Psychiatric Technicians

29-2053.00
Median wage $45,130/yr156,960 employed (US)Rank #860 of 923 scored · top 93% by substitution

Care for individuals with mental or emotional conditions or disabilities, following the instructions of physicians or other health practitioners. Monitor patients' physical and emotional well-being and report to medical staff. May participate in rehabilitation and treatment programs, help with personal hygiene, and administer oral or injectable medications.

Sub-scores

0–100 · band = confidence interval from rater disagreement

Substitution12
Exposure10
Augmentation39

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

16 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%9

panel mean rating 1.4/5 → substitution pressure 9/100

Technical feasibility todayw 20%11

panel mean rating 1.5/5 → substitution pressure 11/100

Cost vs. human wagew 15%13

panel mean rating 1.5/5 → substitution pressure 13/100

Adoption barriersw 20%inverted — strong barriers lower the score15

panel mean rating 4.4/5 (barrier strength) → substitution pressure 15/100

Sector adoption velocityw 10%13

panel mean rating 1.5/5 → substitution pressure 13/100

Task breakdown (16 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.

Take and record measures of patients' physical condition, using devices such as thermometers or blood pressure gauges.

39

CI 30–47 · exposure 30 · augmentation 75 · importance 4.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510014/5Healthcare organizations are rapidly deploying connected vital-sign monitors, patient-worn devices, and EHR integration; hospitals and clinics increasingly automate routine measurement logging, though smaller facilities lag and human oversight remains standard.
Sector adoption velocityclaude-sonnet-52/5Healthcare, especially inpatient psychiatric care, has been slow to adopt full automation of hands-on patient monitoring compared to digital/information-sector tasks.
Augmentation potentialclaude-haiku-4-5-202510015/5Automated vital-sign capture with real-time dashboards and alerts substantially augments technician productivity by eliminating manual recording, flagging abnormal trends, and allowing focus on patient comfort and complication assessment rather than administrative entry.
Augmentation potentialclaude-sonnet-53/5Automated/connected devices can streamline recording and flag abnormal readings, assisting technicians in documentation and trend tracking, though the physical measurement interaction with the patient remains human-led.
Task automatabilityclaude-haiku-4-5-202510012/5Physical measurement capture (temperature, blood pressure) can be partially automated with connected medical devices that log data directly, but the task also requires human judgment to select appropriate measurement sites, position patients correctly, and identify measurement anomalies—elements that current AI cannot reliably perform end-to-end without intervention.
Task automatabilityclaude-sonnet-52/5Recording basic vitals could be partly automated with connected devices, but psychiatric technicians must also observe patient behavior/status during vitals collection, requiring in-person presence and clinical judgment not replaceable end-to-end today.
Adoption barriersclaude-haiku-4-5-202510013/5While measurement devices can auto-record, regulatory and clinical standards often require a licensed or supervised technician to attest to proper patient positioning, device calibration, and anomaly escalation, creating a validation bottleneck that slows substitution.
Adoption barriersclaude-sonnet-53/5No strict licensing requirement mandates a psychiatric technician specifically for vitals, but psychiatric settings often require in-person clinical staff for safety and behavioral monitoring during interactions with vulnerable patients.
Cost vs. human wageclaude-haiku-4-5-202510014/5Connected medical devices that auto-log vitals cost hundreds to low thousands upfront but measure thousands of patients over years; integrated wearable and bedside monitoring systems are substantially cheaper per measurement than dedicated technician time once deployed.
Cost vs. human wageclaude-sonnet-52/5Automated BP cuffs and thermometers exist cheaply, but require a human to apply them, supervise unpredictable patients, and interpret results in context, so net cost savings versus a technician's time are modest.
Technical feasibility todayclaude-haiku-4-5-202510013/5Automated vital-sign monitors and connected devices exist in production healthcare settings and transmit readings to EHRs, but integration is inconsistent across facilities, and interpretation of borderline readings still typically requires human review before clinical documentation.
Technical feasibility todayclaude-sonnet-52/5Automated vital sign monitors and EHR integration exist and are deployed in some settings, but full automation of taking measurements from psychiatric patients (who may resist or be uncooperative) is not a mature deployed product.

Issue medications from dispensary and maintain records in accordance with specified procedures.

28

CI 25–30 · exposure 30 · augmentation 50 · importance 4.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5While large hospital systems and academic medical centers have piloted automated dispensing cabinets (e.g., Pyxus), adoption in psychiatric facilities remains limited and focused on high-volume, low-complexity settings. Broader displacement is slow due to regulatory constraints and the need for human verification.
Sector adoption velocityclaude-sonnet-52/5Healthcare, especially inpatient psychiatric care, has been slower to adopt full automation of clinical/dispensing tasks compared to information-sector work, with automated dispensing systems adopted incrementally over years.
Augmentation potentialclaude-haiku-4-5-202510013/5Pharmacy management software and automated dispensing systems meaningfully assist technicians by reducing manual record-keeping, flagging potential interactions, and organizing inventory, but the human remains central to verification and decision-making. Productivity gains are moderate.
Augmentation potentialclaude-sonnet-53/5Automated dispensing cabinets, barcode verification, and digital record systems meaningfully assist technicians in tracking inventory and reducing errors, though they don't replace the human role.
Task automatabilityclaude-haiku-4-5-202510012/5While medication dispensing involves rule-based logic that could be partially automated (e.g., verifying prescriptions, updating records), the physical handling of medications, patient verification, and the critical safety requirement of human judgment in anomaly detection make full end-to-end automation infeasible today. Current systems cannot achieve 50% time savings at equal safety quality.
Task automatabilityclaude-sonnet-52/5Automated dispensing cabinets and record-keeping software already handle parts of this, but the physical dispensing to patients, verification of identity, and clinical judgment in a psychiatric setting still require a human technician on-site.ate
Adoption barriersclaude-haiku-4-5-202510014/5Strict regulatory requirements (DEA, state pharmacy boards, hospital accreditation) mandate that licensed pharmacists and technicians physically verify and handle controlled substances; liability for medication errors is severe and typically requires human accountability. Legal and clinical governance create strong substitution barriers.
Adoption barriersclaude-sonnet-54/5Medication dispensing in psychiatric/healthcare settings is tightly regulated, requires licensed or certified personnel, and carries high liability for errors, creating strong legal and safety barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5Medication dispensing robots and integrated pharmacy management systems are capital-intensive and require ongoing maintenance, staff training, and compliance integration. The all-in cost per dose remains comparable to or higher than a technician's loaded wage in most organizations.
Cost vs. human wageclaude-sonnet-52/5Automated dispensing hardware and software carry significant capital and integration costs, and human oversight remains mandatory, so savings versus a technician's wage are moderate rather than order-of-magnitude.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed production systems fully automate medication dispensing from a pharmacy by psychiatric technicians; robotics exist in some hospital pharmacies but are narrowly scoped and require significant oversight. Clinical decision support and record-management software exist but do not replace the technician's role at scale.
Technical feasibility todayclaude-sonnet-53/5Automated medication dispensing systems (e.g., Pyxis) and electronic medication administration records are mature and widely deployed in healthcare, but full end-to-end automation of psychiatric medication issuance including patient interaction is not deployed.

Contact patients' relatives to arrange family conferences.

28

CI 25–30 · exposure 25 · augmentation 50 · importance 3.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Healthcare adoption of AI for direct patient/family communication remains cautious and pilot-stage. Psychiatric settings in particular prioritize human relationships; automation in this domain is lagging compared to other sectors.
Sector adoption velocityclaude-sonnet-52/5Healthcare, especially psychiatric and behavioral health settings, has historically been slower to adopt AI-driven communication tools due to privacy, liability, and interpersonal sensitivity concerns.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can usefully assist by identifying and organizing contact information, suggesting optimal times based on availability patterns, and drafting initial communication templates, leaving the technician to personalize and handle the actual conversation.
Augmentation potentialclaude-sonnet-53/5AI can help draft communication scripts, manage scheduling logistics, and send reminders, meaningfully assisting technicians while they retain responsibility for actual family interactions.
Task automatabilityclaude-haiku-4-5-202510012/5While AI could draft outreach messages or identify contact information, the task requires sensitive interpersonal judgment (responding to emotional contexts, scheduling around family constraints, handling refusals) that current systems cannot reliably perform end-to-end. Automation would require significant human oversight and re-engagement.
Task automatabilityclaude-sonnet-52/5Scheduling contact can be partly automated (e.g., calendar tools, automated reminders), but the interpersonal, sensitive nature of coordinating family conferences about a psychiatric patient requires human judgment and empathy that current AI cannot fully replace end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Healthcare settings face regulatory requirements around patient communication and family engagement; there is strong clinical and organizational preference that sensitive family coordination be handled by trained human staff who can navigate emotional and therapeutic dimensions appropriately.
Adoption barriersclaude-sonnet-53/5No strict licensing requirement for this specific contact task, but organizational policies, patient privacy (HIPAA) concerns, and the sensitive nature of psychiatric care create meaningful friction against full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI-assisted contact outreach (email, message templates) has low per-task cost, but comprehensive automation would require human review and correction loops that approach or exceed the cost of a technician making direct contact.
Cost vs. human wageclaude-sonnet-52/5Basic scheduling automation is cheap, but the need for careful, empathetic communication with families of psychiatric patients requires human involvement, limiting cost savings compared to a fully human-staffed process.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed product reliably handles the full task autonomously. AI can assist with contact discovery and message generation, but the interpersonal negotiation, handling objections, and determining appropriate timing require human intervention in current systems.
Technical feasibility todayclaude-sonnet-52/5Scheduling assistants and automated calling/messaging systems exist and are deployed in healthcare settings, but they are not reliably used for sensitive family communications involving psychiatric patients without human oversight.

Interview new patients to complete admission forms, to assess their mental health status, or to obtain their mental health and treatment history.

27

CI 25–29 · exposure 25 · augmentation 50 · importance 4.1/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Adoption of AI-driven psychiatric intake is slow outside pilot programs and research settings. Most psychiatric facilities still rely on technicians for interviews due to liability concerns, patient preference for human contact, and regulatory uncertainty around autonomous assessment systems.
Sector adoption velocityclaude-sonnet-52/5Healthcare, especially psychiatric/behavioral health, has historically been slow to adopt AI for direct patient assessment due to safety, liability, and regulatory concerns, with pilots more common than production use.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist technicians by auto-generating question prompts, flagging high-risk keywords in real-time, and auto-populating routine demographic and history fields, meaningfully reducing documentation burden while keeping the clinician accountable for the assessment itself.
Augmentation potentialclaude-sonnet-53/5AI can assist with pre-interview questionnaires, transcription, and summarization of patient history, meaningfully speeding up documentation while the clinician retains the interview and assessment role.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can generate draft questions and parse responses into structured forms, reliable mental health assessment requires dynamic clinical judgment, rapport-building, and real-time adaptation to patient distress or safety concerns that current systems cannot fully replicate. End-to-end automation with 50% time savings at equal quality is not achievable today.
Task automatabilityclaude-sonnet-52/5Parts of admission form completion could be automated, but the core requires live clinical interviewing, rapport-building, and real-time risk assessment (e.g., suicidality) that current AI cannot reliably perform end-to-end at equal quality.
Adoption barriersclaude-haiku-4-5-202510014/5Mental health intake is a clinical and legal responsibility; most healthcare systems require a licensed clinician or certified technician to conduct and sign off on assessments for clinical and liability reasons. Regulations around patient privacy, informed consent, and duty to warn create significant barriers to full automation.
Adoption barriersclaude-sonnet-54/5Clinical assessment of mental health status and risk typically requires a credentialed professional's judgment and liability accountability, creating strong regulatory and licensing barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510013/5Automated intake infrastructure (chatbots, form-filling systems) can reduce per-interview labor costs, but integration with EHRs, compliance requirements, and necessary human validation offset savings, making costs roughly comparable to a technician on a per-task basis.
Cost vs. human wageclaude-sonnet-52/5AI transcription/intake-form tools are cheap, but because a licensed clinician must still verify mental status and safety risk, overall cost savings are modest rather than order-of-magnitude.
Technical feasibility todayclaude-haiku-4-5-202510012/5Chatbot-based intake systems exist in pilot form at some healthcare organizations, but they remain narrow in scope, often fail to detect crisis indicators, and require heavy human review and correction. No production system reliably conducts psychiatric interviews without significant clinician oversight.
Technical feasibility todayclaude-sonnet-52/5Some AI chatbots and structured intake tools exist for basic history-taking, but no deployed product reliably conducts full psychiatric admission interviews including mental status exams in production clinical settings.

Train or instruct new employees on procedures to follow with psychiatric patients.

21

CI 16–25 · exposure 17 · augmentation 50 · importance 3.9/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Healthcare adoption of AI-led training remains limited; most organizations rely on human trainers and structured onboarding programs with regulatory oversight. While some healthcare systems use e-learning supplements, core psychiatric technician training has not shifted to AI-driven models at scale.
Sector adoption velocityclaude-sonnet-52/5Healthcare training, especially in psychiatric/behavioral health settings, has been slow to adopt AI-driven instruction due to safety concerns, physical/hands-on components, and regulatory environment.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can usefully assist by drafting training scenarios, generating reference materials, creating quizzes, and organizing case studies, which would improve trainer productivity and consistency. However, the core instruction and mentoring role remains human-centered, limiting augmentation impact.
Augmentation potentialclaude-sonnet-53/5AI can assist by generating training manuals, quizzes, scenario simulations, and onboarding checklists, meaningfully supporting trainers while humans still lead hands-on instruction.
Task automatabilityclaude-haiku-4-5-202510011/5This task involves teaching judgment-laden interpersonal procedures to humans, requiring contextual awareness, individual adaptation, and real-time responsiveness that current AI cannot replicate end-to-end. Training on psychiatric patient procedures requires understanding nuance, reading trainees, and handling edge cases that demand human expertise.
Task automatabilityclaude-sonnet-52/5Training new employees on psychiatric patient procedures requires experiential demonstration, hands-on supervision, and adaptive judgment about real patients that current AI cannot deliver end-to-end, though it can support materials creation.
Adoption barriersclaude-haiku-4-5-202510014/5Healthcare organizations have strong licensing and regulatory requirements for trainers, patient safety liability concerns, and accreditation standards that typically mandate qualified human instruction for psychiatric technician training. Organizational culture and legal responsibility create meaningful friction against full automation.
Adoption barriersclaude-sonnet-54/5Healthcare facilities typically require qualified, experienced staff to certify competency in patient-handling procedures, and liability/regulatory concerns around psychiatric care create strong barriers to full AI substitution.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI could help create training content at low cost, but the human trainer remains necessary for delivery, clarification, assessment, and relationship-building, so all-in cost savings are modest. A hybrid model might achieve 20–30% cost reduction, not the order-of-magnitude difference required for a 5.
Cost vs. human wageclaude-sonnet-52/5AI can cheaply generate training materials, but the core supervised, in-person instruction and competency verification still require paid human trainers, keeping overall cost comparable to human-led training.
Technical feasibility todayclaude-haiku-4-5-202510012/5AI can assist with generating training materials or scripting scenarios, but no deployed product reliably performs end-to-end training instruction of new psychiatric technicians in production settings. Existing e-learning platforms are narrow and typically supplement rather than replace human instruction.
Technical feasibility todayclaude-sonnet-52/5Some e-learning and AI-assisted training content tools exist, but no deployed product independently trains staff on hands-on psychiatric patient care procedures in production settings.

Develop or teach strategies to promote client wellness and independence.

18

CI 11–25 · exposure 13 · augmentation 63 · importance 4.0/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Healthcare organizations, particularly psychiatric settings, adopt new clinical technologies slowly due to regulatory burden, liability concerns, and the critical importance of human presence in therapeutic contexts; pilots are rare.
Sector adoption velocityclaude-sonnet-52/5Healthcare and behavioral health sectors adopt AI slowly for direct patient-facing clinical judgment tasks, with most current use limited to administrative or documentation support.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist by drafting educational materials, suggesting evidence-based strategies, or organizing client data to inform human-led teaching, modestly improving technician productivity while the technician remains responsible for delivery and adaptation.
Augmentation potentialclaude-sonnet-54/5AI can help technicians draft psychoeducational materials, suggest evidence-based strategies, and summarize best practices, meaningfully supporting the human in this task.
Task automatabilityclaude-haiku-4-5-202510011/5Developing and teaching wellness strategies requires deep understanding of individual client needs, therapeutic relationship-building, and adaptive instruction based on client response—capabilities that current AI systems cannot reliably perform end-to-end with quality parity to human practitioners.
Task automatabilityclaude-sonnet-52/5Developing personalized wellness/independence strategies requires clinical judgment, relationship-building, and adaptation to individual patient conditions that AI cannot reliably perform end-to-end today.
Adoption barriersclaude-haiku-4-5-202510014/5Psychiatric care is heavily regulated; technicians must be licensed, and teaching therapeutic strategies requires documented human accountability and clinical judgment—liability and regulatory requirements prevent full automation.
Adoption barriersclaude-sonnet-54/5Mental health treatment planning is subject to licensing, clinical supervision, and liability concerns, requiring qualified professionals to develop and validate care strategies.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI-generated wellness content is cheap, but integrating it into clinical care, personalizing it, and overseeing its use by qualified technicians involves significant manual review costs, keeping the all-in ratio unfavorable compared to direct technician labor.
Cost vs. human wageclaude-sonnet-52/5AI could cheaply generate generic wellness content, but the human oversight, assessment, and delivery needed to make it clinically valid keeps overall cost comparable to human-led work.
Technical feasibility todayclaude-haiku-4-5-202510012/5While AI can generate generic wellness resources or educational materials, no deployed product reliably assesses individual psychiatric clients and teaches personalized strategies in a clinical setting; human oversight remains essential for safety and appropriateness.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously designs or delivers individualized behavioral health independence strategies in production settings; this remains a human clinical function.

Monitor patients' physical and emotional well-being and report unusual behavior or physical ailments to medical staff.

16

CI 7–25 · exposure 13 · augmentation 50 · importance 4.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Healthcare adoption of autonomous patient monitoring remains slow despite decades of opportunity; regulatory caution, liability concerns, and entrenched staffing models limit deployment. Pilot programs are common but production-scale replacement of observational duties is rare.
Sector adoption velocityclaude-sonnet-52/5Healthcare, especially inpatient psychiatric care, is a slower-adopting, high-touch physical sector with limited AI deployment for direct patient monitoring.
Augmentation potentialclaude-haiku-4-5-202510013/5AI tools can assist by flagging vital-sign anomalies, alerting to pattern changes, and organizing patient data for review, meaningfully raising technician efficiency in triage and documentation. However, the augmentation is partial—the human must remain the primary observer and clinical decision-maker.
Augmentation potentialclaude-sonnet-53/5AI-enabled sensors, wearables, and alert systems can help flag anomalies (e.g., vital sign changes) to assist technicians, though the core observational and emotional-assessment work remains human-driven.
Task automatabilityclaude-haiku-4-5-202510012/5While AI systems can assist with automated vital sign monitoring and flagging anomalies, the task fundamentally requires real-time observation of emotional states, behavioral nuances, and clinical judgment that are context-dependent. Current AI cannot reliably detect subtle behavioral changes or assess emotional well-being in person without human presence, limiting automation to <50% of meaningful work.
Task automatabilityclaude-sonnet-51/5Direct observation of patients requires physical presence, sensory judgment, and real-time interpretation of behavioral cues that current AI cannot perform end-to-end in a care setting.
Adoption barriersclaude-haiku-4-5-202510014/5Strong regulatory barriers exist: psychiatric technicians are often licensed professionals, and direct patient observation/monitoring is frequently a legal/licensure requirement in healthcare settings. Liability for missed warning signs and the requirement for human clinical judgment create significant adoption friction.
Adoption barriersclaude-sonnet-54/5Patient safety, liability for missed symptoms, and regulatory/clinical oversight requirements create strong barriers to full automation of monitoring vulnerable psychiatric patients.
Cost vs. human wageclaude-haiku-4-5-202510012/5Monitoring infrastructure (sensors, systems) plus human oversight for behavioral assessment means total costs approach or exceed a technician's loaded wage. The safety-critical nature necessitates substantial human validation, eroding cost advantage.
Cost vs. human wageclaude-sonnet-51/5Given the lack of a viable autonomous product, there is no functioning AI cost baseline cheaper than the human technician's wage for this task.
Technical feasibility todayclaude-haiku-4-5-202510012/5Deployed products exist for vital-sign monitoring and basic anomaly detection (wearables, telemetry systems), but reliable end-to-end assessment of emotional well-being and behavioral observation in clinical settings remains immature. Production systems focus on narrow metrics rather than the holistic patient monitoring this task requires.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously monitors psychiatric patients and reports findings to staff; some passive sensor/camera monitoring exists in research or pilot forms but not as a replacement for this task.

Observe and influence patients' behavior, communicating and interacting with them and teaching, counseling, or befriending them.

13

CI 0–25 · exposure 13 · augmentation 38 · importance 4.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Mental health and psychiatric care remain heavily regulated, labor-intensive, and bound to human contact requirements. Adoption of AI for behavioral observation and counseling is limited to pilots and supplementary chatbots in information-rich institutions; displacement remains minimal.
Sector adoption velocityclaude-sonnet-51/5Direct behavioral healthcare work in institutional settings is a low-digitization, high-touch sector with minimal AI agent deployment for hands-on patient interaction.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist by summarizing patient notes, generating psychoeducational content, or flagging behavioral patterns for review, thereby reducing documentation burden and supporting clinical decision-making. However, the core task of observation and therapeutic engagement remains human-centered.
Augmentation potentialclaude-sonnet-52/5AI tools can support documentation, training materials, or decision-support for treatment planning, but they offer little direct assistance to the moment-to-moment task of observing and interacting with patients.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can generate counseling-like text and respond to patient input, genuinely observing behavior nuance, building therapeutic rapport, and adapting emotional support requires human presence and continuity that current AI cannot reliably achieve. At best, AI assists with scripted psychoeducation or documentation, not the core therapeutic relationship.
Task automatabilityclaude-sonnet-51/5This task requires sustained physical presence, real-time behavioral observation, and building therapeutic rapport with vulnerable psychiatric patients, none of which current AI can perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Licensing and legal requirements bind psychiatric technicians to states; liability for adverse mental health outcomes is severe and high-cost; regulatory frameworks (e.g., FDA, state mental health boards) increasingly scrutinize autonomous AI in psychiatric settings. A licensed clinician must supervise and remain accountable.
Adoption barriersclaude-sonnet-55/5Direct patient care in psychiatric settings involves safety-critical judgment, legal duty of care, and often requires credentialed staff physically present, creating strong regulatory and liability barriers to automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI-driven chatbots have low marginal cost, but psychiatric technician oversight, training refinement, liability management, and human backup remain necessary costs. Total system cost is not yet substantially cheaper than staffing the human role.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI product performing this task, so any cost comparison favors the human by default since the AI alternative doesn't functionally exist.
Technical feasibility todayclaude-haiku-4-5-202510012/5Chatbots exist for mental health support, but they are deployed as supplements, not replacements for psychiatric technician observation and intervention. No production system reliably performs the full task of behavioral observation, risk assessment, and therapeutic relationship-building that defines the role.
Technical feasibility todayclaude-sonnet-51/5No deployed product observes and behaviorally influences psychiatric patients in person; AI chatbots for mental health support exist but are narrow adjuncts, not substitutes for hands-on technician interaction.

Encourage patients to develop work skills and to participate in social, recreational, or other therapeutic activities that enhance interpersonal skills or develop social relationships.

3

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

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Psychiatric facilities remain among the lowest-adoption sectors for automation of patient-facing clinical work. Human therapeutic presence is fundamental to mental health practice, and organizational/regulatory barriers prevent substitution with AI systems.
Sector adoption velocityclaude-sonnet-51/5Psychiatric/behavioral health care is a low-digitization, high-touch sector with minimal AI adoption for direct patient engagement tasks.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could potentially assist with scheduling activities, suggesting evidence-based recreational options, or tracking patient participation metrics, but these are peripheral to the core task of interpersonal encouragement and relationship-building, which depends entirely on human interaction.
Augmentation potentialclaude-sonnet-52/5AI could help technicians plan activities, track progress notes, or suggest therapeutic exercises, but it offers limited direct assistance during the interpersonal encouragement itself.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires genuine human interpersonal engagement, emotional attunement, and the ability to motivate and build therapeutic relationships. Current AI cannot replace the human presence, trust-building, and contextual judgment needed to encourage vulnerable patients in real social or recreational settings.
Task automatabilityclaude-sonnet-51/5This requires building trust, in-person motivation, and adaptive interpersonal engagement with psychiatric patients that current AI cannot replicate end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5This task requires direct human contact and therapeutic presence mandated by clinical standards and licensing requirements. Psychiatric care is heavily regulated, and a licensed human technician must be present to engage patients in therapeutic activities; automation is legally and ethically prohibited.
Adoption barriersclaude-sonnet-54/5Patient safety, clinical oversight requirements, and the need for trained staff in psychiatric care settings create strong institutional and regulatory barriers to substitution.
Cost vs. human wageclaude-haiku-4-5-202510011/5AI systems capable of meaningful interpersonal engagement at clinical-grade quality remain speculative and expensive to develop; the loaded cost of such systems would far exceed the hourly wage of a psychiatric technician who performs this work directly.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing this task, so any comparison favors the human technician who provides necessary in-person engagement and supervision.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product reliably performs this task end-to-end. While AI chatbots can offer generic encouragement, they cannot substitute for the human therapeutic alliance, presence, and judgment that psychiatric technicians provide in actual clinical settings with real patients.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously motivates psychiatric patients toward social/vocational participation; this remains a human relational task in clinical settings.

Aid patients in performing tasks, such as bathing or keeping beds, clothing, or living areas clean.

3

CI 0–5 · exposure 0 · augmentation 25 · importance 4.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Psychiatric and healthcare facilities remain among the slowest adopters of automation for direct patient care due to regulatory constraints, union protections, and organizational resistance to replacing hands-on caregiving roles. Current adoption is negligible in this specific function.
Sector adoption velocityclaude-sonnet-51/5Direct care and physical assistance in healthcare/psychiatric settings is a low-digitization, physically embodied task with minimal AI/robotic adoption in production.
Augmentation potentialclaude-haiku-4-5-202510012/5AI offers minimal assistance for hands-on personal care tasks. While monitoring systems or scheduling tools might marginally support workflow, they do not enhance the core task of physically aiding patients in bathing or hygiene, which remains fundamentally human-dependent.
Augmentation potentialclaude-sonnet-52/5AI could support scheduling, reminders, or documentation around care routines, but offers little direct assistance for the physical act of bathing or cleaning tasks.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires direct physical contact, manual dexterity, and real-time responsiveness to patient needs and safety concerns. Current AI systems, including robotics, cannot reliably perform intimate personal care tasks like bathing or clothing patients at the standard required for patient safety and dignity.
Task automatabilityclaude-sonnet-51/5This is hands-on physical assistance with bathing and hygiene requiring direct physical manipulation and safety monitoring of psychiatric patients, which current AI systems cannot perform.
Adoption barriersclaude-haiku-4-5-202510015/5Healthcare regulations, patient privacy laws (HIPAA), liability standards, and accreditation requirements mandate human presence and accountability in personal care. Additionally, the intimate nature of these tasks creates strong legal and institutional barriers to automation, and many patients require human contact as part of therapeutic treatment.
Adoption barriersclaude-sonnet-54/5Direct physical care of vulnerable psychiatric patients involves safety, liability, and often licensing/certification requirements, plus strong preference and need for human contact and judgment.
Cost vs. human wageclaude-haiku-4-5-202510011/5Specialized care robotics capable of safe patient handling are extremely expensive to acquire and maintain, far exceeding the loaded wage of a psychiatric technician. Integration, training, and safety oversight add substantial costs that make AI economically uncompetitive for this task today.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute for physical caregiving, so any hypothetical robotic solution would be vastly more expensive and less capable than a human aide.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed products reliably perform intimate personal care assistance tasks in production clinical settings. While research exists in care robotics, practical systems lack the dexterity, safety assurance, and regulatory clearance to substitute for human psychiatric technicians in hands-on patient care.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product or robot reliably performs physical patient hygiene assistance in psychiatric care settings today; this remains far beyond current robotics capability.

Provide nursing, psychiatric, or personal care to patients with cognitive, intellectual, or developmental disabilities.

0

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

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Healthcare facilities, especially those serving cognitively or developmentally disabled populations, operate under strict regulatory frameworks and have adopted AI primarily for documentation and scheduling, not frontline caregiving. Adoption of AI for direct patient care is minimal.
Sector adoption velocityclaude-sonnet-51/5Healthcare direct-care settings, especially psychiatric and developmental disability care, show minimal AI adoption for hands-on tasks due to physical, regulatory, and safety constraints.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could assist with documentation, activity logging, or behavioral pattern alerts, but has limited augmentative value for the core tasks of hands-on care, de-escalation, and personal support that define the role.
Augmentation potentialclaude-sonnet-52/5AI can support documentation, scheduling, or behavioral pattern tracking, but offers minimal direct assistance during the hands-on caregiving itself.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires direct physical care (bathing, toileting, feeding), emotional support, and real-time behavioral assessment of vulnerable populations. Current AI systems cannot perform end-to-end hands-on caregiving or provide the safety-critical, person-centered human presence essential to this work.
Task automatabilityclaude-sonnet-51/5This is hands-on physical and interpersonal caregiving requiring direct human presence, empathy, and physical intervention that current AI cannot perform.
Adoption barriersclaude-haiku-4-5-202510015/5Healthcare regulations, duty-of-care requirements, and licensing standards legally mandate qualified human oversight for psychiatric and personal care of vulnerable populations. Liability, patient safety, and informed consent requirements create hard barriers to automation.
Adoption barriersclaude-sonnet-55/5Direct patient care of vulnerable populations requires licensed, trained personnel under strict regulatory, ethical, and liability frameworks, with mandated human oversight and hands-on intervention.
Cost vs. human wageclaude-haiku-4-5-202510011/5Psychiatric technicians cost ~$35k–45k annually with benefits; delivering equivalent hands-on care, safety oversight, and emotional attunement through AI would require expensive robotics, infrastructure, and continuous human oversight, making total cost far exceed human labor.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing this physical caregiving task, so cost comparison favors the human by default since no AI alternative exists at scale.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI product performs direct psychiatric or personal care to vulnerable populations. While monitoring and alert systems exist, they support rather than replace the core caregiving function, which remains entirely human-dependent in practice.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product provides direct nursing or personal care to psychiatric/developmental disability patients; this remains entirely research-stage or nonexistent for physical care delivery.

Administer oral medications or hypodermic injections, following physician's prescriptions and hospital procedures.

0

CI 0–0 · exposure 0 · augmentation 25 · importance 4.6/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Healthcare adoption of AI in medication administration remains minimal; clinical settings continue to rely on human technicians, and regulatory bodies have not approved autonomous systems for this task. Adoption velocity is essentially zero.
Sector adoption velocityclaude-sonnet-51/5Healthcare direct patient care, especially psychiatric inpatient settings, has very low AI adoption for hands-on clinical tasks like medication administration.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could marginally assist by verifying medication orders against prescriptions or flagging contraindications, but the core task—physically delivering medication safely—remains entirely human-dependent. Augmentation is limited.
Augmentation potentialclaude-sonnet-52/5AI can support scheduling, medication reminders, or documentation of administration, but offers minimal assistance to the physical act of administering medication itself.
Task automatabilityclaude-haiku-4-5-202510011/5Administering medications and injections requires direct physical contact with patients, real-time clinical judgment to detect adverse reactions, and legal accountability tied to a licensed human. Current AI cannot perform the hands-on injection or oral medication delivery, nor navigate the unpredictable clinical environment needed.
Task automatabilityclaude-sonnet-51/5Administering medications and injections requires physical manipulation, patient assessment, and hands-on care that current AI systems cannot perform; no robotic or software system can substitute for this physical act.
Adoption barriersclaude-haiku-4-5-202510015/5Medication administration is legally restricted to licensed healthcare professionals (psychiatric technicians, nurses, physicians) under state pharmacy and medical practice laws. Patient safety, liability, and regulatory oversight create hard barriers to substitution.
Adoption barriersclaude-sonnet-55/5Administering medications and injections is a licensed clinical act requiring certified personnel under medical supervision, with strict legal, safety, and liability requirements preventing automation.
Cost vs. human wageclaude-haiku-4-5-202510011/5The cost of automating this task (specialized robotics, integration, clinical validation, liability insurance) would vastly exceed the loaded wage of a psychiatric technician. No current AI-based solution is cost-competitive.
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute performing this physical task, so cost comparison is moot—human labor is the only option and thus AI is not cheaper by definition.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI system performs medication administration or injection delivery in production healthcare settings. This task explicitly requires human presence, licensing, and physical capability that no current robotics or AI agent can reliably replicate in a hospital context.
Technical feasibility todayclaude-sonnet-51/5No deployed product administers oral medications or injections to psychiatric patients; this remains outside the scope of AI products, which are limited to information/documentation support.

Restrain violent, potentially violent, or suicidal patients by verbal or physical means as required.

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, especially psychiatric facilities, is a laggard sector for workforce automation due to liability, regulation, and the safety-critical nature of restraint. No evidence of production adoption of AI for physical patient restraint exists.
Sector adoption velocityclaude-sonnet-51/5Healthcare direct-care physical intervention tasks show essentially no AI adoption; this is a laggard, high-touch physical safety task.
Augmentation potentialclaude-haiku-4-5-202510012/5AI can assist in risk flagging and predictive analytics to alert staff, but cannot meaningfully augment the core task of physical or verbal de-escalation. Assistance is limited to pre-incident support, not the restraint event itself.
Augmentation potentialclaude-sonnet-52/5AI could assist with de-escalation scripts, risk prediction, or documentation support, but offers minimal help with the core verbal/physical restraint act itself.
Task automatabilityclaude-haiku-4-5-202510011/5This task fundamentally requires real-time physical intervention, situational judgment under acute stress, and human presence to manage immediate danger. No AI system can physically restrain patients or reliably predict and de-escalate violent behavior in the moment without human agency.
Task automatabilityclaude-sonnet-51/5Physically restraining a violent or suicidal patient requires real-world physical presence, judgment, and human touch that no current AI system can provide.'
Adoption barriersclaude-haiku-4-5-202510015/5This task has hard legal and liability barriers: psychiatric technicians are credentialed health professionals, restraint carries liability and injury risk, and standards of care require a licensed human to assess, decide, and execute physical intervention. Regulatory and duty-of-care requirements prevent automation.
Adoption barriersclaude-sonnet-55/5This task involves legal, clinical, and safety-critical human judgment with licensing and liability requirements; physical intervention must be performed by trained, authorized staff.
Cost vs. human wageclaude-haiku-4-5-202510011/5Physical restraint and immediate crisis response require human presence on-site; AI cannot substitute for staffing costs. Any AI augmentation would add expense (monitoring, alerting) rather than reduce the cost of the human worker required to execute the core task.
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute cost to compare; the human must perform this physically, making AI infinitely more expensive/impossible for the physical component.
Technical feasibility todayclaude-haiku-4-5-202510011/5While AI can assist in risk assessment via data analysis, no deployed product performs autonomous physical restraint or reliable real-time violence prevention. Current systems lack embodied presence, legal authority, and the adaptive judgment needed for safe execution in psychiatric settings.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product performs physical restraint or crisis de-escalation of dangerous patients; this remains entirely a human task.

Lead prescribed individual or group therapy sessions as part of specific therapeutic procedures.

0

CI 0–0 · exposure 0 · augmentation 38 · importance 4.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Mental health care sectors adopt AI very slowly for clinical roles due to regulatory, liability, and patient-preference barriers. Some digital tools assist therapists, but autonomous session leadership is not adopted in production mental health systems.
Sector adoption velocityclaude-sonnet-51/5Healthcare and psychiatric care settings are slow to adopt autonomous AI for direct clinical therapeutic delivery due to regulatory, ethical, and safety constraints.
Augmentation potentialclaude-haiku-4-5-202510012/5AI can assist therapists with administrative tasks, session note generation, or between-session support, but offers limited augmentation to the core act of leading therapy sessions, which depends on licensed clinical judgment and real-time human interaction.
Augmentation potentialclaude-sonnet-53/5AI can assist by helping prepare session materials, track patient progress notes, or suggest therapeutic techniques, but the technician remains fully in control of leading the session.
Task automatabilityclaude-haiku-4-5-202510011/5Leading therapy sessions requires real-time human emotional attunement, therapeutic alliance-building, and clinical judgment that current AI cannot reliably provide. AI cannot substitute for the therapist-client relationship that is itself therapeutic and central to efficacy.
Task automatabilityclaude-sonnet-51/5Leading therapy sessions requires real-time human judgment, empathy, crisis management, and adaptive interpersonal skills that current AI cannot replicate end-to-end for vulnerable psychiatric populations.'
Adoption barriersclaude-haiku-4-5-202510015/5Therapy leadership is legally restricted to licensed mental health professionals (therapists, counselors, psychiatrists) in virtually all jurisdictions. Regulatory and liability barriers are absolute: an unlicensed AI cannot ethically or legally conduct therapy.
Adoption barriersclaude-sonnet-55/5This task requires a credentialed professional to conduct therapy, with legal, licensing, and liability requirements that mandate human oversight and accountability in clinical settings.
Cost vs. human wageclaude-haiku-4-5-202510011/5Therapy requires licensed or credentialed human staff; AI systems for mental health support remain experimental or supplementary, with no cost advantage for autonomous session leadership. Human therapists retain significant cost advantage because they provide legally and clinically required oversight.
Cost vs. human wageclaude-sonnet-51/5Because no viable AI substitute exists for leading sessions, there's no comparable AI cost structure that beats a human technician's wage for this specific task.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI system reliably leads therapy sessions independently or at clinical standard. Research chatbots exist for mental health support, but they lack licensure, clinical validation, and cannot handle crisis intervention or therapeutic complexity.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously leads clinical group or individual therapy sessions for psychiatric patients; chatbot tools remain adjunctive and unsupervised use raises safety concerns.

Collaborate with or assist doctors, psychologists, or rehabilitation therapists in working with patients with cognitive, intellectual, or developmental disabilities to treat, rehabilitate, and return patients to the community.

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, and particularly psychiatric care, has been among the slowest sectors to adopt AI for clinical tasks due to regulatory constraints, liability concerns, and the requirement for human judgment and presence. Adoption remains in pilot phases for decision-support tools, not task automation.
Sector adoption velocityclaude-sonnet-51/5Psychiatric technician work occurs in institutional, direct-care healthcare settings with low digitization and slow AI adoption for hands-on patient treatment.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could provide limited assistance with administrative tasks (scheduling, documentation, outcome tracking) or clinical alerts, but offers minimal productivity enhancement for the core clinical and interpersonal work of psychiatric technicians, where human presence and judgment are irreplaceable.
Augmentation potentialclaude-sonnet-52/5AI can help with documentation, care plan drafting, or scheduling support, but offers minimal assistance to the core interpersonal, therapeutic, and physical care activities described.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires direct human-to-patient interaction, clinical judgment, and adaptive response to complex behavioral and therapeutic needs that current AI cannot perform end-to-end. While AI might assist with documentation or scheduling, the core work of collaboration with clinical staff and hands-on patient care lies firmly outside AI automation today.
Task automatabilityclaude-sonnet-51/5This task requires hands-on patient care, physical assistance, and interdisciplinary clinical collaboration that cannot be executed end-to-end by AI systems today.
Adoption barriersclaude-haiku-4-5-202510015/5Strong legal and regulatory barriers exist: psychiatric technicians often require state licensure or certification, and federal/state healthcare regulations mandate human oversight of patient care and treatment. Liability and duty-of-care requirements further protect this role from substitution.
Adoption barriersclaude-sonnet-55/5Direct patient care for cognitive/developmental disabilities involves licensure, safety regulations, liability, and mandated human staffing ratios in clinical/rehabilitation settings.
Cost vs. human wageclaude-haiku-4-5-202510011/5The cost of AI systems capable of any material part of this task (embodied agents, real-time monitoring, clinical decision support) remains far higher than the loaded wage of a psychiatric technician, especially given liability and regulatory overhead.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing this physical, relational care task, so cost comparison favors the human worker entirely.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI system reliably performs the work of a psychiatric technician: direct patient supervision, behavioral de-escalation, physical assistance, and real-time clinical support. Clinical AI tools exist but do not replace or reliably perform this role in production settings.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product performs direct collaborative patient treatment and rehabilitation work in psychiatric/developmental disability settings; this remains squarely human-delivered care.

Escort patients to medical appointments.

0

CI 0–0 · exposure 0 · augmentation 13 · importance 3.9/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5This task is fundamentally tied to in-person care delivery in healthcare settings with low digitization potential for the physical escort component. Adoption velocity for automation is zero.
Sector adoption velocityclaude-sonnet-51/5Healthcare direct-care and physical patient-handling roles show minimal AI adoption; this is a low-digitization, physical-presence-dependent task.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could assist minimally by scheduling optimizations or navigation aids, but the core task—managing a patient during escort—requires human presence and judgment. Augmentation value is limited.
Augmentation potentialclaude-sonnet-51/5AI offers essentially no assistance to the physical act of escorting a patient, though scheduling logistics around it might be aided marginally.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires physical presence, real-time responsiveness to patient behavior, and navigation of potentially unpredictable human interactions. Current AI systems cannot physically escort patients or reliably manage the safety and behavioral contingencies involved.
Task automatabilityclaude-sonnet-51/5This is a physical escort/transport and supervision task requiring physical presence, situational judgment, and safety monitoring of potentially vulnerable patients; no AI system can perform this bodily task.'
Adoption barriersclaude-haiku-4-5-202510015/5Strong legal and regulatory barriers exist: psychiatric technicians must be physically present and accountable for patient safety and welfare during transport. Liability and duty-of-care requirements create hard barriers to automation.
Adoption barriersclaude-sonnet-55/5Direct patient care, safety supervision, and often legal/facility requirements mandate a trained human staff member physically present, especially for psychiatric patients who may need monitoring or de-escalation.
Cost vs. human wageclaude-haiku-4-5-202510011/5AI systems cannot perform this task at all, so cost comparison is not applicable; the ratio is effectively infinite (impossible divided by human wage).
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute providing this physical service, so AI cost is not comparable—human labor is the only option.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI product performs end-to-end physical escorting of psychiatric patients. This remains entirely within the human domain given the need for physical presence, crisis de-escalation capability, and duty of care.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product physically escorts patients; this remains entirely a human physical-presence function.

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.