Recreational Therapists

29-1125.00
Median wage $61,960/yr14,930 employed (US)Rank #724 of 923 scored · top 78% by substitution

Plan, direct, or coordinate medically-approved recreation programs for patients in hospitals, nursing homes, or other institutions. Activities include sports, trips, dramatics, social activities, and crafts. May assess a patient condition and recommend appropriate recreational activity.

Sub-scores

0–100 · band = confidence interval from rater disagreement

Substitution18
Exposure17
Augmentation50

Substitution — the headline: capability discounted by cost, barriers and adoption.

Exposure — technical capability alone, regardless of whether anyone deploys it.

Augmentation — how much AI assists without replacing. High here + moderate substitution = a changing job, not a disappearing one.

Tasks on the substitution scale

11 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%18

panel mean rating 1.7/5 → substitution pressure 18/100

Technical feasibility todayw 20%15

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

Cost vs. human wagew 15%19

panel mean rating 1.8/5 → substitution pressure 19/100

Adoption barriersw 20%inverted — strong barriers lower the score23

panel mean rating 4.1/5 (barrier strength) → substitution pressure 23/100

Sector adoption velocityw 10%16

panel mean rating 1.6/5 → substitution pressure 16/100

Task breakdown (11 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.

Prepare and submit reports and charts to treatment team to reflect patients' reactions and evidence of progress or regression.

34

CI 25–43 · exposure 33 · augmentation 63 · importance 4.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Recreational therapy is a small, niche clinical specialty with low overall digitization and slow EHR adoption relative to medicine or psychiatry. Production deployment of AI for clinical reporting in this setting is rare; pilots are uncommon.
Sector adoption velocityclaude-sonnet-52/5Healthcare, especially allied health and long-term care settings where recreational therapists work, has been slower to adopt AI documentation tools compared to finance or tech sectors.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist by suggesting chart templates, flagging data entry gaps, and auto-generating boilerplate sections, raising therapist efficiency modestly. However, the core clinical judgment—interpreting patient reactions and assessing progress—remains human-led, limiting transformative impact.
Augmentation potentialclaude-sonnet-54/5AI can meaningfully speed up drafting of reports and organizing observations into structured charts, letting the therapist focus on clinical interpretation and finalize the writeup.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can generate charts and draft reports from structured patient data, the task requires synthesizing qualitative clinical observations and subjective behavioral assessments that demand human clinical judgment. Current AI cannot reliably interpret nuanced patient reactions or make defensible clinical progress judgments without substantial human oversight.
Task automatabilityclaude-sonnet-53/5AI can draft progress notes and summarize observations from structured input, but requires clinician-provided qualitative observations and judgment about patient reactions that aren't easily automated end-to-end."
Adoption barriersclaude-haiku-4-5-202510014/5Strong barriers exist: clinical reports must be signed by a licensed therapist and are part of the legal medical record; liability for inaccurate progress documentation is high; regulatory frameworks (HIPAA, state licensure boards) require therapist accountability for clinical assessments, not AI generation.
Adoption barriersclaude-sonnet-53/5Documentation must reflect clinical judgment and is often legally required to be authored or reviewed by the licensed therapist, creating moderate liability and compliance friction.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI report generation tools require significant integration and clinical review overhead, making the all-in cost (inference + integration + oversight by a trained therapist) roughly comparable to or higher than having a therapist draft the report directly, especially given liability and quality requirements.
Cost vs. human wageclaude-sonnet-53/5AI drafting tools can reduce time spent writing reports, but licensing, integration into EHRs, and mandatory clinician review keep costs roughly comparable to a therapist's own documentation time.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed product reliably performs this task end-to-end in clinical settings. EHR systems can auto-generate templated reports, but they require extensive manual clinical input and fact-checking to be safe and compliant; pure AI drafting of clinical progress assessments remains research-stage or pilot-only.
Technical feasibility todayclaude-sonnet-52/5Clinical documentation AI tools exist (e.g., ambient scribes) in some healthcare settings, but they are not widely deployed specifically for recreational therapy progress charting and require human review for accuracy.

Observe, analyze, and record patients' participation, reactions, and progress during treatment sessions, modifying treatment programs as needed.

25

CI 25–25 · exposure 25 · 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 and therapeutic settings lag in AI automation adoption due to regulatory constraints, liability aversion, and the clinical need for human accountability. Recreational therapy, in particular, operates in settings (hospitals, rehabilitation centers) with strong preference for human oversight and low digitization pressure.
Sector adoption velocityclaude-sonnet-52/5Healthcare and allied health therapy settings are generally slower adopters of AI for direct clinical judgment tasks compared to information-sector professions.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can usefully assist by transcribing session notes, flagging progress metrics, and surfacing patterns in historical data for therapist review. However, augmentation is limited to documentation and post-hoc analysis rather than real-time clinical insight or adaptive intervention support.
Augmentation potentialclaude-sonnet-53/5AI can help by transcribing sessions, flagging patterns in patient data, and drafting progress notes, meaningfully supporting but not replacing the therapist's real-time judgment.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist with recording observations and flagging data patterns, the task requires real-time clinical judgment during sessions—observing subtle emotional and behavioral cues, adjusting interventions on-the-fly, and modifying treatment based on nuanced patient responses. Current AI cannot reliably perform this adaptive, judgment-heavy component end-to-end at equal quality.
Task automatabilityclaude-sonnet-52/5Observing and reacting to patients in real time during therapy sessions requires physical presence, human perception, and clinical judgment that current AI cannot replicate end-to-end; only the documentation portion is automatable.“},
Adoption barriersclaude-haiku-4-5-202510014/5Licensing, patient privacy (HIPAA), and liability concerns create substantial friction. Treatment modification decisions are typically part of the licensed therapist's clinical authority, and automated changes could expose organizations to regulatory and malpractice risk. Human accountability is legally and ethically expected.
Adoption barriersclaude-sonnet-54/5Treatment modification and clinical assessment typically require a licensed therapist's judgment and sign-off, creating strong regulatory and liability barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5Integration of computer vision for session monitoring, data processing, and clinical oversight still requires significant infrastructure and human review. The cost per session remains comparable to or higher than documenting therapist time, given deployment and validation overhead.
Cost vs. human wageclaude-sonnet-52/5AI can cheaply assist with documentation, but the core observation and clinical decision-making still requires a skilled therapist, keeping overall cost comparable to human labor.
Technical feasibility todayclaude-haiku-4-5-202510012/5Some video analysis and note-taking tools exist, but no deployed product reliably combines observation, clinical analysis, and real-time treatment modification. Existing systems work on retrospective data or narrow metrics, not live adaptive decision-making in therapeutic contexts.
Technical feasibility todayclaude-sonnet-52/5Some products assist with clinical note-taking and progress tracking, but no deployed system reliably observes patient behavior in-session and autonomously modifies treatment plans.

Obtain information from medical records, medical staff, family members and the patients, themselves, to assess patients' capabilities, needs and interests.

25

CI 25–25 · 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 adoption of AI is moving but cautiously; assessment and intake processes remain heavily human-centered due to regulatory caution, accreditation standards, and the clinical relationship's importance—adoption in therapy settings is slower than in tech/finance.
Sector adoption velocityclaude-sonnet-52/5Healthcare and allied health therapy settings are historically slow AI adopters relative to information/finance sectors, with pilots for documentation assistance more common than assessment automation.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can usefully assist by summarizing medical records, flagging key data points, and organizing patient-reported information, which raises therapist productivity in synthesis and documentation—but the core assessment work remains human-led.
Augmentation potentialclaude-sonnet-54/5AI can meaningfully assist by quickly extracting and summarizing relevant medical record data, drafting intake questions, and organizing information the therapist then uses alongside direct patient interaction.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can extract structured data from medical records and transcribe conversations, the task fundamentally requires synthesizing subjective information about capabilities, needs, and interests—which demands clinical judgment and nuanced human interaction that current systems cannot reliably perform end-to-end with 50% time savings at equal quality.
Task automatabilityclaude-sonnet-52/5This requires synthesizing information from multiple human sources including sensitive in-person patient interaction, clinical judgment, and rapport-building that current AI cannot fully replicate end-to-end.gene AI could assist with record review but not the interpersonal assessment.rasa
Adoption barriersclaude-haiku-4-5-202510014/5Recreational therapists are often required by regulation to directly assess patients, and many clinical settings have liability requirements that a licensed professional personally gather and interpret assessment data; liability asymmetry is high if errors in needs assessment cause harm.
Adoption barriersclaude-sonnet-54/5Recreational therapists are licensed/certified professionals whose clinical assessments are typically required for care planning and reimbursement, and patient contact combined with liability concerns create strong barriers to substitution.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI-assisted data extraction from EHRs is cheap, but the overhead of human oversight, correction, and the need for licensed staff to validate and conduct interviews means the full-task cost remains comparable to or exceeds direct human performance.
Cost vs. human wageclaude-sonnet-52/5AI can cheaply process records, but the human interaction and clinical judgment components still require the therapist's time, so overall cost savings are limited.
Technical feasibility todayclaude-haiku-4-5-202510012/5Medical record extraction tools exist and work reasonably well on structured data, but no deployed system reliably obtains and integrates information from family members and patients themselves with the contextual sensitivity and adaptive interviewing required for therapeutic assessment.
Technical feasibility todayclaude-sonnet-52/5Products exist for summarizing medical records and clinical notes, but no deployed system reliably conducts the interactive, multi-source patient capability assessment itself in production.

Develop discharge plans for patients.

25

CI 25–25 · exposure 25 · augmentation 50 · importance 4.0/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Healthcare remains slower to adopt AI automation for high-stakes clinical tasks; most systems deployed today support human workers rather than replace them. Discharge planning adoption is hampered by regulatory oversight, multi-disciplinary workflows, and institutional caution around liability.
Sector adoption velocityclaude-sonnet-52/5Healthcare and allied health therapy settings are generally slower adopters of AI-driven clinical decision tools compared to information/finance sectors.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can meaningfully assist therapists by drafting templates, suggesting community resources, organizing patient information, and flagging risk factors, raising the efficiency of plan development. However, augmentation is bounded because the core clinical judgment and accountability must remain with the licensed therapist.
Augmentation potentialclaude-sonnet-53/5AI can assist by summarizing patient records, generating draft plan templates, and flagging relevant considerations, improving therapist efficiency while they retain final judgment.
Task automatabilityclaude-haiku-4-5-202510012/5Developing discharge plans requires synthesizing complex patient data, medical history, psychosocial factors, and community resources into a personalized care transition strategy. While AI can help retrieve and organize information, the task fundamentally demands human judgment about patient-specific needs, preferences, and risk factors that are difficult to automate reliably.
Task automatabilityclaude-sonnet-52/5Discharge planning requires synthesizing patient-specific clinical judgment, family circumstances, and psychosocial factors that AI cannot reliably assess end-to-end, though it can help draft documentation.2 Most of the substantive clinical reasoning still requires a human therapist's evaluation.
Adoption barriersclaude-haiku-4-5-202510014/5Healthcare liability, patient safety requirements, and the need for a licensed therapist to authorize discharge plans create substantial legal and regulatory barriers. Documentation and sign-off must come from a qualified human professional accountable for the plan's outcomes.
Adoption barriersclaude-sonnet-54/5Discharge planning is often a licensed clinical responsibility requiring professional sign-off and coordination with care teams, creating regulatory and liability barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI tools that support discharge planning (templates, document drafting) have modest per-task costs, but the human therapist's time remains essential for safety and quality. The cost advantage is marginal because human oversight and decision-making cannot be eliminated.
Cost vs. human wageclaude-sonnet-52/5AI could cheaply draft portions of documentation, but the clinical assessment and coordination work still requires paid therapist time, keeping overall cost savings modest.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed product reliably generates complete, clinically sound discharge plans end-to-end. AI systems exist to assist with documentation and suggest standard resources, but discharge planning requires multidisciplinary input and individual clinical judgment that current products do not perform autonomously at scale.
Technical feasibility todayclaude-sonnet-52/5No deployed product autonomously creates discharge plans for recreational therapy patients today; EHR systems offer templates but not judgment-driven plan generation.

Develop treatment plan to meet needs of patient, based on needs assessment, patient interests, and objectives of therapy.

24

CI 23–25 · exposure 25 · augmentation 63 · importance 4.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Healthcare adoption of AI for clinical planning remains in pilot or research phases; recreational therapy is a smaller, less digitized specialty compared to major tech-adoption sectors, and regulatory and professional norms favor human-authored clinical documents.
Sector adoption velocityclaude-sonnet-52/5Healthcare and allied health fields, including recreational therapy, have historically been slow to adopt AI for clinical planning tasks due to regulatory and workflow constraints.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can usefully assist by organizing patient data, suggesting evidence-based interventions, and drafting sections of plans, enabling therapists to work more efficiently, though the therapist must retain control and judgment over final plan content and therapeutic strategy.
Augmentation potentialclaude-sonnet-54/5AI can meaningfully assist by summarizing assessments, suggesting evidence-based interventions, and drafting plan language for therapist review and finalization.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist in synthesizing patient data and generating template treatment plans, the task fundamentally requires human clinical judgment to integrate complex patient needs, preferences, and therapeutic objectives. Current systems cannot reliably perform the end-to-end assessment and personalized plan development that meets the ≥50% time-saving threshold at equal quality.
Task automatabilityclaude-sonnet-52/5Drafting a template treatment plan can be partially automated, but integrating nuanced patient interests, clinical judgment, and interpersonal assessment requires human expertise that current AI cannot reliably replace end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Treatment-plan development is legally and clinically scoped to licensed healthcare professionals; patient safety liability, regulatory oversight (state licensing, accreditation standards), and organizational requirements that a qualified therapist approve and sign the plan create substantial adoption barriers.
Adoption barriersclaude-sonnet-54/5Treatment plans typically require sign-off by a licensed therapist and are subject to clinical/regulatory documentation standards, creating strong professional and liability barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI infrastructure and oversight costs for clinical plan development are comparable to or exceed the cost of a recreational therapist performing the task, given the need for reliable outputs and professional review in a regulated healthcare setting.
Cost vs. human wageclaude-sonnet-52/5AI could cheaply generate draft text, but the human clinician must still perform assessment, verify accuracy, and take liability, keeping overall cost savings modest.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed products perform this clinical task reliably in production. Some EHR systems offer treatment-plan templates and decision-support tools, but these remain narrow aids rather than systems that can independently develop comprehensive, individualized recreational therapy plans.
Technical feasibility todayclaude-sonnet-52/5Some clinical documentation and care-plan-drafting tools exist, but no deployed product independently creates individualized recreational therapy treatment plans reliably in production.

Instruct patient in activities and techniques, such as sports, dance, music, art, or relaxation techniques, designed to meet their specific physical or psychological needs.

15

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

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Healthcare sectors are relatively cautious adopters of automation for direct patient care; while telehealth and instructional videos exist, actual substitution of therapist-led instruction remains minimal and adoption of AI-driven automation is slow.
Sector adoption velocityclaude-sonnet-51/5Healthcare and physical/rehabilitative therapy sectors show slow AI adoption for hands-on care tasks, with most AI use confined to administrative or documentation support.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist by suggesting tailored activities, generating instructional scripts, recommending music or art techniques, and tracking patient progress—useful enhancements—but the therapist remains essential for demonstration, real-time adjustment, and therapeutic relationship.
Augmentation potentialclaude-sonnet-53/5AI can help therapists plan sessions, suggest activity ideas tailored to patient goals, and track progress data, but it doesn't materially change the in-session instructional interaction itself.
Task automatabilityclaude-haiku-4-5-202510012/5While AI could generate exercise or activity instructions and create personalized activity plans, the task fundamentally requires real-time adaptation to patient response, physical demonstration, safety monitoring, and emotional attunement—elements that demand human presence and cannot achieve 50% time savings end-to-end today.
Task automatabilityclaude-sonnet-51/5This requires live, hands-on, in-person instruction with physical demonstration, adaptive coaching, and real-time emotional/physical monitoring of patients, which current AI cannot perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Healthcare licensing requirements, liability concerns around patient safety during physical activities, regulatory oversight of therapeutic interventions, and clinical settings' strong preference for human therapist presence create substantial adoption friction.
Adoption barriersclaude-sonnet-54/5Recreational therapy often requires licensed/certified therapists (e.g., CTRS), clinical judgment, and physical presence, creating strong professional and liability barriers to automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5Current AI systems capable of personalized instruction with real-time feedback and safety monitoring are expensive relative to the labor cost of a recreational therapist, and human oversight is still required, making the all-in cost unfavorable.
Cost vs. human wageclaude-sonnet-51/5AI cannot substitute for the physical presence and hands-on guidance needed, so there is no viable AI cost comparison—the human remains necessary.
Technical feasibility todayclaude-haiku-4-5-202510012/5Some narrow components exist (AI-generated activity suggestions, video instruction modules), but deployed products cannot reliably perform the full instructional and adaptive aspects of this task in clinical settings with the required safety oversight and therapeutic responsiveness.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product delivers embodied recreational therapy instruction to patients; this remains firmly a human, physically present clinical activity.

Counsel and encourage patients to develop leisure activities.

15

CI 5–25 · exposure 13 · augmentation 50 · importance 4.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Healthcare and therapeutic sectors remain slow to deploy autonomous AI agents, particularly for direct patient interaction requiring emotional judgment and therapeutic presence. Adoption is limited to pilot chat-based suggestions, not production replacement of counseling.
Sector adoption velocityclaude-sonnet-52/5Healthcare and allied therapy professions are historically slow AI adopters for direct patient-facing counseling tasks, with adoption concentrated in administrative and documentation support rather than core therapy delivery.
Augmentation potentialclaude-haiku-4-5-202510013/5AI could usefully assist recreational therapists by suggesting activities tailored to patient interests, generating resource libraries, or helping track patient progress—providing a productivity lift without replacing the core counseling and encouragement work.
Augmentation potentialclaude-sonnet-53/5AI can help therapists brainstorm activity ideas, track patient progress, or draft educational materials, providing moderate support while the human retains the counseling role.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can suggest leisure activities based on patient profiles and interests, meaningful counseling and encouragement requires understanding individual psychology, building rapport, and adapting to emotional cues in real-time—capabilities current systems lack. The task is not easily decomposable into routine automation and falls below the 50% time-saving threshold.
Task automatabilityclaude-sonnet-51/5Counseling and encouraging patients requires genuine therapeutic rapport, emotional attunement, and adaptive judgment based on patient's physical/mental condition that current AI cannot replicate end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Recreational therapy is typically delivered by licensed professionals in clinical or care settings where regulatory frameworks, institutional liability, and a human-contact requirement for therapeutic efficacy create strong adoption barriers. Licensure and duty-of-care obligations protect this role.
Adoption barriersclaude-sonnet-54/5This is a clinical/therapeutic task often requiring licensed professionals, human relationship-building, and accountability for patient wellbeing, creating strong professional and liability barriers to automation.
Cost vs. human wageclaude-haiku-4-5-202510011/5Current AI systems cannot perform this task with quality comparable to a human therapist, so cost comparison is moot. When accounting for the need for human oversight, correction, and the liability of inadequate therapeutic interaction, AI is not cost-effective.
Cost vs. human wageclaude-sonnet-52/5While AI chat tools are cheap per interaction, they cannot substitute for licensed therapeutic counseling, so the effective cost of achieving equivalent quality outcomes is not favorable to AI.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed product reliably performs patient counseling and encouragement; chatbots can provide activity suggestions but cannot replicate the therapeutic relationship, motivational interviewing, or adaptive emotional support that define this task. Systems exist only at research or narrow-scope prototype stages.
Technical feasibility todayclaude-sonnet-51/5No deployed product performs clinical counseling of patients toward leisure activity development as a substitute for a recreational therapist; this remains outside product-level AI capability.

Encourage clients with special needs and circumstances to acquire new skills and get involved in health-promoting leisure activities, such as sports, games, arts and crafts, and gardening.

15

CI 5–25 · exposure 13 · augmentation 38 · importance 4.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Recreational therapy operates mostly in physical healthcare, nonprofit, and community settings with lower digitization and slower tech adoption; while larger medical systems may pilot activity-scheduling software, therapeutic engagement automation remains nascent and low-priority in these sectors.
Sector adoption velocityclaude-sonnet-51/5Healthcare and therapy services for special-needs populations are among the slowest sectors to adopt AI for direct care delivery, given physical and regulatory constraints.
Augmentation potentialclaude-haiku-4-5-202510013/5AI could usefully assist by generating personalized activity recommendations, organizing schedules, or providing evidence-based activity frameworks that therapists review and adapt; such tools enhance planning efficiency without replacing the human's core role in motivation and therapeutic relationship.
Augmentation potentialclaude-sonnet-52/5AI could help therapists plan activities, track client progress, or suggest personalized leisure programs, but it offers limited assistance during the actual encouragement and engagement process.
Task automatabilityclaude-haiku-4-5-202510012/5While AI could help generate activity suggestions or create structured programs, the core task—encouraging clients with special needs through interpersonal engagement, assessing readiness, adapting approaches, and motivating participation—requires sustained human-client interaction, emotional attunement, and real-time behavioral response that current AI cannot reliably provide end-to-end.
Task automatabilityclaude-sonnet-51/5This requires in-person emotional encouragement, motivation, physical demonstration, and adaptive interpersonal engagement with vulnerable clients, none of which current AI can perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Recreational therapy with special-needs clients typically operates in regulated healthcare or social-service settings where credentialing, liability, and duty-of-care requirements mean a licensed or trained human must remain responsible for therapeutic decisions and participant safety; regulatory and institutional barriers are substantial.
Adoption barriersclaude-sonnet-54/5Recreational therapy often requires certified/licensed practitioners, involves duty-of-care and liability for vulnerable populations, and clients typically expect human contact and trust.
Cost vs. human wageclaude-haiku-4-5-202510012/5Integration of AI tools for activity recommendation or session planning would require oversight by the therapist anyway, and the human labor (building rapport, assessing needs, adjusting in real time) remains the cost driver; total cost savings are minimal.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing this physical, relational task, so cost comparison favors the human by default.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed product reliably performs the full encouragement and motivation function for this population; AI chatbots can provide activity ideas or reminders but cannot replace the therapeutic relationship, behavioral assessment, or adaptive scaffolding that a trained therapist delivers in practice.
Technical feasibility todayclaude-sonnet-51/5No deployed product performs in-person motivational therapy or hands-on leisure activity facilitation for special-needs clients; this remains firmly human-delivered.

Confer with members of treatment team to plan and evaluate therapy programs.

13

CI 5–20 · exposure 8 · 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 remains a slow-adopting, heavily regulated sector for clinical decisions. Direct AI displacement of treatment planning conferences is not observed in production environments due to clinical and legal requirements for human judgment and accountability.
Sector adoption velocityclaude-sonnet-52/5Healthcare and allied health therapy settings show slow, cautious AI adoption for clinical decision-making tasks compared to information-sector norms.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could provide marginal support through summarization of prior notes or reminder of protocol checklists, but the core task—real-time clinical discussion, consensus-building, and mutual accountability—is poorly suited to AI assistance in its current form.
Augmentation potentialclaude-sonnet-54/5AI can meaningfully assist by synthesizing patient records, tracking progress metrics, and drafting discussion summaries to prepare therapists for these team conferences.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires collaborative discussion, real-time judgment synthesis, and interpersonal negotiation among clinicians with diverse expertise and stakes in patient care. Current AI cannot meaningfully participate in or lead such multidisciplinary conferences at quality parity.
Task automatabilityclaude-sonnet-52/5This is a collaborative, interpersonal clinical planning task requiring real-time judgment, negotiation, and reading of team dynamics that current AI cannot substitute for end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Regulatory and liability frameworks require qualified human clinicians to participate directly in treatment planning and evaluation; interdisciplinary coordination and shared clinical responsibility are legally and ethically mandated, not merely preferred.
Adoption barriersclaude-sonnet-54/5Treatment planning in clinical/therapeutic settings typically requires licensed professional input and documented accountability, creating strong regulatory and liability barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510011/5AI would add cost through data prep, integration, and oversight rather than replace the human labor involved in real-time team meetings and clinical decision-making. Eliminating the meeting itself is not viable.
Cost vs. human wageclaude-sonnet-52/5AI could cheaply summarize notes or prep data, but the core conferring activity still requires paid clinician time, so overall cost savings are minimal.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product reliably performs this task; it requires interactive, context-aware collaboration among multiple human professionals to reach consensus on clinical decisions. This remains fundamentally a human coordination task.
Technical feasibility todayclaude-sonnet-51/5No deployed product conducts or leads interdisciplinary treatment-team conferences autonomously; this remains firmly a human clinical activity.

Plan, organize, direct, and participate in treatment programs and activities to facilitate patients' rehabilitation, help them integrate into the community, and prevent further medical problems.

10

CI 0–20 · exposure 8 · augmentation 50 · importance 4.6/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Healthcare and rehabilitation settings adopt AI slowly for clinical decision-making due to liability, regulatory scrutiny, and high stakes for patient outcomes. Adoption remains in the pilot and administrative-support phase rather than displacement of core therapeutic delivery.
Sector adoption velocityclaude-sonnet-51/5Healthcare/rehabilitation therapy services are a low-digitization, physically embodied sector with minimal AI agent deployment for direct patient-facing activity delivery.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist by generating activity recommendations tailored to patient demographics, automating scheduling, drafting documentation, and flagging relevant medical contraindications. However, augmentation is limited to support functions; clinical judgment and interpersonal rapport remain irreplaceable.
Augmentation potentialclaude-sonnet-53/5AI can assist with program planning, documentation, progress tracking, and generating activity ideas, but does not touch the core hands-on facilitation and participation with patients.
Task automatabilityclaude-haiku-4-5-202510012/5While AI could assist with aspects like scheduling and generating activity suggestions, the core task requires ongoing clinical assessment, individualized treatment planning, and real-time responsiveness to patient needs that resist full automation. The participation and facilitation components are fundamentally relational and require human presence and judgment.
Task automatabilityclaude-sonnet-51/5This requires hands-on facilitation, physical presence, real-time adaptation to patient emotional/physical states, and direct interpersonal engagement that current AI cannot perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Recreational therapy typically requires state licensure or certification (NCTRC credential in many jurisdictions), and treatment planning is a clinical function tied to patient safety and legal liability. Regulatory frameworks and liability considerations create hard barriers to autonomous AI substitution.
Adoption barriersclaude-sonnet-54/5Recreational therapy often requires licensed/certified practitioners, direct human contact for safety and rehabilitation, and liability considerations around patient care during physical activities.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI systems for administrative support (scheduling, documentation) are relatively cheap, but they address only peripheral parts of the task. The labor-intensive core—treatment planning and facilitation—still requires qualified human therapists, making overall replacement cost-prohibitive.
Cost vs. human wageclaude-sonnet-51/5AI cannot substitute for the physical participation and supervision required, so there is no viable cost comparison for full task replacement.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed products reliably perform the end-to-end task of planning and directing personalized rehabilitation treatment programs. AI tools for scheduling or activity libraries exist, but clinical program design and patient-centered adaptation remain human-dependent in practice.
Technical feasibility todayclaude-sonnet-51/5No deployed product plans, directs, and participates in physical/recreational therapy sessions with patients; this remains outside current AI product capability.

Conduct therapy sessions to improve patients' mental and physical well-being.

2

CI 0–4 · 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 and mental health sectors remain highly regulated and risk-averse on core clinical delivery; therapy automation is constrained by licensure, liability, and professional standards, resulting in only niche experimental pilots rather than production adoption.
Sector adoption velocityclaude-sonnet-51/5Healthcare and allied therapy fields, especially hands-on rehabilitative and recreational services, show minimal AI adoption for direct treatment delivery, remaining a low-digitization, physical-presence sector.
Augmentation potentialclaude-haiku-4-5-202510012/5AI may assist with session note-taking, patient progress tracking, or supplementary psychoeducational content, but current systems cannot meaningfully augment the live therapeutic relationship itself, which is the core of the task.
Augmentation potentialclaude-sonnet-52/5AI can help with session planning, documentation, or suggesting activity ideas, but offers limited real-time assistance during the actual interpersonal therapy session itself.
Task automatabilityclaude-haiku-4-5-202510011/5Conducting therapy sessions requires real-time empathetic engagement, clinical judgment, and adaptive responsiveness to patient needs that current AI cannot reliably replicate. AI lacks the legal authority and human relational presence fundamental to therapeutic efficacy.
Task automatabilityclaude-sonnet-51/5This requires physical presence, real-time emotional attunement, hands-on activity facilitation, and adaptive human judgment during live therapeutic interaction that current AI cannot replicate end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Recreational therapists are regulated professionals; therapy sessions require direct human contact, clinical licensure, informed consent, and liability responsibility that cannot be legally delegated to AI systems under current healthcare and licensure frameworks.
Adoption barriersclaude-sonnet-55/5Recreational therapy requires licensed/certified practitioners (e.g., CTRS), direct physical and emotional patient contact, and liability for patient safety, making unsupervised AI delivery legally and practically prohibited.
Cost vs. human wageclaude-haiku-4-5-202510012/5While AI inference is cheap, integration into a therapy context requires oversight, liability management, and regulatory compliance that approaches or exceeds the cost of human therapist time, making the all-in cost comparable or unfavorable.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing this task, so cost comparison favors the human therapist entirely; any AI attempt would require extensive human oversight negating savings.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product reliably conducts independent therapy sessions with therapeutic outcomes; AI chatbots exist for psychoeducation but do not meet clinical standards for actual therapy delivery and are not positioned as substitutes for licensed practitioners.
Technical feasibility todayclaude-sonnet-51/5No deployed product conducts recreational therapy sessions with patients; this remains firmly a human-delivered clinical service.

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.