Marriage and Family Therapists
21-1013.00Median wage $66,940/yr66,740 employed (US)Rank #760 of 923 scored · top 82% by substitution
Diagnose and treat mental and emotional disorders, whether cognitive, affective, or behavioral, within the context of marriage and family systems. Apply psychotherapeutic and family systems theories and techniques in the delivery of services to individuals, couples, and families for the purpose of treating such diagnosed nervous and mental disorders.
Sub-scores
0–100 · band = confidence interval from rater disagreement
Substitution16
Exposure16
Augmentation55
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
15 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%16
panel mean rating 1.6/5 → substitution pressure 16/100
Technical feasibility todayw 20%16
panel mean rating 1.6/5 → substitution pressure 16/100
Cost vs. human wagew 15%22
panel mean rating 1.9/5 → substitution pressure 22/100
Adoption barriersw 20%inverted — strong barriers lower the score13
panel mean rating 4.5/5 (barrier strength) → substitution pressure 13/100
Sector adoption velocityw 10%17
panel mean rating 1.7/5 → substitution pressure 17/100
Task breakdown (15 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.
Provide instructions to clients on how to obtain help with legal, financial, and other personal issues.
57CI 34–81 · exposure 58 · augmentation 88 · importance 3.4/5 · click for rater detail
Provide instructions to clients on how to obtain help with legal, financial, and other personal issues.
57| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Mental health organizations are moderately adopting chatbots and information-delivery systems, but adoption remains uneven; many practices still rely on manual resource lists and in-person instruction, reflecting cautious integration of AI in therapeutic contexts. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Mental health services are a moderately digitized but relationship-intensive sector where AI adoption for clinical tasks remains largely in pilot phases rather than deep production use. |
| Augmentation potential | claude-haiku-4-5-20251001 | 5/5 | AI can dramatically augment therapists by instantly surfacing updated legal aid numbers, financial aid eligibility criteria, and local referral options in real time, allowing the therapist to focus on helping clients understand which resources fit their situation rather than spending time researching. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools can efficiently help therapists compile updated referral resources (legal aid, financial counseling services, etc.), meaningfully speeding up this administrative aspect of client support. |
| Task automatability | claude-haiku-4-5-20251001 | 5/5 | AI systems can efficiently retrieve, synthesize, and present information about legal aid, financial assistance programs, and support resources; this is primarily a information-retrieval and presentation task that does not require clinical judgment, achieving well over 50% time savings. |
| Task automatability | claude-sonnet-5 | 2/5 | AI can generate generic referral information and instructions, but tailoring these to a client's specific therapeutic context, emotional state, and case history requires human judgment that current systems cannot reliably replicate end-to-end.dez.: this remains a small, low-complexity slice of a broader clinical relationship. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | While therapists may prefer to maintain the therapeutic relationship and some jurisdictions regulate financial/legal advice, providing information about obtaining help is neither legally restricted nor requires licensure; organizational friction around depersonalization is the primary barrier. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing requirement mandates a therapist personally deliver referral information, but clinical context, liability for poor referrals, and the therapeutic relationship create meaningful friction against pure automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | The cost of AI-generated resource lists and information summaries is substantially lower than the therapist time required to research and verbally explain options; fully automated or semi-automated systems cost a fraction of professional labor. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Generating referral information is cheap via AI, but the overall task is embedded in a paid therapy session, so marginal cost savings from AI assistance are modest relative to the clinician's time already billed. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | AI-powered chatbots and knowledge systems routinely provide information about resources, referrals, and support services with high reliability; mature products exist in healthcare and legal tech that can perform this task, though human oversight for sensitive cases remains standard practice. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Chatbots and search tools can surface generic referral resources today, but no deployed product integrates this into an actual therapeutic workflow with clinical judgment about timing and framing for a specific client. |
Maintain case files that include activities, progress notes, evaluations, and recommendations.
36CI 30–43 · exposure 42 · augmentation 75 · importance 4.6/5 · click for rater detail
Maintain case files that include activities, progress notes, evaluations, and recommendations.
36| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Mental health organizations are cautious adopters of clinical AI due to liability, privacy, and professional standards concerns. While basic documentation tools are spreading, uptake of AI-assisted case-file maintenance remains limited and conservative, typical of regulated healthcare sectors. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Mental health/counseling is a moderately slow-adopting sector for AI due to privacy sensitivity, fragmented small practices, and cautious regulatory environment, though pilots of AI scribes are emerging. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can meaningfully assist therapists by auto-populating templates, summarizing session notes, flagging documentation gaps, and suggesting structured formats for progress notes. These augmentations measurably reduce manual documentation burden while the therapist retains full clinical and legal responsibility. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can meaningfully speed up drafting of progress notes and organizing case file content, letting therapists focus more time on client care while still reviewing and editing outputs. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can help draft progress notes and structure documentation, the task requires clinician judgment about therapeutic activities and patient progress that resists full automation. Meaningful portions (data entry, template filling) are automatable, but case-file maintenance demands human interpretation of therapeutic work and clinical decision-making, falling short of the 50% time-saving threshold for end-to-end automation. |
| Task automatability | claude-sonnet-5 | 3/5 | AI can draft progress notes and summarize session content from transcripts or clinician input, but a therapist must still review, verify clinical accuracy, and finalize entries, limiting full end-to-end automation. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Therapists have professional and legal obligations to maintain accurate, signed clinical records; liability for errors in progress notes or recommendations is high and non-delegable to automation. Licensing bodies and malpractice law require a human clinician's documented judgment, creating a hard requirement for human sign-off. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Clinical documentation is subject to licensing, confidentiality (HIPAA), and legal liability requirements that mandate therapist authorship and sign-off on case records. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI documentation tools reduce drafting time, but the cost of infrastructure, integration, regulatory compliance review, and clinician oversight for error-checking largely offsets savings. For a task already relatively low-cost when done by support staff, the economics are marginal. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI documentation assistants reduce time spent on notes but still require licensed clinician review and correction, so savings are real but not order-of-magnitude given oversight costs. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Documentation templates, note-generation aids, and EHR integrations exist and function in clinical settings, but they require substantial human review and correction. No deployed product reliably generates clinically accurate, liability-safe case files without therapist oversight and editing. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | AI scribe and note-generation products (e.g., ambient documentation tools) are deployed in behavioral health settings, but adoption in marriage/family therapy specifically is narrower and accuracy for nuanced clinical judgment varies. |
Provide public education and consultation to other professionals or groups regarding counseling services, issues, and methods.
28CI 25–30 · exposure 25 · augmentation 63 · importance 3.4/5 · click for rater detail
Provide public education and consultation to other professionals or groups regarding counseling services, issues, and methods.
28| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Mental health and therapeutic services remain among the slowest-adopting sectors for AI automation. Professional licensing requirements, liability sensitivity, and the demand for human credibility mean therapists are cautiously approaching even assistive AI tools, let alone delegating public education and professional consultation. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Mental health services sectors show slower AI adoption for interpersonal and consultative functions compared to information/finance sectors, with pilots more common than deployment. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can meaningfully assist therapists by drafting educational materials, organizing research, outlining consultation frameworks, and generating presentation content that the therapist then refines and delivers. This preserves professional judgment while modestly raising content production speed. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can meaningfully assist in preparing educational content, slides, talking points, and FAQs, significantly speeding up preparation while the therapist still delivers the consultation. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can generate educational content and basic consultation frameworks, this task fundamentally requires real-time professional judgment, contextual understanding of audience needs, and the credibility that comes from licensed expertise. AI cannot reliably substitute for the nuanced, adaptive consultation a licensed therapist provides to other professionals. |
| Task automatability | claude-sonnet-5 | 2/5 | AI can help draft educational content and presentation materials, but delivering public education and professional consultation requires live interaction, adapting to audience questions, and credibility that current AI cannot fully replace end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Strong regulatory and professional barriers exist: marriage and family therapists must be licensed to provide consultation on clinical matters, and professional liability concerns mean organizations typically require a licensed human to sign off on educational content and consultation services. Standards of practice and professional ethics create meaningful friction against full automation. |
| Adoption barriers | claude-sonnet-5 | 3/5 | While not strictly licensed activity in all cases, consultation on counseling methods carries professional credibility and liability expectations that favor a licensed therapist's involvement, though not a hard legal requirement. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | The all-in cost of AI systems capable of generating professional-grade educational content and consultation frameworks remains comparable to or higher than the loaded wage of a therapist when accounting for integration, content review, oversight, and liability. The output still requires substantial human verification. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI can cheaply generate supporting materials, but the actual delivery and consultation still requires paid professional time, so overall cost savings are limited. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed product reliably performs genuine professional consultation and public education in marriage and family therapy at scale. Some AI writing tools can draft educational materials, but they lack the clinical depth, legal accountability, and professional credibility required for consulting with other professionals on clinical issues. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | AI chatbots and content generators exist for drafting educational materials, but no deployed product independently runs public education sessions or professional consultations on counseling methods with reliability. |
Follow up on results of counseling programs and clients' adjustments to determine effectiveness of programs.
27CI 25–29 · exposure 25 · augmentation 63 · importance 3.5/5 · click for rater detail
Follow up on results of counseling programs and clients' adjustments to determine effectiveness of programs.
27| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Mental health organizations have been slow to adopt AI for clinical decision-making due to liability concerns, professional licensing requirements, and resistance from practitioners. Adoption remains largely limited to administrative or supportive tools rather than core clinical judgment tasks. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Mental health services are a lower-digitization, high-human-contact sector; while telehealth and outcome-tracking software have grown, deep AI-driven automation of clinical follow-up remains uncommon. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can meaningfully assist by summarizing client notes, tracking quantitative outcome metrics, and flagging potential concerns for review, enabling therapists to focus clinical attention more efficiently. However, the therapist must remain the primary interpreter of effectiveness. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-based outcome tracking, automated check-in surveys, and data aggregation can meaningfully help therapists monitor client progress and flag patterns for review, improving efficiency of follow-up while the therapist retains clinical decision-making. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Assessing client adjustment and counseling effectiveness requires nuanced evaluation of emotional and relational progress that depends heavily on clinical judgment, contextual understanding, and therapeutic alliance—domains where AI lacks reliable capability. While AI could help organize notes and flag certain metrics, it cannot independently determine counseling program effectiveness. |
| Task automatability | claude-sonnet-5 | 2/5 | Some structured follow-up (sending surveys, tracking symptom scores, scheduling check-ins) can be automated, but clinically interpreting client adjustment and program effectiveness requires human judgment and relational context AI cannot fully replicate today. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Significant regulatory and liability barriers exist: therapists are legally and ethically responsible for treatment outcomes; malpractice risk is high if automation errors influence clinical decision-making. Mental health records are heavily regulated (HIPAA), and treatment decisions require licensed professional judgment. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Assessing therapeutic progress and adjusting treatment plans is generally within licensed therapist scope of practice, with confidentiality and clinical judgment requirements limiting full delegation to AI. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | The cost of AI systems capable of handling nuanced clinical evaluation, combined with required human oversight and clinical validation, is likely comparable to or exceeds the cost of human therapist time spent on follow-up assessment. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Automated survey/reminder tools are cheap, but the interpretive follow-up work still requires a licensed therapist's paid time, keeping overall cost comparable to human-only processes. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed product reliably performs independent assessment of counseling effectiveness at clinical-grade accuracy. Existing tools may offer outcome tracking or survey administration, but the interpretive clinical judgment that defines this task remains human-dependent in practice. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Patient outcome-tracking software and automated symptom questionnaires exist in mental health practice, but they support rather than perform the clinical evaluation of counseling effectiveness itself. |
Confer with other counselors, doctors, and professionals to analyze individual cases and to coordinate counseling services.
23CI 20–25 · exposure 20 · augmentation 50 · importance 4.0/5 · click for rater detail
Confer with other counselors, doctors, and professionals to analyze individual cases and to coordinate counseling services.
23| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Healthcare adoption of AI coordination tools remains slow and fragmented; most case conferences still occur via phone, email, and in-person meetings with minimal AI augmentation, reflecting organizational inertia and clinical skepticism of fully automated coordination. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Healthcare and mental health services are historically slower adopters of AI for interpersonal clinical coordination, with most uptake limited to administrative documentation support rather than case conferencing itself. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can usefully assist by automatically extracting and summarizing case information, flagging relevant clinical history, and organizing documentation before a conference, meaningfully saving preparation time while the therapist maintains full decision-making authority. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can assist by organizing case histories, drafting summaries for case conferences, or suggesting relevant research, meaningfully aiding preparation even though the core collaborative task remains human-led. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | AI can assist with summarizing case details and organizing information, but conferring—which requires nuanced professional judgment, interpersonal negotiation, and real-time responsiveness to other clinicians' expertise—remains fundamentally human-dependent. Current systems cannot reliably replace the back-and-forth clinical reasoning that case conferences demand. |
| Task automatability | claude-sonnet-5 | 2/5 | This requires synthesizing sensitive clinical information across professionals, exercising judgment, and building trust-based collaborative relationships that current AI cannot substitute for end-to-end, though it can help summarize case notes or draft communications. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Conferring and coordinating care often involves multi-professional liability, patient privacy (HIPAA), and professional accountability; clinicians are legally and ethically responsible for coordinating care decisions, and regulatory/licensing requirements protect the human's role in these decisions. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Licensing, confidentiality (HIPAA), liability for clinical coordination decisions, and professional standards of care create strong barriers against AI performing this task autonomously. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | The cost of AI-assisted summarization and documentation is modest, but human clinicians must still conduct the actual conference; AI reduces overhead but does not displace the core labor, making cost savings marginal relative to therapist wages. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI could cheaply draft summaries or facilitate scheduling, but the actual professional consultation and decision-making still requires paid clinician time, so overall cost savings are limited. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While AI can draft summaries or flag clinical patterns from case notes, no deployed product reliably conducts multi-professional case conferences or coordinates care across providers. Existing EHR integrations and decision-support tools offer narrow, fragmented assistance rather than end-to-end conferencing capability. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product independently confers with other clinicians to coordinate care decisions; this remains a human relational and clinical judgment task with only research-stage support tools. |
Collect information about clients, using techniques such as testing, interviewing, discussion, or observation.
16CI 7–25 · exposure 17 · augmentation 63 · importance 4.5/5 · click for rater detail
Collect information about clients, using techniques such as testing, interviewing, discussion, or observation.
16| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Mental health practices have adopted AI for administrative tasks (scheduling, note-taking) but not for clinical assessment itself; actual displacement in therapeutic information-gathering is minimal, and professional resistance and regulatory constraints limit rapid adoption of automation in this function. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Mental health services are a person-heavy, moderately digitized sector; AI adoption for intake automation is growing but assessment still remains largely human-led with limited penetration. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist therapists by drafting intake summaries, generating assessment prompts, or organizing client information post-session, but cannot independently gather or interpret clinical information; these tools offer moderate productivity gains on documentation and organization, not on the core clinical task itself. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can meaningfully assist by administering standardized questionnaires, transcribing and summarizing sessions, flagging risk indicators, and helping organize intake data for the therapist's review. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | AI cannot conduct reliable therapeutic interviews or observations that establish the nuanced trust, rapport, and clinical judgment required to assess family dynamics and client mental states. Information gathering in therapy depends on interpreting non-verbal cues, therapeutic presence, and adaptive responsiveness that current systems cannot replicate meaningfully. |
| Task automatability | claude-sonnet-5 | 2/5 | Interviewing and observing clients in a therapeutic context to assess emotional state and relational dynamics requires nuanced human presence, empathy, and adaptive follow-up that current AI cannot replicate end-to-end. AI can assist with structured intake questionnaires but not the full clinical interview/observation process. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Licensing and legal standards (state licensure, liability, clinical responsibility) require a credentialed therapist to conduct and sign off on client assessments and diagnoses. Regulatory and professional standards mandate human clinical judgment and direct client contact for information gathering. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Clinical assessment is tied to licensure and standards of care; a licensed therapist must typically conduct diagnostic interviews and interpret client presentation, creating strong regulatory and liability barriers. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The human therapist cost per session is moderate ($100–200+), while meaningful clinical assessment requires human expertise and accountability; AI tools that might assist with forms or note-taking are complementary rather than substitutive and do not meaningfully reduce the primary cost. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI-driven intake tools are cheap for basic data collection, but the licensed therapist's clinical interviewing and observational judgment still requires costly human time, so overall savings are limited. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While AI can assist with intake questionnaires or generate structured interview templates, no deployed product reliably performs the clinical assessment core of this task. Existing systems lack the ability to conduct genuine therapeutic interviews or accurately interpret client communication and family interactions in real time. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some digital intake forms and chatbot-based screening tools exist, but no deployed product reliably conducts full clinical assessment interviews or behavioral observation for therapy at production scale. |
Ask questions that will help clients identify their feelings and behaviors.
14CI 4–25 · exposure 13 · augmentation 50 · importance 4.7/5 · click for rater detail
Ask questions that will help clients identify their feelings and behaviors.
14| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Adoption of AI for core therapeutic questioning is extremely low due to regulatory constraints, professional gatekeeping, and institutional risk aversion in mental health settings. Pilot projects exist but production deployment at scale is negligible. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Mental health services are a low-digitization, high-human-contact sector; AI tools are used adjunctively (notes, scheduling) but adoption of AI performing core therapeutic questioning is minimal and slow. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist therapists by suggesting question frames, organizing client responses, or highlighting patterns in notes, which can modestly boost therapist productivity. However, the human therapist remains central to the questioning and interpretation process. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help therapists prepare question frameworks, review session notes for patterns, or suggest reflective prompts, offering moderate assistance without replacing the clinician's real-time judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can generate questions and prompt text, effective therapeutic questioning requires genuine responsiveness to client emotional states, subtle behavioral cues, and dynamic relationship building that current systems cannot replicate reliably. The task demands moment-to-moment adaptive depth that falls well short of the 50% time-saving threshold. |
| Task automatability | claude-sonnet-5 | 1/5 | This task requires live clinical judgment, empathy, and adaptive real-time responsiveness to a client's emotional state, which is core to the therapeutic relationship and cannot be automated end-to-end today without losing quality and safety. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Strong regulatory and ethical barriers protect this task: therapy licensing laws, malpractice liability, professional ethics codes, and the legal requirement in most jurisdictions that a licensed human therapist hold clinical responsibility. Client safety and informed consent further restrict substitution. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Marriage and family therapy is a licensed profession with legal requirements for who may practice therapy, strong liability exposure for harm, and clients generally require human therapeutic rapport, making this a hard-barrier task. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | The cost of infrastructure, fine-tuning, oversight, and liability insurance for therapeutic AI likely exceeds the wage cost of a trained therapist, especially given the low-volume, highly personalized nature of each client interaction. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | While AI inference is cheap, the liability, oversight, and quality requirements for genuine therapeutic work mean any viable AI-assisted approach still requires costly human clinical oversight, keeping all-in costs comparable to or higher than a human alone for this specific task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Chatbots can produce generic prompting questions, but no deployed product reliably performs therapeutic-grade questioning that identifies feelings and behaviors with the nuance and safety margins required in clinical practice. Products exist but with significant gaps in contextual sensitivity and clinical appropriateness. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs licensed therapeutic questioning autonomously in production; chatbot wellness apps exist but are not substitutes for a licensed therapist conducting this clinical task. |
Develop and implement individualized treatment plans addressing family relationship problems, destructive patterns of behavior, and other personal issues.
10CI 0–20 · exposure 13 · augmentation 63 · importance 4.6/5 · click for rater detail
Develop and implement individualized treatment plans addressing family relationship problems, destructive patterns of behavior, and other personal issues.
10| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Healthcare and mental health sectors are among the slowest to adopt AI automation due to regulatory constraints, liability concerns, and the irreplaceable need for human clinical judgment and therapeutic alliance in this domain. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Behavioral health is a slower-adopting sector for clinical decision-making AI due to regulatory, ethical, and trust concerns, though administrative AI tools are spreading somewhat faster. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI could usefully assist by drafting plan templates, organizing intake information, suggesting evidence-based interventions, or analyzing session notes—but the therapist remains central to the decision-making process and clinical responsibility. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can meaningfully assist therapists by suggesting evidence-based intervention frameworks, drafting documentation, and organizing session notes, improving efficiency while the therapist retains full clinical control. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Developing individualized treatment plans requires deep clinical judgment, understanding of complex family dynamics, and therapeutic expertise that current AI cannot reliably replicate. The task demands integration of diagnostic assessment, treatment goal setting, and personalized intervention design—beyond what AI systems can perform end-to-end at clinically safe quality standards. |
| Task automatability | claude-sonnet-5 | 2/5 | AI can help draft generic treatment plan templates but cannot reliably tailor plans requiring clinical judgment, ongoing relational assessment, and nuanced ethical decision-making at a quality bar equal to a licensed therapist.atable |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | This task is heavily protected by licensing requirements—only licensed marriage and family therapists can legally develop and implement treatment plans. Regulatory boards, malpractice liability, and ethical standards create hard barriers against substitution or unsupervised AI automation. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Treatment planning is a licensed clinical activity requiring professional judgment, confidentiality safeguards, and legal accountability; only a licensed therapist can develop and implement such plans. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Even if partial automation were possible, the liability costs, oversight requirements, and need for licensed human sign-off mean the all-in cost of AI would exceed or equal the cost of a human therapist performing the task independently. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | While AI drafting assistance is cheap, the clinical judgment, liability, and human oversight required keep the effective all-in cost of a compliant plan closer to human-level costs. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product reliably performs the full scope of this task in clinical practice today. While AI may assist with documentation or provide psychoeducational content, the core responsibility of developing and implementing a treatment plan must remain with a licensed therapist due to the complexity and liability involved. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | No deployed product autonomously creates and implements individualized family therapy treatment plans in clinical practice; existing AI mental health tools are narrow, largely supportive or administrative, not clinical plan authors of record. |
Encourage individuals and family members to develop and use skills and strategies for confronting their problems in a constructive manner.
7CI 4–11 · exposure 5 · augmentation 50 · importance 4.8/5 · click for rater detail
Encourage individuals and family members to develop and use skills and strategies for confronting their problems in a constructive manner.
7| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Despite interest in mental health AI, actual displacement of therapists' core functions remains minimal. The sector moves cautiously due to regulatory requirements, liability concerns, and professional standards that prioritize human therapeutic relationship. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Mental health/counseling remains a high-touch, in-person-preferred field with slow, cautious AI adoption largely limited to administrative support and screening tools rather than core therapeutic tasks. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist by suggesting evidence-based coping strategies, offering prompt templates, or providing psychoeducational materials that therapists then contextualize and encourage clients to use. This represents meaningful but limited augmentation within a human-directed therapeutic process. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help therapists prepare psychoeducational materials, suggest coping strategies, or draft session notes, but the core relational encouragement work is not meaningfully augmented within-session. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task fundamentally requires real-time, adaptive engagement with human emotions and interpersonal dynamics. Current AI systems cannot develop the contextual understanding, emotional attunement, and relational presence needed to genuinely encourage behavioral change and skill-building in therapeutic settings. |
| Task automatability | claude-sonnet-5 | 1/5 | Encouraging clients toward constructive behavioral change requires real-time relational attunement, trust, and clinical judgment that current AI cannot replicate end-to-end at equal quality. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Licensing laws in most jurisdictions require licensed marriage and family therapists to perform or directly supervise therapeutic interventions. Liability concerns, ethical standards, and the inherent requirement for human clinical judgment create hard regulatory and professional barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Marriage and family therapy is a licensed profession with legal and ethical requirements for a qualified human to provide treatment, especially given liability for mental health harm. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI-generated therapeutic guidance is cheap to produce, but integration into clinical practice, liability coverage, and necessary human oversight add significant costs. For genuine therapeutic encouragement requiring professional judgment, the cost advantage remains marginal compared to human therapists. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | While AI inference is cheap, the human clinician remains necessary for safe, effective delivery, so realistic cost comparison still favors the human when accounting for liability and oversight needed to deploy AI safely. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While AI chatbots can provide generic coping strategies and psychoeducational content, no deployed product reliably performs the nuanced work of encouraging skill development across diverse family situations with therapeutic competence. Existing mental health AI tools serve as supplements, not substitutes for this core therapeutic function. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed clinical product autonomously delivers therapeutic encouragement and coaching in live family therapy sessions; chatbot mental health tools are adjunct, narrow, and not equivalent to licensed therapy practice. |
Confer with clients to develop plans for posttreatment activities.
7CI 0–15 · exposure 8 · augmentation 50 · importance 4.0/5 · click for rater detail
Confer with clients to develop plans for posttreatment activities.
7| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Mental health and therapy sectors have been slow to adopt AI-driven automation, with strong professional and regulatory resistance to replacing licensed therapists in core clinical functions. Adoption remains confined to ancillary administrative tasks, not client-facing therapeutic work. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Mental health services remain a high-touch, in-person-preferred sector with slow AI adoption for core clinical tasks, though administrative and note-taking AI tools are spreading in the periphery. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI could assist by suggesting evidence-based posttreatment activity templates, summarizing treatment progress, or drafting initial plan outlines for the therapist to refine with the client. However, the human therapist must remain the primary architect of the plan through direct dialogue with the client. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help draft aftercare plan templates, summarize session notes, or suggest resource lists, giving therapists useful support while they retain full control of the actual conference and decision-making. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires deep interpersonal dialogue, empathetic understanding of individual client circumstances, and collaborative goal-setting—domains where current AI cannot reliably replace human judgment or build genuine therapeutic rapport. The task is inherently relational and requires the therapist to synthesize complex emotional and behavioral information to co-create meaningful, personalized plans. |
| Task automatability | claude-sonnet-5 | 2/5 | Planning posttreatment activities requires interpreting a specific client's clinical history, relational dynamics, and readiness for change, which demands real-time judgment and rapport that current AI cannot replicate end-to-end. AI could draft generic aftercare templates, but not conduct the actual client conference. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Marriage and family therapy is a licensed profession with legal and ethical requirements that the therapist—not an AI—must perform assessment, establish therapeutic relationship, and develop treatment plans. Liability, professional ethics, and regulation create hard barriers to full automation or unsupervised delegation to AI. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Marriage and family therapy is a licensed profession with legal and ethical requirements that a qualified human must conduct treatment planning and sign off on client care, creating a hard regulatory and liability barrier. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI inference is cheap, but the integration and human oversight required to ensure therapeutic safety and quality would consume most of the cost savings. Any AI tool used would require a therapist to review, validate, and refine output, making the combined cost only marginally lower than direct human provision. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | Since AI cannot perform the core conferencing and clinical judgment task independently, the effective cost comparison favors the human therapist, whose licensed judgment is required for liability and quality reasons. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI system reliably performs therapeutic planning conversations with clients in production clinical settings. While chatbots can offer generic advice or psychoeducation, they cannot substitute for a licensed therapist's assessment, relationship-building, or accountability in a real therapeutic context. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product autonomously conducts client conferences to establish posttreatment plans; existing AI mental-health tools are largely research-stage or adjunct chatbots, not substitutes for licensed therapist-led planning sessions. |
Determine whether clients should be counseled or referred to other specialists in such fields as medicine, psychiatry, or legal aid.
7CI 0–15 · exposure 8 · augmentation 50 · importance 4.0/5 · click for rater detail
Determine whether clients should be counseled or referred to other specialists in such fields as medicine, psychiatry, or legal aid.
7| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Mental health and therapeutic services remain among the most heavily regulated and human-contact-intensive sectors. Adoption of autonomous AI for clinical decision-making is minimal, with the field maintaining strong professional and legal requirements for human involvement. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Mental health services remain a mostly high-touch, professionally regulated field with cautious, slow AI adoption, especially for clinical decision-making rather than administrative tasks. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI could assist by summarizing client symptoms, flagging potential red flags for psychiatric or medical comorbidities, or organizing information about specialist availability and insurance coverage. However, the final referral decision must remain with the licensed therapist. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help by summarizing intake information, flagging risk indicators, or suggesting referral resources, usefully supporting the therapist's own judgment without replacing it. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires nuanced clinical judgment, understanding of complex client situations, and professional gatekeeping responsibilities that demand human expertise. Current AI systems cannot reliably make binding referral decisions involving medical, psychiatric, or legal domains without human oversight. |
| Task automatability | claude-sonnet-5 | 2/5 | This requires clinical judgment integrating nuanced client history, risk assessment, and professional discretion that current AI cannot reliably replicate end-to-end; at most AI can support triage suggestions, not perform the determination itself with equal quality time savings. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Strong legal and professional barriers exist: marriage and family therapists must hold licensure to practice, are bound by professional ethical codes, carry malpractice liability for referral decisions, and are legally responsible for duty-to-refer determinations. Regulatory frameworks require licensed human professionals to make these gatekeeping decisions. |
| Adoption barriers | claude-sonnet-5 | 5/5 | This is a core licensed clinical judgment task with direct patient safety and legal liability implications, typically requiring a credentialed professional to make and be accountable for the determination. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The cost of AI errors in clinical referral decisions (missed diagnoses, inappropriate referrals, liability exposure) vastly exceeds the human labor cost. Integration, oversight, and error remediation would likely exceed the cost of human clinical judgment. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | While AI screening tools are cheap to run, the liability and oversight costs of using AI for a clinical referral decision (with human review required) largely erode cost savings versus a licensed therapist making the call. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product reliably performs autonomous clinical triage and specialist referral decisions for therapy clients. While AI can assist in information gathering, the high-stakes nature of therapeutic referrals and professional liability requirements mean no production system operates end-to-end in this capacity. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed clinical product autonomously makes referral/treatment-pathway decisions for therapy clients in production; existing tools are decision-support or screening aids at best, not autonomous deciders. |
Gather information from doctors, schools, social workers, juvenile counselors, law enforcement personnel, and others to make recommendations to courts for resolution of child custody or visitation disputes.
7CI 0–15 · exposure 8 · augmentation 38 · importance 3.3/5 · click for rater detail
Gather information from doctors, schools, social workers, juvenile counselors, law enforcement personnel, and others to make recommendations to courts for resolution of child custody or visitation disputes.
7| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Legal, ethical, and professional standards in family law and mental health preclude AI substitution for custody evaluations. No meaningful adoption of AI for this task exists in practice. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Mental health and family therapy sectors are slow AI adopters overall, especially for legally consequential assessments, with pilots for administrative support far more common than adoption for substantive recommendations. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI might assist marginally with document summarization or scheduling interviews across multiple sources, but the core task of gathering sensitive information, interpreting it professionally, and making custody recommendations requires human clinical expertise and cannot be meaningfully augmented by current systems without removing human judgment from the decision point. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help therapists organize, summarize, and cross-reference reports from various parties (medical records, school reports, law enforcement notes), improving efficiency in preparing for and drafting recommendations. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires synthesizing sensitive information from multiple professional sources, exercising clinical judgment about child welfare, and formulating legally-consequential recommendations. Current AI cannot reliably collect confidential information from external professionals, assess credibility and context, or make custodial recommendations that meet professional standards of care. |
| Task automatability | claude-sonnet-5 | 2/5 | While AI can help aggregate and summarize documents from multiple sources, the task requires synthesizing sensitive multi-party information, exercising clinical judgment, and forming professional recommendations that current AI cannot reliably do end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Strong legal and regulatory barriers protect this task: court systems require licensed mental health professionals to conduct custody evaluations, sign reports, and testify; liability for incorrect recommendations is substantial; and jurisdictions have explicit licensing and qualification requirements for those making such recommendations. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Court-facing custody recommendations typically require licensed therapists or evaluators whose findings carry legal weight and liability, making this a hard-barrier task requiring credentialed human sign-off. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | AI cannot perform this task independently, so comparative cost is not meaningful. Human therapists with court credentials, training in custody evaluation, and professional liability coverage remain the only feasible option. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI could cheaply assist with document summarization and note organization, but the overall task still requires expensive human labor (interviews, judgment, legal liability), keeping all-in costs comparable to or only marginally below human costs. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI product performs court-ordered child custody evaluations or recommendations. This task requires licensed mental health expertise, legal knowledge, and accountability that cannot be delegated to AI systems; courts and jurisdictions require qualified human professionals to sign off on recommendations. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs this integrated custody-evaluation function; this remains a specialized clinical/legal task performed by licensed professionals with no production AI substitute. |
Counsel clients on concerns, such as unsatisfactory relationships, divorce and separation, child rearing, home management, or financial difficulties.
5CI 4–6 · exposure 0 · augmentation 38 · importance 4.5/5 · click for rater detail
Counsel clients on concerns, such as unsatisfactory relationships, divorce and separation, child rearing, home management, or financial difficulties.
5| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Adoption of AI for actual therapy delivery remains negligible in clinical practice. The profession is regulated, liability-averse, and human-contact intensive; organizations have not adopted AI agents for primary counseling roles. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Behavioral health is a highly relational, regulated sector with slow AI adoption for direct clinical care, though some digital mental health tools are gaining pilot use. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI might assist therapists with documentation, session summaries, or evidence-based psychoeducational materials, but the core therapeutic work—building alliance, facilitating insight, intervening in family dynamics—resists meaningful augmentation. The therapeutic relationship itself is the primary intervention. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help therapists with session notes, treatment planning suggestions, psychoeducation materials, and between-session client resources, but does not replace the core counseling interaction. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires genuine therapeutic rapport, empathetic listening, real-time emotional calibration, and personalized intervention based on complex human dynamics. Current AI systems cannot provide the nuanced, context-dependent counseling that addresses deeply personal and often sensitive interpersonal issues at the standard expected of a licensed therapist. |
| Task automatability | claude-sonnet-5 | 1/5 | Therapeutic counseling requires nuanced human judgment, trust-building, and real-time emotional attunement across sensitive family dynamics that current AI cannot reliably replicate end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Marriage and family therapy is a licensed profession in most jurisdictions; a licensed LMFT, LCSW, or equivalent must legally conduct therapy. Insurance reimbursement, liability, and ethical standards all require a credentialed human to perform and sign off on treatment decisions, creating hard legal and regulatory barriers. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Licensed therapists are legally required to provide this counseling; strong liability, ethical, and regulatory frameworks (state licensure, scope-of-practice laws) prevent AI from independently performing this task. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Even if basic AI-generated advice were acceptable (which it is not clinically), the cost of oversight, validation, and liability management would be substantial. Licensed therapists command significant wages, but the regulatory and professional indemnity burden of AI substitution keeps all-in costs uncompetitive. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Even if AI chat support is cheap per interaction, achieving comparable therapeutic outcomes requires human oversight and liability coverage, negating cost advantage for actual counseling delivery. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product reliably performs marriage and family therapy as a primary intervention. While chatbots offer general coping suggestions, they lack the diagnostic acumen, clinical judgment, and accountability required for actual therapeutic counsel on divorce, child-rearing conflicts, or family financial crises. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product independently conducts licensed marriage/family counseling; chatbot mental health tools remain adjunct, unlicensed, and not substitutes for clinical therapy. |
Supervise other counselors, social service staff, and assistants.
4CI 0–7 · exposure 5 · augmentation 50 · importance 3.4/5 · click for rater detail
Supervise other counselors, social service staff, and assistants.
4| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Healthcare and mental-health sectors show laggard adoption of autonomous AI in clinical roles; supervision remains a protected human responsibility with strong regulatory and organizational resistance to delegation. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Behavioral health and social services are relatively slow to adopt AI for supervisory or people-management functions, with pilots limited mostly to administrative support. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist supervisors by organizing case notes, scheduling, flagging documentation gaps, and generating compliance reports, but the human supervisor remains essential for clinical judgment and therapeutic oversight. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help supervisors with documentation, case note review, scheduling, and identifying training needs, but does not replace the interpersonal supervisory relationship. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Supervision requires real-time judgment of clinical performance, interpersonal feedback, ethical oversight, and adaptive coaching—tasks that demand understanding of human dynamics and contextual decision-making that current AI cannot reliably perform end-to-end. |
| Task automatability | claude-sonnet-5 | 1/5 | Clinical supervision requires nuanced judgment, live observation, mentorship, and interpersonal accountability that current AI cannot perform end-to-end.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Clinical supervision of counselors is legally and ethically mandated to be performed by a licensed mental-health professional in most jurisdictions; this creates a hard barrier that AI cannot cross regardless of technical capability. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Clinical supervision of licensed/associate counselors is typically legally mandated to be performed by a qualified, licensed supervisor, creating hard regulatory and liability barriers. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Supervision is inherently labor-intensive and requires a licensed professional's accountability; the all-in cost of any AI system plus mandatory human oversight would exceed the cost of direct human supervision. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute for supervisory judgment and liability oversight, so cost comparison favors the human by default. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While AI can assist with scheduling, documentation review, and flagging compliance gaps, no deployed product reliably handles the core supervisory functions of evaluating clinical competence, providing corrective feedback, and ensuring therapeutic quality at scale. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs clinical/staff supervision of counselors autonomously; AI is at best a note-taking or scheduling aid in this context. |
Write evaluations of parents and children for use by courts deciding divorce and custody cases, testifying in court if necessary.
0CI 0–0 · exposure 0 · augmentation 38 · importance 3.2/5 · click for rater detail
Write evaluations of parents and children for use by courts deciding divorce and custody cases, testifying in court if necessary.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Adoption is effectively zero in this regulated clinical and legal domain. Family courts have not moved toward AI-authored evaluations; the professional licensing system and legal requirements make automation infeasible at present. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Family courts and forensic mental health evaluation are highly conservative, low-digitization domains with essentially no AI adoption for this specific function. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could marginally assist with administrative tasks (scheduling, literature review, report formatting), but the core assessment—interviewing parties, observing family dynamics, forming clinical opinions—remains exclusively human. Augmentation is minimal because the human must perform the critical diagnostic work and carry full accountability. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help draft report templates, organize notes, summarize session transcripts, or research relevant case law, giving moderate assistance while the clinician retains full evaluative and testimonial responsibility. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires deep clinical judgment, human-centered assessment, and legal accountability that current AI cannot perform end-to-end. Writing court-admissible custody evaluations demands synthesizing sensitive psychological information, cross-examining inconsistencies, and making high-stakes determinations about child welfare—all legally and ethically tied to a licensed clinician's professional judgment. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires clinical judgment based on in-person observation, interviews, and psychological assessment of family dynamics, plus live courtroom testimony under oath—none of which current AI can perform end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Courts require evaluations to be authored and signed by a licensed mental-health professional; the evaluator must testify under oath and accept legal liability for findings. Regulatory and legal barriers are absolute: a human license and professional standing are mandatory, not optional. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Courts require a qualified, often court-appointed licensed professional to conduct evaluations and testify under oath, with legal accountability and cross-examination requirements that only a human can satisfy. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | AI cannot yet substitute for this task, making cost comparison premature. Even if partial drafting existed, the oversight burden and liability exposure would keep total cost well above hiring a licensed therapist, whose work commands $150–300/hour in this legal context. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | AI cannot substitute for the licensed evaluator or witness, so there is no viable AI cost basis to compare against the human expert's fee. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI product can independently write valid court evaluations for custody cases or testify in proceedings. Such work requires licensed professional judgment, court acceptance of the evaluator's credentials, and legal accountability; AI outputs lack the professional standing and legal signature required. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product conducts custody evaluations or provides expert court testimony; this remains entirely a licensed clinician's responsibility. |
Related occupations — Community & Social Service
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.