licensed nurse/physician must still perform the diagnostic act, the AI cost is additive rather than substitutive, keeping overall costs comparable or higher due to integration and liability oversight.
Technical feasibility todayclaude-haiku-4-5-202510012/5While AI tools exist for limited diagnostic support (e.g., pattern matching on vital signs or ECG interpretation in narrow contexts), no deployed product reliably performs comprehensive acute diagnosis with the safety margins required in ICU or emergency settings. Clinical decision support exists in production but is narrow in scope and always requires human validation before action.
Technical feasibility todayclaude-sonnet-52/5Clinical decision-support tools (sepsis alerts, early-warning scores) are deployed in hospitals but function only as adjuncts flagging risk, not as autonomous diagnosticians making final acute-care determinations.

Discuss illnesses and treatments with patients and family members.

9

CI 3–16 · exposure 5 · augmentation 50 · importance 4.9/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Healthcare remains a low-digitization, highly regulated sector where clinical decisions and patient communication are tightly guarded by professional licensing. Observed AI adoption in hospitals focuses on documentation and administrative tasks, not clinical discussion; patient-facing adoption is minimal and cautious.
Sector adoption velocityclaude-sonnet-52/5Healthcare overall adopts AI slowly for direct patient-facing clinical communication due to regulatory, liability, and trust barriers, though administrative AI use is growing.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can draft talking points, summarize medical concepts, or flag questions a nurse should address, moderately enhancing preparation and consistency. However, the nurse remains central to delivery, and AI's assistive role is limited by the need for clinical accuracy and genuine human presence.
Augmentation potentialclaude-sonnet-53/5AI can help nurses prepare patient-friendly explanations, summarize records, or draft educational materials to support these conversations, but doesn't replace the interpersonal exchange itself.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires nuanced human communication, empathy, and the ability to tailor explanations to individual emotional and cognitive states. Current AI cannot reliably replace the interpersonal judgment, trust-building, and dynamic responsiveness that patients and families need during vulnerable moments.
Task automatabilityclaude-sonnet-51/5This requires real-time, empathetic, high-stakes clinical communication with patients and grieving/anxious family members that demands trust, physical presence, and licensed judgment; AI cannot substitute end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Nursing regulations and scope-of-practice laws in most jurisdictions require a licensed nurse to provide patient education and counseling. Additionally, patients and families strongly prefer human contact for sensitive health discussions, and legal liability for misstatement is asymmetrically borne by the organization if AI provides incorrect clinical guidance.
Adoption barriersclaude-sonnet-55/5Nursing licensure, scope-of-practice laws, and the fundamental human-contact/trust requirement in acute care make this task legally and ethically reserved for a human clinician.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI-generated health explanations have low infrastructure cost but require significant human oversight, fact-checking, and liability management. The all-in cost remains higher than simply having a nurse conduct the conversation, particularly given error-cost sensitivity in clinical contexts.
Cost vs. human wageclaude-sonnet-51/5Human nurses remain necessary for licensure, liability, and trust reasons, so AI cannot meaningfully substitute the labor cost for this task despite being cheap per query.
Technical feasibility todayclaude-haiku-4-5-202510012/5While chatbots can generate generic health information, no deployed system reliably handles the full scope of emotionally-aware, context-sensitive clinical discussion with patients and families in real healthcare settings. Limited demos exist but production systems remain narrow and lack clinical credibility.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously conducts clinical discussions about illness/treatment with patients and families in acute care settings; chatbots exist for basic FAQs but not for this core relational task.

Manage patients' pain relief and sedation by providing pharmacologic and non-pharmacologic interventions, monitoring patients' responses, and changing care plans accordingly.

9

CI 3–16 · exposure 13 · augmentation 50 · importance 4.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Healthcare organizations are slowly adopting AI-assisted documentation and clinical decision tools, but production-level autonomous pain and sedation management remains rare. Most deployment is in pilot or advisory phases due to high-stakes clinical and regulatory constraints.
Sector adoption velocityclaude-sonnet-52/5Healthcare, especially bedside acute care, adopts AI slowly due to regulatory, safety, and workflow integration challenges, with pilots more common than production deployment.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can meaningfully assist by suggesting evidence-based pain scores, dosing recommendations, and monitoring alerts that help nurses make faster decisions; however, the task remains fundamentally human-centered, and augmentation is incremental rather than transformative.
Augmentation potentialclaude-sonnet-53/5AI-based clinical decision support and monitoring analytics (e.g., early warning scores, sedation scales) can help nurses track responses and suggest adjustments, aiding but not replacing judgment.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist in pain-assessment protocols and drug-dosing recommendations, the task requires real-time bedside patient monitoring, physical administration of interventions, and adaptive clinical judgment in response to patient state changes—none of which current AI can perform autonomously. At most, AI-assisted decision support covers part of the planning phase.
Task automatabilityclaude-sonnet-51/5This requires hands-on physical assessment, medication administration, and real-time clinical judgment in a bedside setting that current AI cannot perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Nursing regulations, scope of practice laws, and clinical liability frameworks require a licensed registered nurse to assess pain, order/administer medications, monitor sedation, and adjust care plans. Legal and professional accountability is non-delegable to automated systems.
Adoption barriersclaude-sonnet-55/5Administering sedatives/analgesics and adjusting care plans requires licensed nursing/physician authority, strict regulatory oversight, and direct liability for patient safety.
Cost vs. human wageclaude-haiku-4-5-202510011/5The task requires continuous bedside presence, physical intervention capability, and accountability that AI cannot replicate; human nurses remain essential. The marginal cost of AI tools is small compared to the loaded wage of a registered nurse, and AI does not displace that cost.
Cost vs. human wageclaude-sonnet-51/5AI cannot substitute for the licensed nurse's physical presence and dosing authority, so there is no viable cost comparison for full task replacement.
Technical feasibility todayclaude-haiku-4-5-202510012/5Clinical decision-support systems exist for pain management and sedation protocols, but no deployed product reliably performs the full cycle of monitoring, assessment, intervention delivery, and adaptive adjustment without human clinicians in primary control. Products remain advisory rather than autonomous.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product independently manages pain relief or sedation for acute care patients; decision-support tools exist but are not autonomous performers of this task.

Set up, operate, or monitor invasive equipment and devices, such as colostomy or tracheotomy equipment, mechanical ventilators, catheters, gastrointestinal tubes, and central lines.

4

CI 0–7 · exposure 5 · augmentation 50 · importance 4.6/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Healthcare organizations have not meaningfully adopted AI for autonomous invasive device setup or operation due to patient safety requirements, regulatory constraints, and the physical nature of the work; adoption remains minimal despite digitization of other hospital functions.
Sector adoption velocityclaude-sonnet-52/5Healthcare bedside/physical care tasks show slow AI adoption compared to administrative or diagnostic support functions, constrained by regulation and physical necessity.
Augmentation potentialclaude-haiku-4-5-202510013/5AI-driven monitoring systems and predictive alerts can help nurses detect equipment problems earlier and optimize parameters, improving their situational awareness and decision-making, though the core manual and clinical skills remain firmly human-driven.
Augmentation potentialclaude-sonnet-53/5AI-enabled monitoring systems and smart alarms can assist by flagging abnormal readings from ventilators or catheters, improving situational awareness while the nurse remains responsible for hands-on care.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires direct physical manipulation of invasive medical equipment, real-time clinical assessment of patient responses, and immediate manual adjustments that only a present human can perform. AI cannot currently set up catheters, adjust ventilator parameters based on tactile feedback, or respond to equipment failures at the bedside.
Task automatabilityclaude-sonnet-51/5This requires hands-on physical setup, insertion, and continuous monitoring of invasive medical equipment on patients, which current AI cannot physically perform or safely execute end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Strong legal and regulatory barriers mandate that licensed nurses perform or directly oversee invasive equipment management; liability for patient harm from equipment mismanagement is severe, and clinical credentialing laws require human accountability.
Adoption barriersclaude-sonnet-55/5Licensed nursing scope-of-practice laws, patient safety regulations, and liability requirements mandate that only trained, licensed clinical staff perform invasive procedures and monitoring.
Cost vs. human wageclaude-haiku-4-5-202510011/5Invasive equipment setup and monitoring require licensed clinical personnel with years of training; the fully-loaded cost of an acute care nurse is substantially lower than any viable robotic or autonomous system capable of performing this task safely in a hospital setting.
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute performing this physical task, so cost comparison favors the human nurse entirely; AI cannot deliver the output at all.
Technical feasibility todayclaude-haiku-4-5-202510012/5While AI can monitor equipment telemetry and alert nurses to anomalies, no deployed system can autonomously operate or set up invasive devices in clinical practice. Some hospital IT systems provide monitoring dashboards, but the core task of physical setup and real-time operation remains exclusively human.
Technical feasibility todayclaude-sonnet-51/5No deployed product physically operates or manages invasive devices like ventilators, catheters, or central lines; this remains entirely within human clinical practice.

Adjust settings on patients' assistive devices, such as temporary pacemakers.

1

CI 0–3 · exposure 0 · augmentation 25 · importance 4.7/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5No measurable AI adoption exists for this task because regulatory frameworks and clinical standards explicitly require human operators. Healthcare remains highly regulated and slow to substitute licensed clinical judgment with automated systems.
Sector adoption velocityclaude-sonnet-52/5Acute care nursing involves direct physical patient care in hospital settings where AI adoption for hands-on device manipulation remains minimal despite AI's broader growth in healthcare documentation and diagnostics.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could marginally assist by providing real-time data visualization or recommending parameter ranges based on patient history, but current systems offer limited practical augmentation for a task dominated by tactile adjustment, immediate patient feedback interpretation, and clinical accountability.
Augmentation potentialclaude-sonnet-52/5AI can support monitoring and alerting on device parameters or patient vitals to inform nurse decisions, but it does not meaningfully assist the physical act of adjusting the device itself.
Task automatabilityclaude-haiku-4-5-202510011/5Adjusting pacemaker settings requires real-time assessment of patient physiology, manual dexterity, and immediate clinical judgment in response to patient feedback—tasks that current AI systems cannot perform autonomously. No off-the-shelf AI can reliably handle the complex sensor integration and rapid decision-making needed for safe device adjustment.
Task automatabilityclaude-sonnet-51/5This requires hands-on physical manipulation of medical devices with real-time clinical judgment based on patient response; no AI system can perform this physical, high-stakes intervention end-to-end today.
Adoption barriersclaude-haiku-4-5-202510015/5Hard regulatory and licensing barriers mandate that only credentialed clinical staff (nurses, physicians, technicians) can adjust life-critical devices like pacemakers. FDA regulations, clinical liability, and patient safety laws make autonomous or unsupervised AI adjustment legally and ethically prohibited.
Adoption barriersclaude-sonnet-55/5Adjusting temporary pacemaker settings is an invasive, high-risk clinical intervention that legally and ethically requires a licensed, trained nurse or physician physically present and accountable.
Cost vs. human wageclaude-haiku-4-5-202510011/5Current AI systems cannot reduce the cost of this task because they cannot perform it at all; a human nurse must remain fully responsible for the clinical assessment and manual adjustment, making AI cost-additive rather than cost-reducing.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing this physical task, so any comparison is moot—the human is the only current option, making AI effectively infinitely costlier or simply inapplicable.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI product reliably performs pacemaker or assistive device adjustment independently; this remains entirely within clinical staff domain. Regulatory bodies (FDA) classify pacemaker programming as requiring trained human operators with specific clinical certification.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously adjusts pacemaker or assistive device settings on patients; this remains firmly a licensed clinician task performed manually.

Prescribe medications and observe patients' reactions, modifying prescriptions as needed.

1

CI 0–3 · exposure 0 · augmentation 50 · importance 4.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5There is no adoption of AI-driven prescription automation in acute care settings because it is not permitted. Healthcare organizations and regulators maintain strict gatekeeping on prescribing authority to preserve patient safety and professional accountability.
Sector adoption velocityclaude-sonnet-52/5Healthcare broadly has slow, cautious AI adoption for direct clinical decision-making, especially prescribing, due to regulatory and liability constraints.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist nurses in monitoring and reporting medication reactions (flagging adverse events, suggesting when to notify the prescriber) and can support evidence-based recommendations to present to the prescriber, but the nurse does not retain autonomous prescribing authority in this augmented workflow.
Augmentation potentialclaude-sonnet-53/5AI can assist with drug interaction checks, dosing calculators, and flagging abnormal vitals/labs, supporting the nurse's decision-making without replacing judgment.
Task automatabilityclaude-haiku-4-5-202510011/5Prescribing medications is a legally restricted clinical decision that requires a licensed physician or nurse practitioner, not a nurse. Acute care nurses administer and monitor medications but do not prescribe them; this task is outside their scope of practice and cannot be automated by AI without changing fundamental regulatory and professional boundaries.
Task automatabilityclaude-sonnet-51/5Prescribing medication and clinically observing patient reactions requires licensed medical judgment, physical assessment, and legal authority that current AI cannot exercise end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Prescribing is protected by hard regulatory barriers: state medical boards, federal DEA regulations, and malpractice liability require a licensed human clinician (MD, DO, NP, or PA) to write and take responsibility for all prescriptions. Automation is legally and ethically prohibited.
Adoption barriersclaude-sonnet-55/5Prescribing is tightly regulated and requires licensure, DEA/state authority, and legal accountability, making this one of the most protected clinical tasks.
Cost vs. human wageclaude-haiku-4-5-202510011/5This comparison is moot because the task itself (prescribing) is not within nursing scope. AI cannot perform the gated clinical-legal function regardless of cost; a physician or advanced practice provider must remain in control.
Cost vs. human wageclaude-sonnet-51/5AI cannot legally substitute for the licensed nurse/provider performing this task, so there is no viable cost comparison—the human must be paid regardless.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI product can legally or clinically prescribe medications. While AI can support diagnostic reasoning and drug interaction checking, actual prescription authority requires a licensed clinician to make the final decision and bear legal liability, which current systems cannot do.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously prescribes and adjusts medications for acute care patients; clinical decision support exists but always requires a licensed prescriber to act.

Assess the needs of patients' family members or caregivers.

1

CI 0–3 · 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 remains a conservative, heavily regulated sector with slow AI adoption for clinical decision-making. Family/caregiver need assessment is deeply integrated into nursing practice and clinical relationships, showing minimal automation adoption in production settings.
Sector adoption velocityclaude-sonnet-52/5Healthcare overall has lagged in AI adoption for direct patient/family interaction tasks, with AI more used for documentation and diagnostics support than interpersonal assessment.
Augmentation potentialclaude-haiku-4-5-202510012/5Current AI offers minimal augmentation for this task. While chatbots might draft standardized screening forms, the core work—eliciting trust, reading emotional cues, and synthesizing complex family dynamics—remains almost entirely human-dependent; any AI assistance is marginal.
Augmentation potentialclaude-sonnet-52/5AI tools like documentation assistants or care coordination software can offload administrative burden, indirectly giving nurses more time for family engagement, but they don't directly enhance the assessment interaction itself.
Task automatabilityclaude-haiku-4-5-202510011/5Assessing family/caregiver needs requires nuanced interpersonal judgment, emotional intelligence, and contextual understanding of complex social dynamics. Current AI systems lack the ability to conduct reliable, empathetic assessments of psychosocial needs in real clinical settings without human oversight.
Task automatabilityclaude-sonnet-51/5This requires in-person empathetic assessment of emotional, informational, and logistical needs of family members, relying on nuanced human observation and rapport that current AI cannot replicate end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Strong legal and professional barriers exist: acute care nurses are licensed clinicians whose assessments of family needs are part of professional nursing judgment and documentation. Healthcare regulation, liability concerns, and the requirement for a qualified nurse to evaluate and respond to family needs create hard barriers to substitution.
Adoption barriersclaude-sonnet-55/5Licensed nursing scope of practice, hospital regulations, and the inherently human-contact nature of family support create hard barriers against automation.
Cost vs. human wageclaude-haiku-4-5-202510011/5AI systems cannot yet perform this task autonomously, making any cost comparison speculative. The loaded wage of a nurse performing this task is likely lower than the overhead of integrating experimental AI systems with meaningful oversight.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI system replacing this human relational task, so cost comparison favors the human nurse who is already required for care delivery.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI product reliably performs autonomous assessment of family/caregiver needs in acute care environments. This task fundamentally depends on clinical conversation, observation of emotional states, and relationship understanding—domains where AI has no production-grade solutions.
Technical feasibility todayclaude-sonnet-51/5No deployed clinical products perform in-person family needs assessment; this remains a core nursing interpersonal function without AI substitutes in production.

Participate in patients' care meetings and conferences.

1

CI 0–3 · exposure 0 · augmentation 38 · importance 4.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Healthcare adoption of AI for clinical collaboration remains slow; care conferences remain predominantly human-driven, with AI playing no meaningful role in actual meeting participation across sectors.
Sector adoption velocityclaude-sonnet-52/5Healthcare adoption of AI for clinical documentation and decision support is growing, but direct participation in care meetings remains largely untouched by AI deployment.
Augmentation potentialclaude-haiku-4-5-202510012/5AI might assist nurses by preparing summaries or extracting key patient data before a meeting, but does not substantially augment their performance during active participation, discussion, and decision-making in the conference itself.
Augmentation potentialclaude-sonnet-53/5AI can assist by summarizing patient records, generating meeting notes, or flagging relevant clinical data beforehand, improving the nurse's preparation and efficiency during the meeting.
Task automatabilityclaude-haiku-4-5-202510011/5Participating in care meetings and conferences requires real-time interaction, clinical judgment, and collaborative decision-making with interdisciplinary teams. Current AI cannot meaningfully contribute to or replace a nurse's presence and voice in these high-stakes collaborative discussions.
Task automatabilityclaude-sonnet-51/5Participating in care conferences requires physical presence, clinical judgment, real-time interpersonal communication, and accountability that current AI cannot replicate end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Healthcare regulations, professional licensing, and patient safety standards require licensed nurses to participate in care decisions. Legal liability, informed-consent norms, and the requirement for professional accountability create hard barriers to substitution.
Adoption barriersclaude-sonnet-55/5Clinical decision-making and accountability in patient care conferences require a licensed nurse's presence and judgment, with legal, regulatory, and liability requirements mandating human involvement.
Cost vs. human wageclaude-haiku-4-5-202510011/5Even if partial automation were possible (e.g., pre-meeting summarization), the cost of AI tools plus oversight would exceed the cost of a nurse's time spent in actual meetings, and would not deliver equivalent clinical value.
Cost vs. human wageclaude-sonnet-51/5Since AI cannot perform the task itself, there is no viable cost comparison for full substitution; any AI use is supplementary, not replacing the human cost.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product reliably performs a nurse's participatory role in care meetings; AI might summarize meeting notes or assist preparation, but cannot substitute for human professional judgment and accountability in real conferences.
Technical feasibility todayclaude-sonnet-51/5No deployed product substitutes a nurse's participation in multidisciplinary care meetings; at best AI provides note-taking or summarization support, not participation itself.

Perform emergency medical procedures, such as basic cardiac life support (BLS), advanced cardiac life support (ACLS), and other condition-stabilizing interventions.

0

CI 0–0 · exposure 0 · augmentation 25 · importance 4.8/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Hospitals have not adopted AI to replace nurses performing BLS/ACLS procedures. The sector remains dependent on trained human staff for emergency response, with no measurable displacement of this task by AI systems in production environments.
Sector adoption velocityclaude-sonnet-51/5Hands-on acute care procedures in hospital settings show minimal AI displacement; adoption in this specific physical task domain is essentially nonexistent.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could provide limited assistance through real-time protocol reminders, vital sign monitoring dashboards, or decision support (e.g., alerting nurses to medication timing), but the core physical interventions and judgment remain squarely with the human clinician. The augmentation upside is modest because the task is already highly protocolized.
Augmentation potentialclaude-sonnet-52/5AI can support decision protocols, monitoring alerts, or documentation around the event, but offers negligible assistance during the actual physical execution of life-saving procedures.
Task automatabilityclaude-haiku-4-5-202510011/5Emergency medical procedures require real-time physical intervention on a patient's body (chest compressions, intubation, defibrillation). Current AI systems cannot perform these hands-on, time-critical interventions end-to-end; they lack embodied robotics integration and the legal/safety framework to operate autonomously in emergency settings.
Task automatabilityclaude-sonnet-51/5This is a hands-on physical emergency intervention requiring manual dexterity, real-time physical manipulation, and split-second judgment; current AI systems cannot physically perform CPR, defibrillation, or intubation.
Adoption barriersclaude-haiku-4-5-202510015/5Emergency medical procedures are heavily regulated and require a licensed nurse or physician to legally perform and sign off on interventions. Liability for errors during cardiac resuscitation or stabilization is extremely high, and patient safety regulations explicitly require credentialed human providers to execute these time-critical, life-saving procedures.
Adoption barriersclaude-sonnet-55/5Licensed clinical personnel are legally required to perform emergency medical procedures, with strict certification (BLS/ACLS) and liability requirements making substitution essentially impossible.
Cost vs. human wageclaude-haiku-4-5-202510011/5The capital cost of even prototype emergency medical robotics, combined with required oversight and integration, far exceeds the labor cost of a trained nurse performing these interventions. No cost-effective AI alternative exists for this task at scale.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing this physical task, so no meaningful cost comparison exists; a human clinician is the only current option.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI product reliably performs BLS, ACLS, or emergency stabilization procedures in clinical practice today. Research exists on robotic CPR devices, but they operate under human supervision and are not autonomous AI systems capable of independent clinical decision-making and execution in emergency contexts.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product performs physical resuscitation or stabilization procedures; robotics for this remains research-stage at best.

Administer blood and blood product transfusions or intravenous infusions, monitoring patients for adverse reactions.

0

CI 0–0 · exposure 0 · augmentation 50 · importance 4.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Healthcare adoption of AI remains slow and cautious, particularly for direct patient care tasks involving medication and transfusion administration. No measurable displacement of nurses performing infusion administration has occurred; the sector prioritizes human oversight and maintains strong regulatory and cultural resistance to autonomous clinical interventions.
Sector adoption velocityclaude-sonnet-51/5Bedside physical nursing tasks in acute care remain among the least digitized/automated aspects of healthcare, with essentially no AI displacement occurring.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist by providing real-time monitoring alerts, flagging vital sign changes, and helping nurses recognize patterns of adverse reactions faster, improving vigilance. However, the core clinical judgment and physical task remain nurse-dependent, limiting transformative productivity gains.
Augmentation potentialclaude-sonnet-53/5AI-enabled monitoring systems and alert algorithms can help flag abnormal vital signs or reaction risk, assisting nurses' vigilance during transfusions, though the core physical task remains manual.
Task automatabilityclaude-haiku-4-5-202510011/5Administering blood transfusions and IV infusions requires physical manipulation (insertion, monitoring, adjustment of equipment) and real-time clinical judgment to detect subtle adverse reactions that demand immediate intervention. Current AI systems cannot reliably perform the hands-on clinical actions or make split-second safety decisions in dynamic bedside settings.
Task automatabilityclaude-sonnet-51/5This requires physical manipulation of IV lines, blood products, and hands-on patient monitoring for adverse reactions like anaphylaxis, which current AI cannot physically perform.
Adoption barriersclaude-haiku-4-5-202510015/5Administering blood products and monitoring for transfusion reactions is a regulated clinical task that legally requires a licensed nurse to perform and be accountable. Liability for adverse outcomes, patient safety requirements, and professional licensing laws create hard barriers to automation.
Adoption barriersclaude-sonnet-55/5This is a licensed clinical procedure with strict regulatory, safety, and liability requirements mandating a qualified nurse or clinician to administer and monitor transfusions.
Cost vs. human wageclaude-haiku-4-5-202510011/5The labor cost of an acute care nurse administering infusions is low relative to potential liability and oversight requirements for AI systems. Deploying AI to replace this task would require extensive validation, regulatory approval, and continuous supervision, making it more expensive than retaining human nurses.
Cost vs. human wageclaude-sonnet-51/5AI cannot perform the physical administration or hands-on monitoring at all, so there is no viable cost comparison—human nursing labor is required.
Technical feasibility todayclaude-haiku-4-5-202510011/5While AI can assist with monitoring alerts and data analysis, no deployed product autonomously administers transfusions or makes independent clinical decisions about adverse reactions in production. The task requires licensed human clinical judgment and direct patient contact that remains legally and practically non-automatable today.
Technical feasibility todayclaude-sonnet-51/5No deployed product administers transfusions or infusions or performs the physical monitoring required; this remains entirely outside current AI product capability.

Obtain specimens or samples for laboratory work.

0

CI 0–0 · exposure 0 · augmentation 25 · importance 4.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Healthcare remains a conservative, regulation-heavy sector with strong barriers to automating patient-contact tasks; no meaningful production adoption of automated specimen collection exists.
Sector adoption velocityclaude-sonnet-51/5Physical bedside nursing tasks like specimen collection show minimal AI/robotic adoption; healthcare's hands-on procedural tasks are among the least digitized and automated areas.
Augmentation potentialclaude-haiku-4-5-202510012/5AI offers minimal direct augmentation for the core specimen-collection task itself, though digital ordering systems and lab tracking provide peripheral assistance rather than enhancing the collection act.
Augmentation potentialclaude-sonnet-52/5AI can assist with logistics like labeling, tracking, or flagging when samples are due, but offers negligible assistance with the core physical act of obtaining the specimen itself.
Task automatabilityclaude-haiku-4-5-202510011/5Obtaining specimens (blood draws, cultures, urine samples, etc.) requires physical presence, dexterity, and direct patient contact that current AI and robotics cannot reliably perform end-to-end in uncontrolled clinical environments.
Task automatabilityclaude-sonnet-51/5Physically obtaining specimens (blood draws, swabs, catheter samples) from patients requires hands-on manipulation and physical dexterity that current AI systems cannot perform without robotic embodiment, which is not deployed in this context.
Adoption barriersclaude-haiku-4-5-202510015/5Specimen collection requires direct patient contact, clinical licensing, sterile technique certification, and often legal/regulatory authorization; only licensed healthcare workers can legally obtain many specimens.
Adoption barriersclaude-sonnet-55/5Specimen collection requires licensed clinical personnel, involves direct patient contact, infection control protocols, and liability concerns, making this a hard-barrier task reserved for trained humans.
Cost vs. human wageclaude-haiku-4-5-202510011/5Specialized robotic systems for phlebotomy are experimental, expensive, and not cost-competitive with trained nurses; AI offers no material cost advantage for this physical task.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute for physical specimen collection, so cost comparison favors the human nurse entirely; any robotic alternative would be far more expensive and unproven.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed clinical system performs routine specimen collection autonomously; this remains firmly a human nurse task in all production healthcare settings.
Technical feasibility todayclaude-sonnet-51/5No deployed AI or robotic product autonomously performs specimen collection from acute care patients in production settings; this remains a manual clinical task.

Treat wounds or superficial lacerations.

0

CI 0–0 · exposure 0 · augmentation 25 · 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 automation due to regulatory constraints, clinical conservatism, and the requirement for human accountability; wound care automation is not on the adoption horizon in any mainstream healthcare setting.
Sector adoption velocityclaude-sonnet-51/5Direct physical patient care tasks in acute care settings show minimal AI adoption; healthcare hands-on procedures remain a laggard area for automation.
Augmentation potentialclaude-haiku-4-5-202510012/5AI can support wound assessment through imaging analysis and documentation, but meaningful augmentation is limited because the nurse must execute treatment manually and retain full clinical judgment; AI tools are peripheral rather than transformative to the core task.
Augmentation potentialclaude-sonnet-52/5AI can assist with documentation, wound photography analysis, or triage decision support, but offers minimal direct assistance to the physical act of treating the wound itself.
Task automatabilityclaude-haiku-4-5-202510011/5Wound treatment requires direct physical manipulation, sterile technique, real-time assessment of tissue condition, and adaptive decision-making that current AI cannot perform end-to-end. No meaningful automation of the core task is feasible without full embodied robotics under constrained conditions.
Task automatabilityclaude-sonnet-51/5Wound treatment requires physical manipulation, assessment of tissue, and hands-on dexterity that current AI systems, lacking embodied robotic capability at scale, cannot perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Wound treatment is a core clinical nursing function governed by scope-of-practice regulations, licensure requirements, and liability law that mandate a licensed healthcare provider perform or directly supervise the intervention. Legal and regulatory barriers are substantial.
Adoption barriersclaude-sonnet-55/5Wound care is a licensed clinical procedure requiring nursing credentials, direct patient contact, and liability accountability, creating hard regulatory and legal barriers to automation.
Cost vs. human wageclaude-haiku-4-5-202510011/5Current AI has no practical cost advantage because it cannot perform the task; any system capable of autonomous wound care would require expensive robotic infrastructure, ongoing oversight, and specialized sensors far exceeding the loaded cost of a nurse.
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute performing this physical task, so the human remains the only cost-effective option; AI cost is essentially infinite/inapplicable.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product reliably performs wound treatment or laceration care autonomously in clinical settings. While wound imaging AI exists for assessment support, it does not meet the bar of performing the treatment task itself.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously treats wounds or lacerations in clinical practice today; this remains a purely manual nursing task.

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.