Clinical and Counseling Psychologists

19-3033.00
Median wage $100,580/yr75,990 employed (US)Rank #633 of 923 scored · top 69% by substitution

Assess, diagnose, and treat mental and emotional disorders of individuals through observation, interview, and psychological tests. Help individuals with distress or maladjustment understand their problems through their knowledge of case history, interviews with patients, and theory. Provide individual or group counseling services to assist individuals in achieving more effective personal, social, educational, and vocational development and adjustment. May design behavior modification programs and consult with medical personnel regarding the best treatment for patients.

Sub-scores

0–100 · band = confidence interval from rater disagreement

Substitution22
Exposure22
Augmentation61

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

30 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%21

panel mean rating 1.8/5 → substitution pressure 21/100

Technical feasibility todayw 20%23

panel mean rating 1.9/5 → substitution pressure 23/100

Cost vs. human wagew 15%27

panel mean rating 2.1/5 → substitution pressure 27/100

Adoption barriersw 20%inverted — strong barriers lower the score15

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

Sector adoption velocityw 10%24

panel mean rating 2.0/5 → substitution pressure 24/100

Task breakdown (30 tasks)

Substitution pressure per task, weighted by O*NET importance in the composite. Expand a task for the full rater audit trail — every rating, every model, every rationale.

Document patient information including session notes, progress notes, recommendations, and treatment plans.

46

CI 30–62 · exposure 50 · augmentation 88 · importance 4.8/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Adoption of documentation AI in clinical psychology remains limited and cautious; most practices still rely on manual or minimally AI-assisted documentation due to liability concerns, regulatory uncertainty, and clinician skepticism about handing over clinical record-keeping to automated systems.
Sector adoption velocityclaude-sonnet-53/5Healthcare and behavioral health are adopting AI documentation tools steadily but more cautiously than fully digital sectors like finance or tech, with pilots more common than universal deployment.
Augmentation potentialclaude-haiku-4-5-202510014/5AI-assisted documentation tools significantly augment clinician productivity by automating transcription, suggesting structure, and generating first drafts that clinicians then refine, enabling faster documentation without replacing clinical judgment or sign-off authority.
Augmentation potentialclaude-sonnet-55/5AI scribes and note-drafting assistants significantly speed up documentation while the clinician remains in control, reviewing and finalizing content for clinical accuracy and compliance.
Task automatabilityclaude-haiku-4-5-202510012/5AI can draft session notes and structure documentation from transcripts or summaries, but clinical documentation requires nuanced clinical judgment, accurate risk assessment, and legally defensible record-keeping that necessitates substantial human review and modification. Significant setup and oversight prevent reaching the 50% time-saving threshold for the full task.
Task automatabilityclaude-sonnet-54/5AI scribes/ambient documentation tools can draft session notes, progress notes, and treatment plans from transcripts or clinician summaries with substantial time savings, though clinician review remains necessary for accuracy and clinical judgment.
Adoption barriersclaude-haiku-4-5-202510014/5Clinical documentation is subject to strict regulatory requirements (HIPAA, state practice acts, licensing boards), must be legally defensible, and the psychologist bears professional and legal liability for accuracy and completeness. Malpractice exposure and licensing board standards create hard barriers to full automation without clinician sign-off.
Adoption barriersclaude-sonnet-53/5Clinical documentation must be reviewed and signed by a licensed clinician who bears legal and ethical responsibility, and confidentiality/HIPAA concerns add friction, though the drafting itself isn't regulated.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI transcription and templating tools reduce some documentation labor, but the cost of oversight, error correction, compliance review, and liability insurance considerations means total cost savings remain marginal relative to clinician wages, especially given the high cost of documentation errors.
Cost vs. human wageclaude-sonnet-54/5AI documentation tools cost a small subscription/usage fee compared to clinician time spent writing notes, offering substantial savings even after factoring in review time.
Technical feasibility todayclaude-haiku-4-5-202510013/5Clinical documentation AI tools and EHR-integrated transcription products exist and are used in some practices, but they produce outputs with material error rates in clinical accuracy, liability exposure, and completeness that require heavy human revision. No mature system reliably handles the full scope of clinical documentation without substantial clinician rework.
Technical feasibility todayclaude-sonnet-53/5AI medical/therapy scribe products (e.g., ambient documentation tools) are deployed in some behavioral health practices but adoption is uneven and error-checking is still required, so reliability in production varies.

Refer clients to other specialists, institutions, or support services as necessary.

46

CI 25–67 · exposure 50 · augmentation 75 · importance 3.7/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510013/5Healthcare systems are adopting clinical decision support and EHR-integrated referral tools at moderate pace; pilots and early deployment are common in major health systems, but production adoption remains unevenly distributed. Regulatory compliance requirements and clinician resistance to algorithm-based referral slow velocity compared to purely administrative sectors.
Sector adoption velocityclaude-sonnet-52/5Healthcare and mental health services are historically slow adopters of AI for clinical decision-making tasks, with pilots more common than production deployment for referral management.
Augmentation potentialclaude-haiku-4-5-202510015/5AI excels at augmenting this task: rapid querying of specialist networks, insurance coverage, wait times, and service offerings; automatic flagging of co-occurring needs (e.g., substance abuse services alongside depression); and generation of referral templates that clinicians review and customize. This substantially accelerates clinician productivity while preserving professional judgment.
Augmentation potentialclaude-sonnet-53/5AI can help by searching for appropriate specialists, summarizing client history for referral letters, or flagging matches to resources, meaningfully aiding but not replacing the psychologist's judgment.
Task automatabilityclaude-haiku-4-5-202510014/5Current AI systems can reliably parse client presentations, match them against a knowledge base of specialist types and services, and generate appropriate referral recommendations with high accuracy. This task involves structured matching rather than novel judgment, enabling ≥50% time savings in documentation and initial triage, though final human sign-off remains standard practice.
Task automatabilityclaude-sonnet-52/5Referral requires clinical judgment about a client's specific needs, diagnosis, insurance, and rapport with other providers, which AI cannot yet reliably assess end-to-end.
Adoption barriersclaude-haiku-4-5-202510013/5While no legal barrier requires a licensed psychologist to personally execute referrals, professional liability, standards of care, and malpractice risk mean clinicians must review and approve any AI-generated referral before sending. Organizational protocols, EHR integration requirements, and professional liability insurance create meaningful friction against full automation.
Adoption barriersclaude-sonnet-54/5Referral decisions are part of licensed clinical practice with liability implications; a psychologist must exercise professional judgment and often sign off, creating strong regulatory and ethical barriers.
Cost vs. human wageclaude-haiku-4-5-202510014/5AI inference for referral matching costs negligible per case (fractions of a cent), while clinician time spent researching available services, verifying credentials, and generating referral paperwork may consume 15–30 minutes per referral. The labor cost for this task in the US is approximately $20–40 per referral; AI reduces this by 60–80%.
Cost vs. human wageclaude-sonnet-52/5AI could cheaply generate a list of local specialists, but the clinical judgment, liability, and follow-through required still necessitate a licensed professional, keeping overall cost comparable to human labor.
Technical feasibility todayclaude-haiku-4-5-202510014/5Deployed products (EHR systems, clinical decision support, scheduling/referral platforms) already perform referral matching and documentation at scale in healthcare organizations. Systems like clinical registry lookups and AI-assisted diagnosis coding demonstrate reliable performance, though human clinicians retain final authority over referral decisions.
Technical feasibility todayclaude-sonnet-52/5Some clinical decision-support tools can suggest referral resources or flag risk, but no deployed product autonomously manages the clinical judgment and relational aspects of referral at scale.

Maintain current knowledge of relevant research.

46

CI 32–59 · exposure 42 · augmentation 88 · importance 4.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510013/5Many clinical psychologists use AI tools for research discovery and summarization as supplements, but meaningful adoption of AI-driven knowledge maintenance remains uneven across practice settings. Academic and larger organizations lead, while solo practitioners lag.
Sector adoption velocityclaude-sonnet-53/5Healthcare and academic/clinical psychology sectors show moderate AI tool adoption for research support, with growing but not yet ubiquitous use of AI literature tools among practitioners.
Augmentation potentialclaude-haiku-4-5-202510014/5AI significantly enhances productivity by rapidly filtering and summarizing vast literature, flagging relevant studies, and organizing findings—allowing psychologists to focus on critical appraisal and clinical integration. This is a strong augmentation use case within professional bounds.
Augmentation potentialclaude-sonnet-55/5AI significantly boosts efficiency in staying current by filtering, summarizing, and flagging relevant new studies, letting psychologists focus limited time on the most pertinent findings.
Task automatabilityclaude-haiku-4-5-202510012/5AI can retrieve and summarize research papers, but the task requires critical evaluation of methodologies, contextual judgment about relevance to specific clinical populations, and integration with prior knowledge—functions that demand human expertise. AI can assist with literature discovery and synthesis but cannot independently maintain clinical currency.
Task automatabilityclaude-sonnet-53/5AI can summarize and surface new research (literature reviews, alerts, synthesis of journal articles), covering much of the discovery and synthesis work, but true 'currency' requires judgment about clinical relevance and integration into practice that AI cannot fully replace.
Adoption barriersclaude-haiku-4-5-202510014/5Professional licensure and ethics codes require psychologists to maintain competence; regulatory bodies expect direct engagement with research evidence. Liability exposure for relying solely on automated summaries, combined with professional duty to stay current, creates strong barriers to full automation.
Adoption barriersclaude-sonnet-52/5No licensing requirement mandates a human personally read all research, but professional ethics and continuing education requirements create some expectation of individual engagement with the literature.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI-powered research tools are relatively inexpensive, but when factoring in required human oversight to verify accuracy and relevance, the all-in cost remains high relative to the modest time savings on initial screening and summarization.
Cost vs. human wageclaude-sonnet-54/5AI-based literature scanning and summarization tools are inexpensive subscriptions compared to the hours a psychologist would spend manually reading and tracking journals.
Technical feasibility todayclaude-haiku-4-5-202510013/5Tools like literature search engines, summarization APIs, and AI-powered research platforms exist and are used, but they produce variable quality, miss nuanced relevance judgments, and require substantial human validation. No deployed system fully replaces a psychologist's own reading and critical appraisal.
Technical feasibility todayclaude-sonnet-53/5Products like AI-powered literature search, summarization tools, and research digest services (e.g., Elicit, Consensus, PubMed AI tools) are deployed and used by professionals today, though accuracy and completeness vary and require verification.

Write reports on clients and maintain required paperwork.

44

CI 25–62 · exposure 45 · augmentation 75 · importance 4.6/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Clinical psychology remains a regulated, human-centered discipline with strong professional norms favoring human accountability. While EHR adoption is growing, use of AI for autonomous report generation is still uncommon and cautious in production practice.
Sector adoption velocityclaude-sonnet-53/5Healthcare/behavioral health lags top-adopting sectors like finance or tech, but AI scribe tools are seeing growing pilot and production use in clinical practices.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist by drafting templates, summarizing test data, or suggesting report structure, which may reduce time spent on routine formatting and clerical elements. However, the core diagnostic and clinical reasoning must remain with the psychologist, limiting the augmentation ceiling.
Augmentation potentialclaude-sonnet-55/5AI can substantially speed up drafting of session notes and reports from clinician input or recordings, letting the psychologist focus on review and clinical accuracy rather than writing from scratch.
Task automatabilityclaude-haiku-4-5-202510012/5AI can draft routine sections of clinical reports (demographics, test summaries) but cannot replace the core judgment required to synthesize client presentation, formulate diagnostic impressions, or document clinical reasoning without substantial human review. The legal and clinical accountability attached to psychologist-authored reports prevents full automation.
Task automatabilityclaude-sonnet-54/5Drafting clinical reports and progress notes from session data/transcripts is largely templatable and current LLMs can generate compliant drafts, though a clinician must review and finalize for accuracy and clinical judgment.'},'
Adoption barriersclaude-haiku-4-5-202510014/5Psychologists are legally and ethically required to author, review, and sign clinical reports; liability for misstatement or inadequacy falls directly on the licensed provider. Professional standards and liability frameworks create hard barriers to full automation or delegation to unlicensed AI systems.
Adoption barriersclaude-sonnet-53/5Clinical documentation must reflect the licensed clinician's professional judgment and is subject to confidentiality (HIPAA) and liability concerns, requiring human sign-off even if drafting is assisted.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI drafting tools are inexpensive, but the required clinician time to review, modify, and sign off on AI-generated clinical documentation may not materially reduce the loaded wage cost of the psychologist's time spent on report writing.
Cost vs. human wageclaude-sonnet-54/5AI-assisted documentation tools cost a small fraction of clinician time saved on paperwork, offering substantial per-note cost savings even after review overhead.
Technical feasibility todayclaude-haiku-4-5-202510012/5While AI writing tools exist and some EHR systems offer templating, no deployed product reliably produces clinically defensible or compliant psychological reports without expert human revision. Products that draft sections exist but require substantial clinician oversight and modification to meet standards.
Technical feasibility todayclaude-sonnet-53/5AI scribe and note-generation tools are deployed in behavioral health settings today, but adoption is uneven and outputs still require clinician editing for accuracy and liability reasons.

Provide occupational, educational, or other information to individuals so that they can make educational or vocational plans.

42

CI 32–51 · exposure 38 · augmentation 75 · importance 3.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Adoption of AI for vocational counseling remains limited to small pilots and supplementary tools within existing practices. Psychology and counseling are human-contact-dependent fields with slow digitization relative to information sectors, and trust in the profession centers on the licensed practitioner relationship.
Sector adoption velocityclaude-sonnet-52/5Mental health and counseling fields have been slower to adopt AI tools compared to sectors like finance or general information services, partly due to trust, regulatory, and ethical concerns.
Augmentation potentialclaude-haiku-4-5-202510014/5AI can substantially assist psychologists by retrieving occupation data, summarizing labor-market trends, and organizing educational pathways, allowing the human to focus on interpreting those options in light of the client's values and circumstances. Such assistance raises productivity while the psychologist retains decision-making authority.
Augmentation potentialclaude-sonnet-54/5AI can efficiently research, summarize, and present occupational and educational data, significantly speeding up the informational component of a psychologist's guidance work while they retain the interpretive and relational role.
Task automatabilityclaude-haiku-4-5-202510012/5AI can retrieve and present educational/vocational information at scale, but the core task requires understanding individual circumstances, aspirations, and constraints in context—elements that demand human judgment and rapport. Current systems cannot reliably conduct the discovery and personalization needed for effective career planning.
Task automatabilityclaude-sonnet-53/5AI can compile and present occupational/educational information effectively, but tailoring it to an individual's psychological profile, motivations, and context within a counseling relationship requires human judgment that current systems only partially replicate.
Adoption barriersclaude-haiku-4-5-202510014/5Significant barriers exist: clients typically seek counseling from licensed professionals for trust and accountability; liability concerns attach to vocational guidance that goes wrong; and organizational/insurance structures assume human provider involvement. Regulatory framework does not yet permit autonomous AI to substitute for licensed psychologist recommendations.
Adoption barriersclaude-sonnet-53/5This sub-task itself isn't strictly licensed, but it typically occurs within a broader clinical relationship where the psychologist's credentialed judgment and liability considerations create moderate friction against pure automation.
Cost vs. human wageclaude-haiku-4-5-202510014/5Information-delivery via AI (chatbot, knowledge base) costs a fraction of a human psychologist's billable hour, and at scale the marginal cost per interaction is minimal, making AI potentially an order of magnitude cheaper for the pure information component.
Cost vs. human wageclaude-sonnet-54/5Generating occupational/educational information via AI is extremely cheap compared to a licensed psychologist's time, even though the psychologist must still integrate it with clinical judgment.
Technical feasibility todayclaude-haiku-4-5-202510012/5While chatbots can dispense generic career information, no deployed product reliably performs the full task (discovery, assessment, contextual advice, and follow-up) with the quality expected in clinical/counseling psychology. Demos exist but real therapeutic/counseling settings still require human psychologists.
Technical feasibility todayclaude-sonnet-53/5Career guidance chatbots and AI-assisted counseling tools exist and are used for information delivery, but psychologist-delivered integrated advice combining clinical insight with vocational planning is not reliably automated in production.

Consult reference material, such as textbooks, manuals, or journals, to identify symptoms, make diagnoses, or develop approaches to treatment.

36

CI 25–46 · exposure 42 · augmentation 75 · importance 4.0/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5While healthcare is digitizing, clinical psychology adoption of autonomous AI decision systems remains limited due to regulatory constraints, liability concerns, and professional norms emphasizing human judgment. Most deployments are assistive rather than replacement.
Sector adoption velocityclaude-sonnet-52/5Healthcare and mental health fields have historically been slower and more cautious adopters of AI tools compared to sectors like finance or general professional services, due to regulatory, liability, and trust concerns.
Augmentation potentialclaude-haiku-4-5-202510014/5AI-powered literature search, symptom matching, and evidence-based treatment suggestions significantly augment a clinician's ability to access and synthesize current knowledge quickly, allowing the psychologist to focus on patient interaction and contextual judgment.
Augmentation potentialclaude-sonnet-54/5AI tools can quickly surface relevant literature, summarize diagnostic criteria (e.g., DSM/ICD), and suggest evidence-based treatment options, meaningfully speeding up a clinician's research and decision support while the clinician retains final judgment.
Task automatabilityclaude-haiku-4-5-202510012/5AI can retrieve and summarize relevant literature quickly, but diagnosis and treatment planning require integration of patient-specific context, clinical judgment, and nuanced interpretation that current systems cannot reliably perform end-to-end. The task demands synthesis of complex, multi-faceted clinical information rather than pure information retrieval.
Task automatabilityclaude-sonnet-53/5AI can rapidly search and synthesize reference materials to suggest differential diagnoses or treatment approaches, but a licensed psychologist must still verify and apply clinical judgment, so only part of this research/consultation subtask is automatable at equal quality.
Adoption barriersclaude-haiku-4-5-202510015/5Diagnosis and treatment planning are core responsibilities regulated by licensing law; a licensed psychologist must legally make the diagnostic and treatment decisions. Malpractice liability, standard-of-care requirements, and patient safety create hard barriers to full automation.
Adoption barriersclaude-sonnet-54/5Diagnosis and treatment planning are subject to licensure, malpractice liability, and professional standards requiring a qualified human to make and be accountable for clinical decisions, strongly limiting full automation even though reference lookup itself is low-risk.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI literature search and summarization is inexpensive, but the human psychologist's time remains necessary for clinical judgment and patient interaction. The cost of AI oversight and validation of AI outputs may approach or exceed the time saved by automation.
Cost vs. human wageclaude-sonnet-54/5Querying an AI system for symptom/diagnosis/treatment references costs a fraction of a cent to a few dollars per query versus the loaded hourly cost of a psychologist's research time, making AI substantially cheaper for the information-retrieval portion.
Technical feasibility todayclaude-haiku-4-5-202510013/5Products like clinical decision support systems and AI-assisted literature search tools exist and are deployed in some healthcare settings, but they typically support rather than replace the clinician and have material error rates in diagnosis or treatment recommendation. Fully autonomous diagnostic systems remain research-stage or narrowly scoped.
Technical feasibility todayclaude-sonnet-53/5Deployed clinical decision-support and LLM-based literature tools exist and are used for reference lookup and drafting, but they are not fully reliable for autonomous diagnostic reasoning and are typically used as adjuncts rather than primary tools.

Plan, supervise, and conduct psychological research and write papers describing research results.

30

CI 28–32 · exposure 25 · augmentation 75 · importance 2.9/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510013/5Academic and clinical psychology labs increasingly use AI for literature review, statistical analysis, and manuscript drafting, but adoption remains in the pilot and support phase. Human oversight of research design and conduct is not being displaced at scale.
Sector adoption velocityclaude-sonnet-53/5Academic and clinical research settings are adopting AI writing and analysis tools at a moderate pace, with pilots common but full-scale production automation of research supervision still limited.
Augmentation potentialclaude-haiku-4-5-202510014/5AI significantly augments psychologists' productivity in literature synthesis, data analysis, statistical visualization, and manuscript drafting. These tools allow researchers to focus more time on conceptual design, interpretation, and ethical oversight rather than clerical work.
Augmentation potentialclaude-sonnet-54/5AI substantially aids literature synthesis, drafting, statistical coding, and editing, meaningfully boosting researcher productivity while the psychologist retains responsibility for design and interpretation.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist with literature review, data analysis, and initial drafting of papers, the core work—designing rigorous research protocols, supervising human subjects, interpreting nuanced psychological findings, and making methodological decisions—requires human expertise and accountability. AI cannot independently plan ethically sound research or supervise participants.
Task automatabilityclaude-sonnet-52/5AI can assist with literature review, drafting, and statistical analysis, but planning novel research, exercising scientific judgment, and supervising human researchers require domain expertise and oversight that current AI cannot fully replace end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Institutional Review Boards (IRBs) legally require human researchers to design protocols, obtain informed consent, and oversee participant welfare. Professional licensure and liability for research conduct create hard barriers to full automation or unsupervised AI deployment.
Adoption barriersclaude-sonnet-53/5No licensing law mandates a human write research papers, but ethical/IRB requirements, authorship responsibility, and academic integrity norms create moderate friction against full AI substitution.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI tools for data analysis and writing assistance are cheap per use, but they supplement rather than replace the expensive labor of research design, participant management, and expert interpretation. The all-in cost remains dominated by the psychologist's time and liability.
Cost vs. human wageclaude-sonnet-52/5AI can cheaply produce drafts and analyses, but the overall task still requires expensive expert oversight, supervision of staff, and research design judgment, keeping costs comparable to human-led work for the full task.
Technical feasibility todayclaude-haiku-4-5-202510012/5Products exist for literature mining and statistical analysis, but no deployed system reliably performs end-to-end psychological research planning, supervision, and write-up. IRB approval, participant safety oversight, and clinical judgment remain human-required functions that AI cannot substitute.
Technical feasibility todayclaude-sonnet-52/5AI writing/analysis tools are used in research support (e.g., drafting assistance, data analysis) but no deployed product independently plans and supervises psychological research programs reliably.

Advise clients on how they could be helped by counseling.

27

CI 25–29 · exposure 25 · augmentation 50 · importance 3.9/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Mental health and psychology sectors have been cautious and slow in AI adoption; most deployment remains research pilots or indirect support (administrative) rather than frontline clinical decision-making. Regulatory, liability, and trust barriers limit production use.
Sector adoption velocityclaude-sonnet-52/5placeholder
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist clinicians by surfacing relevant treatment modalities, generating initial assessments of common presentations, or organizing client information—useful support tools. However, the core task of advising clients requires human clinical judgment, so augmentation is meaningful but not transformative.
Augmentation potentialclaude-sonnet-53/5placeholder
Task automatabilityclaude-haiku-4-5-202510012/5While AI can provide generic information about counseling modalities and benefits, this task requires understanding individual client circumstances, needs, and therapeutic fit—nuanced judgment that current systems struggle with reliably. Limited segments (psychoeducational guidance, referral matching) could be partially automated, but genuine counseling-fit assessment remains primarily human.
Task automatabilityclaude-sonnet-52/5This requires building trust, assessing individual clinical needs, and making nuanced judgments about treatment fit that current AI cannot reliably replicate end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Clinical judgment about treatment suitability often requires licensure and carries liability risk; improper advice can cause harm. Regulatory frameworks and professional standards expect human clinician accountability, and many jurisdictions legally require a licensed psychologist for treatment recommendations.
Adoption barriersclaude-sonnet-54/5placeholder
Cost vs. human wageclaude-haiku-4-5-202510012/5Basic AI-driven psychoeducation is cheaper than human clinician time, but for the full task of meaningful advice tailored to an individual's presentation, oversight and integration costs remain high relative to modest automation gains. Cost advantage is marginal.
Cost vs. human wageclaude-sonnet-53/5placeholder
Technical feasibility todayclaude-haiku-4-5-202510012/5Chatbots and mental health apps exist that offer counseling information and basic screening, but they operate in narrow, controlled contexts and cannot reliably assess complex client needs or therapeutic suitability. No mature product performs this task at the depth and accuracy expected in clinical practice.
Technical feasibility todayclaude-sonnet-52/5placeholder

Obtain and study medical, psychological, social, and family histories by interviewing individuals, couples, or families and by reviewing records.

25

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

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Mental health care remains heavily regulated with strong professional gatekeeping; adoption of AI for core clinical interviewing and history-taking is minimal in production settings. Most adoption remains limited to administrative tasks like scheduling or notes summarization, not the core assessment function.
Sector adoption velocityclaude-sonnet-52/5Healthcare and mental health services are relatively slow adopters of AI for core clinical interactions, though administrative and documentation support is growing.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can meaningfully assist by transcribing interviews, flagging relevant prior medical records, and suggesting follow-up questions based on incomplete information, but the human clinician must remain central to the diagnostic relationship. Tools that organize and synthesize historical data do provide modest productivity gains.
Augmentation potentialclaude-sonnet-54/5AI tools (ambient scribes, transcription, summarization, record review assistants) meaningfully speed up documentation and history review, letting clinicians focus more on the interview itself.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can extract and summarize information from medical records automatically, conducting effective clinical interviews requires nuanced rapport-building, real-time clinical judgment, and adaptive questioning that current systems cannot reliably perform. The task fundamentally involves interpersonal assessment and integration of complex contextual information where AI falls short of the 50% time-saving threshold.
Task automatabilityclaude-sonnet-52/5AI can help structure intake forms and summarize records, but conducting a nuanced clinical interview to elicit sensitive history requires rapport, judgment, and adaptive follow-up that current systems cannot reliably replicate end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Significant regulatory and professional barriers exist: clinical assessment and diagnosis require licensed mental health professionals who must legally obtain informed consent, maintain therapeutic relationship, and be held liable for assessment accuracy. Liability asymmetry and ethical requirements create strong legal and organizational friction against full automation.
Adoption barriersclaude-sonnet-54/5Licensure, confidentiality (HIPAA), and clinical liability strongly favor a human psychologist conducting and interpreting history-taking interviews, especially for diagnosis-relevant information.
Cost vs. human wageclaude-haiku-4-5-202510012/5Integration costs for secure health record access, clinical-grade quality control, and necessary human oversight make AI solutions comparable to or more expensive than human clinicians for this sensitive task. The liability and error-cost burden further increases effective AI expenses.
Cost vs. human wageclaude-sonnet-52/5AI can cheaply assist with documentation and pre-screening, but the interviewing and clinical judgment portions still require a licensed professional, keeping overall cost savings modest.
Technical feasibility todayclaude-haiku-4-5-202510012/5AI can assist with record review and transcription, but no deployed product reliably conducts or replaces clinical interviews with the therapeutic alliance and diagnostic accuracy required in real clinical settings. Current systems produce high error rates in complex case formulation and miss critical contextual cues essential to psychological assessment.
Technical feasibility todayclaude-sonnet-52/5Some intake chatbots and transcription/summarization tools are deployed in clinical settings, but no product reliably conducts full psychosocial history-taking interviews independently at scale.

Develop, direct, and participate in training programs for staff and students.

25

CI 25–25 · exposure 25 · augmentation 75 · importance 3.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5While some universities and training programs use AI for supplementary materials and scheduling, core training and supervision remain dominated by human psychologists due to professional norms, accreditation standards, and the critical need for personalized mentoring in clinical skill development.
Sector adoption velocityclaude-sonnet-52/5Healthcare and clinical training environments are historically slow AI adopters due to regulatory, ethical, and accreditation constraints, with pilots more common than production deployment.
Augmentation potentialclaude-haiku-4-5-202510014/5AI can substantially augment this task by generating draft training content, summarizing research literature, organizing schedules, and providing supplemental learning materials, allowing psychologist-instructors to focus more on live facilitation, feedback, and mentoring.
Augmentation potentialclaude-sonnet-54/5AI tools can meaningfully assist in generating training curricula, case simulations, quizzes, and administrative materials, improving efficiency while the psychologist remains the direct trainer/supervisor.
Task automatabilityclaude-haiku-4-5-202510012/5AI can assist with content generation, scheduling, and administrative logistics for training programs, but the task of developing coherent curricula, directing live instruction, and providing personalized feedback requires human expertise and judgment that current systems cannot replace end-to-end at equal quality.
Task automatabilityclaude-sonnet-52/5Designing and delivering training programs for staff/students involves curriculum design, live facilitation, mentorship, and adaptive teaching that current AI cannot fully replace, though AI can help draft materials.rating reflects partial support only.
Adoption barriersclaude-haiku-4-5-202510014/5Regulatory and licensing frameworks require that psychology training, particularly clinical supervision, be delivered by qualified licensed professionals who bear responsibility for student competency and patient safety; delegation to AI is legally and ethically constrained.
Adoption barriersclaude-sonnet-54/5Training future clinicians/counselors typically requires licensed, credentialed supervisors per accreditation and licensing board standards, creating strong regulatory and professional barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI can reduce content creation time and administrative overhead, but the ongoing cost of human instructors, curriculum oversight, and personalized mentoring remains the dominant expense, limiting AI cost advantage.
Cost vs. human wageclaude-sonnet-52/5AI can cheaply produce draft materials, but the substantial human oversight, live instruction, and supervision components mean overall cost savings versus a psychologist's time are modest.
Technical feasibility todayclaude-haiku-4-5-202510012/5Deployed products can generate training materials and handle scheduling, but no mature system reliably directs comprehensive staff/student training programs or manages the interpersonal and pedagogical nuances required in psychology training at scale.
Technical feasibility todayclaude-sonnet-52/5Products exist for generating training content, slides, and quizzes, but no deployed system reliably directs or delivers full clinical/counseling training programs including supervision and interpersonal skill-building.

Provide consulting services, including educational programs, outreach programs, or prevention talks to schools, social service agencies, businesses, or the general public.

25

CI 25–25 · exposure 25 · augmentation 75 · importance 3.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Adoption of AI for live consulting or public-facing prevention programs remains minimal; most organizations view this as a core human service. Pilots exist for content support, but production displacement is rare and slow in sectors that value clinical credentials and direct human contact.
Sector adoption velocityclaude-sonnet-52/5Psychology and social services sectors are slower AI adopters, especially for public-facing advisory and educational engagements which remain largely human-delivered.
Augmentation potentialclaude-haiku-4-5-202510014/5AI can significantly assist by drafting talking points, generating customized educational materials, summarizing research for audience-appropriate language, and helping organize content—allowing psychologists to focus on delivery, engagement, and real-time responsiveness. This augmentation is high even though full automation is not feasible.
Augmentation potentialclaude-sonnet-54/5AI can meaningfully assist with drafting presentations, educational content, and research summaries, enhancing psychologists' productivity in preparing outreach programs.
Task automatabilityclaude-haiku-4-5-202510012/5Educational and prevention talks require nuanced audience engagement, real-time adaptation to emotional cues, and credibility-building that current AI systems cannot reliably deliver end-to-end. While AI can assist with content drafting and slide generation, the live delivery, personal connection, and clinical judgment in responding to audience needs remain beyond current automation.
Task automatabilityclaude-sonnet-52/5AI can help draft materials and content for these programs, but delivering consulting, outreach, and prevention talks requires live human presence, adaptability, and credibility that current AI cannot replace end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Consulting and educational talks, especially in clinical and school contexts, often involve implicit or explicit expectations that a licensed clinical psychologist leads the engagement. Liability, institutional credibility, and client trust create strong friction against full automation; regulatory bodies and schools typically require human accountability.
Adoption barriersclaude-sonnet-54/5Professional licensure, liability for psychological advice, and institutional preference for credentialed human experts create strong barriers to full automation of consulting/outreach roles.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI-assisted content creation has low per-unit cost, but integrating it into a consulting program (with human oversight and follow-up) still requires substantial human expertise. The all-in cost for a viable consulting service remains comparable to or higher than a mid-career psychologist's billable rate.
Cost vs. human wageclaude-sonnet-52/5While AI can cheaply generate supporting content, the actual delivery and consulting relationship still requires paid human time, limiting overall cost savings.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed products reliably conduct full consulting engagements or deliver talks as trained psychologists in production settings. Chatbots can provide educational content, but cannot substitute for the clinical presence, adaptability, and accountability required in actual outreach programs.
Technical feasibility todayclaude-sonnet-52/5AI content-generation tools exist for drafting presentations or educational materials, but no deployed product independently delivers psychological consulting or outreach talks to institutions.

Collect information about individuals or clients, using interviews, case histories, observational techniques, and other assessment methods.

23

CI 20–25 · exposure 25 · augmentation 63 · importance 4.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Despite digitization in healthcare, clinical psychology has been slow to adopt fully automated intake systems. Adoption remains largely pilot-stage with heavy human oversight; organizational culture and risk-aversion in mental health settings slow velocity.
Sector adoption velocityclaude-sonnet-52/5Mental health care remains a relatively low-digitization, high-touch field where AI adoption for core clinical assessment is still in early pilot stages, not widespread production use.
Augmentation potentialclaude-haiku-4-5-202510014/5AI substantially augments clinicians through automated transcription, structured note-taking, pattern flagging in symptom data, and literature retrieval to inform assessment. These tools meaningfully accelerate data organization and hypothesis generation while the clinician retains full diagnostic and relational authority.
Augmentation potentialclaude-sonnet-53/5AI can help with structured intake questionnaires, transcription, and summarizing case histories, offering moderate productivity gains while the psychologist remains central to the interview and judgment process.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist in transcribing interviews and structuring case history data, it cannot reliably conduct the nuanced, adaptive interpersonal interactions required for clinical assessment. Human psychologists must read nonverbal cues, build rapport, and adjust questioning in real-time—capabilities that current AI systems lack, making end-to-end automation with 50% time savings infeasible.
Task automatabilityclaude-sonnet-52/5Clinical interviewing and observational assessment require real-time rapport, nonverbal cue reading, and adaptive follow-up questioning that current AI cannot reliably replicate end-to-end; at best AI can assist with intake forms or transcription.atability is limited beyond partial documentation support.
Adoption barriersclaude-haiku-4-5-202510015/5Strong legal and licensing barriers exist: only licensed psychologists can conduct clinical assessments in most jurisdictions, and liability for misdiagnosis or missed clinical risk falls directly on the credentialed clinician. Regulatory frameworks require human professional judgment and accountability.
Adoption barriersclaude-sonnet-54/5Clinical assessment is typically bound by licensure, ethical codes, and legal standards requiring a qualified professional to gather and interpret client information, creating strong regulatory and liability barriers.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI-assisted tools (transcription, note generation) reduce overhead costs but require significant human oversight and validation. The total cost of AI infrastructure plus necessary clinician supervision likely approaches or exceeds the cost of direct clinical assessment by trained staff.
Cost vs. human wageclaude-sonnet-52/5While automated screening tools are cheap to run, they cannot substitute for the full clinical interview, so any real cost comparison must include the human clinician's time for verification and deeper assessment, keeping the ratio modest.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed product reliably performs independent clinical intake or assessment interviews at production scale. AI tools can assist with documentation and transcription, but they cannot replace the clinician's real-time judgment, therapeutic relationship, and diagnostic decision-making that define this task.
Technical feasibility todayclaude-sonnet-52/5Some chatbot-based intake tools and symptom-screening apps exist, but they are narrow, not validated for full clinical assessment, and rarely used as the primary data-gathering method in licensed practice.

Develop therapeutic and treatment plans based on clients' interests, abilities, or needs.

23

CI 20–25 · exposure 25 · augmentation 63 · importance 4.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Healthcare and mental health sectors remain slower in AI adoption due to regulatory constraints, liability concerns, and cultural preference for human clinician ownership of treatment decisions; pilots exist but production deployment of autonomous or semi-autonomous planning remains rare.
Sector adoption velocityclaude-sonnet-52/5Mental health care remains a relatively low-digitization, high-touch sector with cautious, slow adoption of AI for core clinical decisions, though administrative AI tools are spreading faster.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist psychologists by drafting preliminary plans, organizing client data, and suggesting evidence-based interventions, moderately raising efficiency—but the psychologist retains full responsibility for clinical formulation and plan adequacy.
Augmentation potentialclaude-sonnet-54/5AI can meaningfully assist by synthesizing client information, suggesting evidence-based interventions, and speeding documentation, enhancing clinician efficiency while the psychologist retains final judgment.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can generate template treatment plans and summarize client information, developing truly personalized therapeutic plans requires nuanced clinical judgment, understanding of complex psychological dynamics, and real-time adaptation based on client presentation—tasks current AI cannot perform reliably end-to-end at equal quality with 50% time savings.
Task automatabilityclaude-sonnet-52/5AI can draft generic treatment plan templates from intake notes, but individualized clinical judgment integrating client history, risk, and therapeutic rapport is not reliably automatable end-to-end today.
Adoption barriersclaude-haiku-4-5-202510014/5Significant legal and regulatory barriers exist: licensed psychologists must develop or formally approve treatment plans; malpractice liability attaches to the clinician, not the AI; and professional standards require human clinical judgment for care planning.
Adoption barriersclaude-sonnet-55/5Treatment planning is a core clinical function typically requiring a licensed psychologist's judgment and sign-off, with strong liability, ethical, and regulatory constraints against full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI infrastructure costs (model access, integration, oversight workflows) remain comparable to or potentially exceed the hourly cost of a psychologist drafting a plan, especially when accounting for required human review and liability.
Cost vs. human wageclaude-sonnet-52/5AI drafting assistance is cheap, but the human clinician still must review, adapt, and take responsibility, so overall cost savings versus the licensed professional's time are modest.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed clinical product reliably generates treatment plans independent of psychologist oversight; research prototypes and decision-support tools exist but require substantial human validation and modification before clinical use.
Technical feasibility todayclaude-sonnet-52/5Some AI documentation/decision-support tools exist that suggest treatment plan elements, but no deployed product independently creates clinically sound, personalized treatment plans without clinician authorship.

Select, administer, score, and interpret psychological tests to obtain information on individuals' intelligence, achievements, interests, or personalities.

23

CI 20–25 · exposure 25 · augmentation 63 · importance 4.0/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Healthcare and clinical psychology remain relatively low-adoption sectors for automation due to regulatory constraints, professional licensing requirements, and liability aversion. While EHR integration of scoring tools is growing, clinicians retain skepticism about delegating judgment to AI, and adoption remains pilot-stage rather than mainstream.
Sector adoption velocityclaude-sonnet-52/5Healthcare and clinical psychology adopt digital tools slowly due to regulatory, liability, and training constraints, with AI-assisted scoring seeing modest uptake but full task automation rare.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist by automating score calculation, generating normative comparisons, and surfacing relevant test patterns for clinician review, improving efficiency in the clerical and interpretive preparation phases. However, the augmentation is limited to support functions; the clinician must still own final interpretation and test selection decisions.
Augmentation potentialclaude-sonnet-54/5AI-assisted scoring, normative comparisons, and report drafting meaningfully speed up psychologists' workflow while they retain responsibility for test selection and clinical interpretation.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can score standardized tests and generate initial interpretations from test results, the selection of appropriate tests, adaptive administration based on client response, and nuanced interpretation requiring clinical judgment remain largely human-dependent. The task critically depends on contextual clinical decision-making that AI cannot reliably perform end-to-end at equal quality.
Task automatabilityclaude-sonnet-52/5Scoring standardized tests can be automated, but selecting appropriate instruments, administering them with clinical rapport, and interpreting results in context of a client's history require professional judgment AI cannot yet replicate end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Significant regulatory and liability barriers exist: psychologists must be licensed to administer and interpret psychological tests, and test publishers impose strict licensing requirements. Clinical liability for misinterpretation and incorrect test selection creates strong asymmetric error costs, and professional standards (APA) legally bind practitioners to personal accountability for these decisions.
Adoption barriersclaude-sonnet-55/5Psychological testing and interpretation is a licensed clinical activity with strict ethical, legal, and professional standards requiring a qualified psychologist to administer and sign off on results.
Cost vs. human wageclaude-haiku-4-5-202510012/5Automated scoring reduces clerical costs, but the professional clinical expertise required for test selection and interpretation remains expensive. The all-in cost of an AI system with required oversight and validation is comparable to or higher than the time saved in scoring alone, given low volume per clinician.
Cost vs. human wageclaude-sonnet-52/5Scoring software is cheap, but the interpretive and administrative components still require paid clinician time, keeping overall cost comparable to or only modestly below human-only delivery.
Technical feasibility todayclaude-haiku-4-5-202510012/5Scoring algorithms for standardized tests exist and work reliably, but no mature product currently handles full end-to-end test selection, adaptive administration, and clinically-valid interpretation in production clinical settings. Deployed systems are limited to scoring; the interpretive and selection components remain primarily manual.
Technical feasibility todayclaude-sonnet-52/5Software exists for automated scoring and even computer-adaptive testing, but no deployed product reliably selects instruments and interprets results holistically without licensed psychologist oversight.

Identify psychological, emotional, or behavioral issues and diagnose disorders, using information obtained from interviews, tests, records, or reference materials.

20

CI 20–20 · exposure 25 · augmentation 75 · 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 mental health sectors show slow, cautious adoption of AI tools for diagnosis due to liability concerns, professional skepticism, and regulatory scrutiny. Most adoption remains pilot-stage or limited to screening, not diagnosis.
Sector adoption velocityclaude-sonnet-52/5Healthcare/behavioral health is a historically slow-adopting, highly regulated sector; AI pilots exist for screening but production-level diagnostic use is rare.
Augmentation potentialclaude-haiku-4-5-202510014/5AI can meaningfully assist psychologists by surfacing patterns in interview transcripts, flagging relevant diagnostic criteria, summarizing test results, and retrieving reference materials, thereby increasing clinician efficiency while the expert remains central to decision-making.
Augmentation potentialclaude-sonnet-54/5AI can meaningfully assist by analyzing intake data, summarizing records, flagging risk indicators, and suggesting differential diagnoses for clinician review, enhancing efficiency and thoroughness.
Task automatabilityclaude-haiku-4-5-202510012/5AI can assist with pattern recognition in test data and interview transcripts, but diagnosis requires clinical judgment, nuanced interpretation of context, and integration of multiple information sources that current systems struggle with reliably. The task involves significant human interpretation and accountability that cannot be fully automated while maintaining diagnostic accuracy.
Task automatabilityclaude-sonnet-52/5Diagnosis requires nuanced clinical interviewing, observation of affect/behavior, and integration of context that current AI cannot reliably replicate end-to-end, though it can assist with structured intake and screening data synthesis.'
Adoption barriersclaude-haiku-4-5-202510015/5Clinical diagnosis requires a licensed mental health professional to legally perform and sign off on the assessment; liability, regulatory requirements (state licensing boards), and ethical standards create hard legal barriers to AI autonomy in this task.
Adoption barriersclaude-sonnet-55/5Diagnosing psychological disorders is a licensed clinical act with strict legal, ethical, and liability requirements mandating a qualified professional to perform and sign off on the diagnosis.
Cost vs. human wageclaude-haiku-4-5-202510012/5Current AI diagnostic support systems require clinician review, integration with EHR systems, and ongoing oversight, making all-in costs comparable to or higher than the clinician's time savings. Full replacement is not feasible, limiting cost advantage.
Cost vs. human wageclaude-sonnet-52/5AI screening tools are cheap to run, but since they cannot independently produce a valid diagnosis, the human clinician's time is still required, keeping overall cost comparable to fully human-delivered care.
Technical feasibility todayclaude-haiku-4-5-202510012/5While AI tools exist for symptom screening and preliminary assessment support, no deployed product reliably performs full diagnostic formulation at the standard required for clinical practice. Products show material error rates and typically function as decision aids rather than independent diagnostic systems.
Technical feasibility todayclaude-sonnet-52/5Some AI-based screening tools (e.g., depression/anxiety chatbots, symptom checkers) exist but are not deployed as primary diagnostic instruments in clinical practice; formal diagnosis remains clinician-driven.

Evaluate the effectiveness of counseling or treatments and the accuracy and completeness of diagnoses, modifying plans or diagnoses as necessary.

20

CI 20–20 · exposure 25 · augmentation 63 · importance 4.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Mental health and clinical psychology sectors have adopted AI tools for administrative tasks and supportive analytics, but actual substitution in diagnosis and treatment planning remains rare; organizational and regulatory caution keeps automation to pilots and augmentation rather than replacement.
Sector adoption velocityclaude-sonnet-52/5Mental health care remains a high-touch, heavily regulated sector with slow, cautious AI adoption limited mostly to administrative and screening support rather than diagnostic judgment.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can meaningfully assist by summarizing patient progress data, flagging outcome metrics, and suggesting diagnostic or treatment considerations, but the clinician must interpret context and make final judgments, offering useful but not transformative productivity gains.
Augmentation potentialclaude-sonnet-54/5AI can meaningfully assist by summarizing session notes, tracking symptom trends, cross-referencing diagnostic criteria, and flagging inconsistencies, improving clinician efficiency while the clinician retains final judgment.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist in analyzing symptom patterns and treatment outcomes data, the task fundamentally requires nuanced clinical judgment, contextual understanding of individual patient circumstances, and the ability to modify complex treatment plans—capabilities current systems cannot reliably perform end-to-end with equal quality at 50% time savings.
Task automatabilityclaude-sonnet-52/5This requires integrating clinical judgment, patient history, observed behavior, and ethical responsibility for treatment decisions—current AI can assist with data synthesis but cannot autonomously evaluate treatment efficacy or revise diagnoses with equal quality.
Adoption barriersclaude-haiku-4-5-202510015/5Strong regulatory and legal barriers exist: only licensed psychologists can render formal diagnoses and modify treatment plans; malpractice liability, state licensing requirements, and standard-of-care expectations place the human clinician in non-delegable positions of authority.
Adoption barriersclaude-sonnet-55/5Diagnosis and treatment modification are legally restricted to licensed psychologists/psychiatrists, with strong liability, ethical, and regulatory requirements for human sign-off.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI systems for mental health assessment and outcome tracking exist but require significant human oversight, training, and integration costs that approach or exceed the cost of a clinician's direct time on case review and diagnostic modification.
Cost vs. human wageclaude-sonnet-52/5AI tools can cheaply process notes and flag patterns, but the human oversight, liability, and clinical judgment required keep all-in costs comparable to or only modestly below human-only review.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed AI product reliably performs independent evaluation and modification of diagnoses or treatment plans in clinical practice; existing tools may flag patterns or suggest considerations but clinical psychologists must retain full evaluative authority due to liability and clinical complexity.
Technical feasibility todayclaude-sonnet-52/5Some clinical decision-support tools exist to flag inconsistencies or suggest differential diagnoses, but no deployed product independently evaluates treatment effectiveness and revises psychological diagnoses in production settings.

Conduct research to develop or improve diagnostic or therapeutic counseling techniques.

18

CI 11–25 · exposure 13 · augmentation 63 · importance 3.1/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Academic psychology and clinical research sectors adopt AI for auxiliary functions (data processing, literature management) slowly and cautiously. The core task of developing new techniques remains researcher-driven, with limited evidence of production-level AI adoption for the generative, conceptual phase.
Sector adoption velocityclaude-sonnet-52/5Academic and clinical psychology research is a slower-adopting sector for full AI-driven research pipelines, though AI writing/analysis tools are increasingly used as aids.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can meaningfully assist by automating literature search, organizing study data, suggesting statistical approaches, and drafting sections. However, the psychologist must retain control over hypothesis formation, technique design, and validity assessment, making this a genuine augmentation of productivity rather than replacement.
Augmentation potentialclaude-sonnet-54/5AI substantially helps with literature reviews, data analysis, hypothesis generation, and drafting manuscripts, meaningfully boosting researcher productivity while humans retain control over design and interpretation.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist with literature review, data analysis, and draft synthesis of findings, the core activities—designing novel therapeutic techniques, interpreting nuanced clinical outcomes, and developing conceptual frameworks—require human clinical judgment, creativity, and iterative refinement that AI cannot yet perform end-to-end with 50% time savings at equal quality.
Task automatabilityclaude-sonnet-51/5Original psychological research requires designing studies, recruiting human subjects, ethical oversight, and interpreting results in a clinical context that current AI cannot autonomously conduct end-to-end.'
Adoption barriersclaude-haiku-4-5-202510014/5Clinical research involving new diagnostic or therapeutic techniques faces regulatory (IRB, FDA pathway for devices/interventions), credentialing (requires licensed psychologist authorship), and liability barriers. Institutional and professional standards mandate human clinical expertise in validating and publishing novel techniques.
Adoption barriersclaude-sonnet-54/5Clinical research involving human subjects requires licensed professionals, IRB approval, and ethical review, creating strong institutional and regulatory barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI can reduce some research overhead (literature review, basic analysis), but a qualified clinical psychologist must direct the research, validate findings, and conceptualize improvements. The all-in cost of AI infrastructure and human oversight remains comparable to or higher than direct human research effort for this knowledge-intensive task.
Cost vs. human wageclaude-sonnet-52/5AI can cheaply assist with literature synthesis or statistical analysis, but the overall research process still requires expensive human expertise, IRB compliance, and subject recruitment, keeping costs comparable to human-led research.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed product reliably conducts independent research to develop or improve diagnostic/therapeutic techniques. AI tools exist for literature mining and statistical analysis, but the integrative, novel-technique-development phase lacks production-grade automation; most systems remain research-stage.
Technical feasibility todayclaude-sonnet-51/5No deployed product independently designs and executes clinical/counseling research programs; AI is at best a research-stage tool for literature review or data analysis assistance.

Provide psychological or administrative services and advice to private firms or community agencies regarding mental health programs or individual cases.

17

CI 9–25 · exposure 17 · augmentation 63 · importance 2.9/5 · 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 cautious about AI substitution for licensed clinical roles. Adoption of AI in clinical psychology is largely pilot-stage and focused on administrative support, not autonomous service delivery; organizational and legal resistance is substantial.
Sector adoption velocityclaude-sonnet-52/5Healthcare and mental health services sectors have historically been slower to adopt AI for judgment-intensive advisory work compared to information/finance sectors, though administrative-support AI adoption is growing.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can usefully assist with administrative tasks (scheduling, documentation, initial intake screening), literature synthesis, and draft report generation, which can modestly raise a psychologist's productivity. However, augmentation is limited to peripheral tasks; the core clinical and advisory work remains fundamentally human.
Augmentation potentialclaude-sonnet-54/5AI can meaningfully assist by drafting program proposals, summarizing case histories, generating reports, and suggesting evidence-based interventions, significantly boosting psychologist productivity while they retain final judgment.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires nuanced clinical judgment, therapeutic relationships, confidentiality handling, and situational understanding of organizational context that current AI cannot reliably perform end-to-end. AI cannot conduct genuine psychological assessment or therapy, nor can it navigate the relational and ethical complexities of advising firms and agencies on sensitive mental health matters.
Task automatabilityclaude-sonnet-52/5This blends clinical judgment, organizational consultation, and case-specific advice that requires contextual understanding, professional accountability, and relationship-building that current AI cannot fully replicate end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Legal and regulatory barriers are stringent: psychological services and clinical advice must be provided by licensed psychologists in most jurisdictions, and organizational liability for mental health guidance falls on the credentialed professional. Liability asymmetry and licensing requirements create hard barriers to full automation.
Adoption barriersclaude-sonnet-54/5Advising on mental health programs and individual cases typically requires licensure, clinical judgment, and liability accountability, creating strong professional and legal barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI tools for administrative support are inexpensive, but they cannot replace the loaded cost of a clinical psychologist ($80k–$150k+ annually) because the task requires licensed clinical judgment. Any cost savings would be marginal and limited to back-office support, not the primary service delivery.
Cost vs. human wageclaude-sonnet-52/5While AI could cheaply draft materials, the consulting value depends on licensed expertise and liability coverage that still requires expensive human oversight, keeping overall cost comparable to or only modestly below human-only delivery.
Technical feasibility todayclaude-haiku-4-5-202510012/5While AI can assist with administrative support, literature review, and drafting guidance documents, no deployed product reliably performs the core clinical and advisory functions independently. Existing AI tools lack the clinical licensure, accountability, and ability to engage in genuine therapeutic or complex organizational consultation that this role demands.
Technical feasibility todayclaude-sonnet-52/5AI tools exist for drafting policy documents or summarizing case data, but no deployed product independently provides authoritative mental health program consulting or individual case advice to organizations.

Consult with other professionals, agencies, or universities to discuss therapies, treatments, counseling resources or techniques, and to share occupational information.

16

CI 7–25 · exposure 13 · augmentation 50 · importance 3.6/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Clinical psychology remains largely relationship and expertise-driven; adoption of AI for core consultation work is minimal. Most adoption to date focuses on administrative support (scheduling, note-taking) rather than replacing the consultation itself.
Sector adoption velocityclaude-sonnet-52/5Healthcare and mental health services adopt AI more slowly than other professional sectors, especially for interpersonal, judgment-heavy collaborative tasks like this.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can usefully assist by retrieving evidence-based treatment guidelines, organizing prior consultation notes, and drafting agendas, enabling psychologists to prepare more thoroughly. However, the augmentation is limited to preparation and follow-up, not the consultation itself.
Augmentation potentialclaude-sonnet-53/5AI can help summarize case notes, prepare talking points, or search literature to support these consultations, offering moderate assistance without replacing the interaction itself.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can draft summaries, facilitate scheduling, and prepare background materials for consultations, the task fundamentally requires expert judgment, nuance in discussing therapeutic approaches, and real-time dialogue with peers. Current AI cannot reliably conduct the full bidirectional expert consultation that defines this task.
Task automatabilityclaude-sonnet-51/5This task requires synchronous professional dialogue, relationship-building, and case-specific judgment across institutions that current AI cannot conduct autonomously end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Professional norms, licensing, liability, and institutional trust strongly favor human-to-human consultation for therapeutic guidance. Organizations and regulatory bodies expect licensed psychologists to engage directly with peers; automated substitution faces high professional and organizational friction.
Adoption barriersclaude-sonnet-54/5Professional consultation often involves confidential clinical information, licensure-bound judgment, and interagency trust relationships that create strong practical and ethical barriers to automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5The marginal cost of using AI to support a consultation is modest, but AI cannot eliminate the human expert's time; it only reduces preparation and documentation overhead. The cost of AI plus human remains higher than the human alone for this collegial task.
Cost vs. human wageclaude-sonnet-51/5There is no AI product performing this consultative role, so cost comparison favors the human entirely; any AI use is just a minor support tool, not a replacement.
Technical feasibility todayclaude-haiku-4-5-202510012/5AI can assist with information retrieval and meeting coordination, but no deployed product reliably replaces human expert-to-expert consultation. Chatbots lack the clinical credibility and contextual sophistication to serve as genuine consultants in therapeutic discussions.
Technical feasibility todayclaude-sonnet-51/5No deployed product substitutes for a psychologist consulting with other professionals or agencies; this remains a human interpersonal and professional networking activity.

Interact with clients to assist them in gaining insight, defining goals, and planning action to achieve effective personal, social, educational, or vocational development and adjustment.

15

CI 6–24 · exposure 13 · augmentation 63 · importance 4.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Adoption of AI in clinical psychology is slow and cautious. While some clinicians use AI for administrative tasks (note-taking, scheduling), deploying AI to conduct client sessions or replace the therapeutic relationship remains rare and controversial in professional practice.
Sector adoption velocityclaude-sonnet-52/5Healthcare and mental health services are historically slower adopters of AI for core clinical interactions due to regulatory, ethical, and trust barriers, though adjunct digital tools are growing.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist clinicians by drafting session notes, organizing client data, or suggesting psychoeducational resources, raising some administrative efficiency. However, the core interaction—listening, responding, building therapeutic alliance—remains fundamentally human, limiting AI's role to meaningful but peripheral support.
Augmentation potentialclaude-sonnet-54/5AI can meaningfully assist psychologists with session note-taking, treatment planning drafts, psychoeducation materials, and between-session client support tools, enhancing productivity while the clinician remains central.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires sustained empathetic engagement, real-time emotional attunement, and responsive clinical judgment that current AI cannot reliably perform. The goal of helping clients gain insight through dialogue depends on human presence, trust-building, and adaptive understanding of individual psychological states—elements that remain far beyond what deployed systems can deliver.
Task automatabilityclaude-sonnet-52/5Core therapeutic interaction requires nuanced empathy, real-time judgment, and relationship-building that current AI cannot fully replicate end-to-end; some chatbot-based support exists but doesn't meet the equal-quality bar for licensed clinical work.
Adoption barriersclaude-haiku-4-5-202510015/5Significant legal and regulatory barriers protect this task: psychologists must be licensed, therapy involves a duty of care and liability for harm, and mental health regulations often explicitly require a qualified human practitioner. Professional ethics codes and state licensing laws make it difficult or illegal to fully substitute AI for this work.
Adoption barriersclaude-sonnet-55/5Clinical and counseling psychology is a licensed profession with strict legal, ethical, and liability requirements mandating a qualified human professional for diagnosis and treatment planning.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI inference costs remain negligible, but integrating an AI system into a therapeutic practice requires clinical oversight, regulatory compliance, liability infrastructure, and human fallback—overhead that erodes any unit cost advantage. A human clinician remains substantially cheaper when all-in costs are counted.
Cost vs. human wageclaude-sonnet-53/5AI chat tools are far cheaper per interaction than a psychologist's session, but they only handle a narrow subset of the task, so cost comparison for equivalent full-service output is roughly comparable once oversight and liability are factored in.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product reliably performs therapeutic interaction and insight-generation with clients at production quality. While chatbots can simulate conversation, they lack the clinical competence, liability insurance, and proven safety record required in mental health practice, and they cannot be deployed as replacements for this core clinical function.
Technical feasibility todayclaude-sonnet-52/5Mental health chatbots (e.g., Woebot, Wysa) exist and are used for adjunctive support, but no deployed product reliably performs full clinical/counseling interaction independently at scale in licensed practice.

Use a variety of treatment methods, such as psychotherapy, hypnosis, behavior modification, stress reduction therapy, psychodrama, or play therapy.

13

CI 6–20 · exposure 13 · augmentation 50 · importance 4.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Adoption of AI in clinical psychology is slow and cautious. While some clinicians use AI for administrative tasks (documentation, treatment planning support) and psychoeducation, actual displacement or automation of the therapeutic session itself is minimal in production. Regulatory and ethical concerns, plus demand for human practitioners, keep velocity low.
Sector adoption velocityclaude-sonnet-52/5Healthcare and mental health services adopt AI slowly due to regulatory, ethical, and trust concerns, with pilots more common than full production deployment for actual treatment delivery.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can modestly assist psychologists by suggesting treatment techniques, automating progress notes, recommending evidence-based interventions, or triaging intake questionnaires. However, the core therapeutic act—the live, empathic, adaptive interaction—remains with the human, so augmentation is helpful but limited in scope.
Augmentation potentialclaude-sonnet-53/5AI can assist with treatment planning, session notes, psychoeducation content, and between-session support tools, meaningfully aiding clinicians without replacing the therapeutic relationship.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can generate therapy scripts, psychoeducational content, and suggest evidence-based techniques, the interactive, real-time judgment and dynamic adjustment required in live therapeutic relationships—detecting subtle emotional cues, adapting in the moment to client state, building therapeutic alliance—remains largely beyond current AI. No AI system can reliably execute the full suite of these methods end-to-end with 50% time savings at equal therapeutic quality.
Task automatabilityclaude-sonnet-51/5Delivering actual therapeutic interventions requires real-time relational judgment, ethical responsibility, and adaptive human presence that current AI cannot perform end-to-end at equal quality.
Adoption barriersclaude-haiku-4-5-202510015/5Hard legal and regulatory barriers: clinical diagnosis, psychotherapy, and hypnosis for therapeutic purposes require licensed mental health professionals in virtually all jurisdictions. Liability, duty of care, and malpractice risk are extreme if an unlicensed entity (human or AI) conducts psychotherapy. State licensure boards explicitly restrict these tasks to licensed psychologists or psychiatrists.
Adoption barriersclaude-sonnet-55/5Psychotherapy and related clinical treatments require licensure, legal accountability, and are tightly regulated; only credentialed professionals may legally provide these services.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI-based self-help or adjunctive tools (chatbots, guided meditations) are cheap, but they cannot replace the full clinical scope; when AI is used alongside human therapists for documentation or session notes, the all-in cost (including oversight, regulatory compliance, and liability) approaches or exceeds the human wage for the clinical work itself.
Cost vs. human wageclaude-sonnet-52/5AI chat tools are cheap per interaction, but they cannot substitute for the full scope of licensed treatment, so cost comparison for the actual task is not favorable once liability and efficacy are considered.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI product reliably delivers clinical psychotherapy, hypnosis, or play therapy as a substitute for a licensed psychologist. AI chatbots exist for psychoeducation and supportive dialogue, but they do not meet clinical standards for diagnosing, treating, or managing mental health conditions in production healthcare settings.
Technical feasibility todayclaude-sonnet-52/5Some chatbot-based mental health apps exist (e.g., Woebot, Wysa) but they are narrow, supplementary tools, not substitutes for licensed clinicians delivering psychotherapy or hypnosis.

Develop and implement individual treatment plans, specifying type, frequency, intensity, and duration of therapy.

13

CI 5–20 · exposure 17 · augmentation 63 · importance 4.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Mental health care remains heavily regulated and human-centered; even highly digitized healthcare systems retain treatment planning as a licensed clinician responsibility. Adoption of autonomous AI treatment planning is minimal given legal restrictions and clinical risk aversion.
Sector adoption velocityclaude-sonnet-52/5Healthcare and mental health services are historically slow adopters of AI for core clinical decision-making due to regulatory, liability, and trust concerns, though administrative AI tools are spreading.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can meaningfully assist with literature retrieval, outcome measure suggestions, template scaffolding, and evidence-based recommendations for therapy modality and frequency, supporting clinician planning workflows. However, the clinician retains and must execute final judgment on the individualized plan.
Augmentation potentialclaude-sonnet-54/5AI can meaningfully assist by synthesizing patient history, suggesting evidence-based treatment options, and drafting plan documentation, improving clinician efficiency while the psychologist retains final judgment.
Task automatabilityclaude-haiku-4-5-202510011/5Treatment planning requires nuanced clinical judgment, differential diagnosis integration, patient-specific contraindications, and tailored therapeutic matching that current AI cannot perform end-to-end. AI cannot reliably assess complex psychopathology, establish therapeutic alliance requirements, or make the individualized clinical decisions that define adequate treatment planning.
Task automatabilityclaude-sonnet-52/5AI can draft treatment plan templates and suggest evidence-based interventions given intake data, but clinical judgment on individualized therapy type, frequency, and duration for a specific patient requires nuanced human assessment that current systems cannot reliably replicate end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Treatment planning is legally and professionally restricted to licensed mental health providers; regulatory standards (licensing boards, accreditation bodies, malpractice liability) explicitly require licensed clinicians to develop and sign off on treatment plans. Liability asymmetry is severe: AI errors in treatment planning can harm patients, creating unmovable gatekeeping.
Adoption barriersclaude-sonnet-55/5Treatment plans typically must be authored or approved by a licensed mental health professional under legal and ethical practice standards, with significant liability for errors, making this a hard-barrier task.
Cost vs. human wageclaude-haiku-4-5-202510011/5Comprehensive oversight, validation, and liability management would exceed the cost of human clinician planning given the high stakes and current AI error rates in clinical decision-making. The loaded cost of ensuring safe AI-generated plans exceeds direct human clinician time.
Cost vs. human wageclaude-sonnet-52/5AI drafting assistance is cheap, but the requirement for licensed clinician review, liability, and oversight means overall cost savings versus a human psychologist doing this task remain modest.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed products perform this task reliably in production; AI can assist with template generation or literature review but cannot independently develop clinically sound, individualized treatment plans. Existing systems lack the clinical validation and liability coverage required for autonomous treatment planning.
Technical feasibility todayclaude-sonnet-52/5Some clinical decision-support tools and AI scribes assist with documentation, but no deployed product autonomously creates and finalizes individualized treatment plans in production clinical settings.

Consult with or provide consultation to other doctors, therapists, or clinicians regarding patient care.

13

CI 5–20 · exposure 17 · augmentation 63 · importance 4.0/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Clinical psychology and psychiatry remain highly regulated, hierarchical professions with strong norms requiring human-to-human professional consultation; adoption of AI for peer consultation in production settings is negligible.
Sector adoption velocityclaude-sonnet-52/5Healthcare and behavioral health remain slow to adopt AI for clinical judgment tasks due to regulation, liability, and trust concerns, with pilots more common than production use in this specific interaction.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist by summarizing case notes, flagging relevant literature, or organizing differential diagnoses, potentially speeding preparation for consultation, but the consultation itself remains a human professional responsibility.
Augmentation potentialclaude-sonnet-54/5AI can meaningfully assist by summarizing patient records, surfacing relevant research, drafting consultation notes, or suggesting differential diagnoses, enhancing the consulting clinician's efficiency while they retain decision authority.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires nuanced clinical judgment, synthesis of complex medical/psychological histories, and professional accountability that current AI systems cannot reliably perform end-to-end. Consultation inherently demands a licensed professional's direct engagement and responsibility.
Task automatabilityclaude-sonnet-52/5This requires nuanced clinical judgment, integration of patient history, and professional accountability that current AI cannot autonomously replicate for peer consultation; only limited sub-parts (literature lookup, summarization) are automatable.
Adoption barriersclaude-haiku-4-5-202510015/5Consultation on patient care is legally restricted to licensed mental health professionals; liability, malpractice exposure, and regulatory requirements (state licensing boards, HIPAA) prevent substitution with unlicensed or unaccountable AI systems.
Adoption barriersclaude-sonnet-55/5Clinical consultation involves licensed professional judgment, liability for patient outcomes, and confidentiality/regulatory requirements (e.g., HIPAA), all of which legally require human clinician involvement and sign-off.
Cost vs. human wageclaude-haiku-4-5-202510011/5The all-in cost of AI systems with necessary human oversight and validation exceeds the cost of direct clinician-to-clinician consultation, which is already embedded in practice workflow.
Cost vs. human wageclaude-sonnet-52/5Given liability, need for expert oversight, and low reliability for this specific interpersonal/clinical judgment task, AI does not yet offer substantial all-in cost savings over a qualified psychologist's time.
Technical feasibility todayclaude-haiku-4-5-202510012/5While AI can draft summaries or suggest differential diagnoses from case materials, no deployed product reliably performs genuine peer consultation—the task requires real-time professional judgment, negotiation, and co-responsibility that existing systems cannot assume.
Technical feasibility todayclaude-sonnet-52/5AI tools exist to summarize case notes or suggest differential considerations, but no deployed product functions as a reliable clinical consultant that peers actually rely on for patient-care decisions.

Observe individuals at play, in group interactions, or in other contexts to detect indications of cognitive, intellectual, or developmental disabilities.

13

CI 0–25 · exposure 13 · augmentation 38 · importance 3.0/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Clinical psychology remains a human-expert-intensive, regulated domain with slow digitization of core diagnostic work. Adoption of AI for disability screening or observation is in early pilot stages, with no evidence of significant production displacement in mainstream practice.
Sector adoption velocityclaude-sonnet-52/5Clinical psychology involving direct behavioral observation is a low-digitization, high-touch field where AI adoption for this specific observational task is minimal and largely confined to pilot research.
Augmentation potentialclaude-haiku-4-5-202510013/5AI-assisted video annotation, flagging of atypical behavioral markers, or real-time prompting during observation could meaningfully assist a clinician's structured assessment. However, the clinical judgment and pattern recognition required limit how transformative such assistance becomes relative to standard practice.
Augmentation potentialclaude-sonnet-52/5Some tools (e.g., video-based behavioral coding software, eye-tracking) can assist by flagging patterns for later review, but they play a minor supportive role rather than transforming the observational process itself.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can analyze behavioral video and detect some statistical anomalies, observing play contexts requires nuanced interpretation of intent, affect regulation, social reciprocity, and developmental appropriateness—judgments that depend heavily on contextual knowledge and clinical training. Current systems cannot reliably replace this observational work end-to-end with 50% time savings at equal quality.
Task automatabilityclaude-sonnet-51/5This requires physically observing real-time behavior in naturalistic settings and synthesizing subtle nonverbal cues into clinical judgment—current AI cannot perform this end-to-end task at equal quality with major time savings.
Adoption barriersclaude-haiku-4-5-202510014/5This task sits at the core of clinical diagnosis and treatment planning, requiring licensed psychologists in most jurisdictions. Liability, standards of care, and legal/regulatory requirements that a qualified human clinician perform or directly oversee disability assessment create strong adoption barriers.
Adoption barriersclaude-sonnet-55/5Diagnosing developmental or cognitive disabilities requires a licensed psychologist's clinical judgment and carries high liability, with legal and ethical requirements for human evaluation and sign-off.
Cost vs. human wageclaude-haiku-4-5-202510012/5Current video analysis and annotation systems still require significant human oversight, clinical validation, and integration labor. The all-in cost (inference, labeling, clinical review) remains comparable to or higher than direct clinician observation for reliable outputs.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing this observational task, so cost comparison favors the human clinician entirely; any AI video analysis tool would add cost without replacing the clinician's judgment.
Technical feasibility todayclaude-haiku-4-5-202510012/5Products exist for video analysis and behavior classification, but they operate at low specificity for clinical disability detection in real-world play settings and lack the longitudinal, contextual judgment clinicians apply. No mature production system reliably performs this task as a deployed clinical tool.
Technical feasibility todayclaude-sonnet-51/5No deployed clinical product autonomously observes individuals in play or group settings to detect developmental disabilities; this remains research-stage (e.g., experimental video analysis tools) rather than production practice.

Counsel individuals, groups, or families to help them understand problems, deal with crisis situations, define goals, and develop realistic action plans.

7

CI 4–11 · exposure 5 · augmentation 38 · importance 4.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Mental health practice remains stubbornly human-centric and risk-averse; adoption of AI in active counseling is minimal outside research pilots. Healthcare organizations prioritize liability containment and therapist-client relationship integrity, slowing any shift toward algorithmic delivery.
Sector adoption velocityclaude-sonnet-52/5Healthcare/mental health services adopt AI slowly due to regulation, liability, and the sensitive human-contact nature of the work; adjunct tools are piloted but core counseling remains human-led.
Augmentation potentialclaude-haiku-4-5-202510012/5AI can assist with documentation, session note generation, and psychoeducational materials, but offers limited augmentation to the core counseling act itself. Current systems lack the real-time emotional reasoning and relational depth needed to materially enhance a clinician's in-session effectiveness.
Augmentation potentialclaude-sonnet-53/5AI can help with session note-taking, symptom tracking, psychoeducation materials, and between-session support, but for the core counseling interaction itself the productivity gain is moderate given the domain's inherently relational focus.
Task automatabilityclaude-haiku-4-5-202510011/5Counseling requires genuine empathy, nuanced emotional understanding, and real-time adaptive response to human vulnerability that current AI cannot reliably deliver. While AI can generate therapeutic language, it cannot replicate the therapeutic alliance, contextual judgment, and moment-to-moment attunement essential to crisis intervention and meaningful behavior change.
Task automatabilityclaude-sonnet-51/5Effective counseling requires sustained therapeutic alliance, clinical judgment, crisis risk assessment, and adaptive human presence that current AI cannot replicate at equal quality, especially for crisis intervention where errors carry severe consequences.
Adoption barriersclaude-haiku-4-5-202510015/5Counseling is a licensed professional service; most jurisdictions legally require a licensed mental health provider to deliver or directly supervise therapeutic interventions. Liability exposure for algorithmic harm in crisis contexts, combined with regulatory oversight of mental health delivery, creates hard regulatory barriers to full automation.
Adoption barriersclaude-sonnet-55/5Licensure laws, malpractice liability, confidentiality/HIPAA rules, and mandatory reporting/crisis duty-of-care requirements make this a legally protected human function in most jurisdictions.
Cost vs. human wageclaude-haiku-4-5-202510012/5Even if scaled widely, AI counseling oversight, liability management, and integration with clinical workflows would remain costly. Human clinical psychologists command substantial labor costs, but the malpractice and duty-of-care infrastructure around counseling keeps total AI cost-to-benefit analysis unfavorable today.
Cost vs. human wageclaude-sonnet-52/5AI chat tools are cheap per interaction, but the oversight, liability, and quality assurance needed to make them clinically safe substitutes raise effective cost, and they still don't match licensed therapist output.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product reliably performs full counseling sessions in production healthcare settings. Chatbots exist but are explicitly limited to psychoeducation and triage, not the integrated diagnostic and therapeutic work this task describes; clinicians do not deploy them as substitutes for counseling.
Technical feasibility todayclaude-sonnet-52/5Some chatbot apps (e.g., Woebot) offer scripted CBT-style support, but no product reliably performs full psychotherapy counseling including crisis management in production at clinical standards.

Prepare written evaluations of individuals' psychological competence for court hearings.

7

CI 0–15 · exposure 8 · augmentation 38 · importance 3.0/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Legal and healthcare sectors remain highly regulated with strong human-credential requirements; there is no meaningful adoption of AI for autonomous court evaluations, and courts actively resist algorithmic substitution for licensed professional judgment.
Sector adoption velocityclaude-sonnet-52/5Forensic and clinical psychology adopts AI slowly due to licensing, liability, and courtroom evidentiary standards, with most current AI use limited to administrative note-taking rather than evaluation substance.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could marginally assist with note compilation, literature review, or report formatting, but the core clinical assessment, diagnostic reasoning, and legal attestation must remain under human psychologist control, limiting augmentation potential.
Augmentation potentialclaude-sonnet-53/5AI can help psychologists draft report language, organize test results, and summarize case history, moderately speeding up the writing process while the clinician retains full responsibility for conclusions.
Task automatabilityclaude-haiku-4-5-202510011/5Preparing psychological competence evaluations for court requires clinical judgment, synthesis of complex patient history, direct observation, and legally defensible clinical reasoning that current AI cannot perform end-to-end. The task fundamentally requires a licensed psychologist's professional judgment and legal accountability, which AI cannot substitute.
Task automatabilityclaude-sonnet-52/5Drafting portions of a forensic psychological evaluation report can be assisted by AI, but the core task requires clinical interviews, test administration, professional judgment, and legal accountability that AI cannot perform end-to-end today.
Adoption barriersclaude-haiku-4-5-202510015/5Hard legal and regulatory barriers exist: only licensed psychologists can produce evaluations admissible in court, the task requires direct patient contact and clinical judgment, and there is substantial liability for incorrect assessments affecting legal proceedings. Judicial and licensing authorities would not accept AI-generated evaluations.
Adoption barriersclaude-sonnet-55/5Court-admissible competency evaluations legally require a licensed psychologist's professional opinion and signature, with strong liability, ethical, and judicial admissibility requirements preventing AI substitution.
Cost vs. human wageclaude-haiku-4-5-202510011/5AI currently cannot perform this task at all, making cost comparison moot; a human psychologist must perform the evaluation regardless, incurring full professional fees without AI substitution.
Cost vs. human wageclaude-sonnet-52/5AI could cheaply help format or summarize notes, but the bulk of cost is the licensed psychologist's assessment time, testimony preparation, and liability exposure, which AI cannot substitute for at lower cost.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI product can independently produce court-admissible psychological competence evaluations; this requires licensed clinical assessment and carries significant legal liability. Existing AI lacks the authority, accountability, and clinical credibility required in judicial contexts.
Technical feasibility todayclaude-sonnet-51/5No deployed product independently produces court-admissible psychological competency evaluations; this remains a licensed clinician's exclusive domain with only drafting/documentation support tools available.

Conduct assessments of patients' risk for harm to self or others.

6

CI 0–11 · exposure 5 · augmentation 50 · importance 5.0/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Adoption of autonomous AI for risk assessment is minimal in clinical practice. Structured screening tools and decision-support software exist, but they are used as inputs to human judgment, not replacements. Liability, licensing law, and professional standards create strong headwinds against displacement.
Sector adoption velocityclaude-sonnet-52/5Mental healthcare is a moderately digitized but cautious sector; AI adoption for high-stakes clinical risk judgments remains in early pilot/research phases rather than broad production use.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist clinicians by organizing data, flagging structured risk factors, suggesting assessment frameworks, and prompting completeness; however, the core interpretive and judgment task—weighing dynamic context, reading interpersonal cues, and making the ultimate risk determination—remains human-centered. Useful support, but not transformative.
Augmentation potentialclaude-sonnet-53/5AI can assist by flagging risk indicators in patient language, aggregating history, or supporting structured screening tools, but the clinician must interpret and make the final judgment.
Task automatabilityclaude-haiku-4-5-202510011/5Risk assessment for self-harm or harm to others requires nuanced clinical judgment, integration of complex historical context, observation of subtle behavioral and affective cues, and dynamic reassessment. Current AI cannot reliably replace this end-to-end with 50% time savings at equal quality; it lacks the real-time interpersonal assessment and legal accountability required.
Task automatabilityclaude-sonnet-51/5Risk assessment for self-harm or violence requires clinical judgment, direct interpersonal observation, and legal/ethical accountability that current AI cannot replicate end-to-end; no off-the-shelf system can perform this autonomously at equal quality with major time savings.
Adoption barriersclaude-haiku-4-5-202510015/5Strong legal and regulatory barriers protect this task: clinicians are liable for risk assessment decisions, many jurisdictions require a licensed mental health professional to conduct and sign off on risk assessments, and malpractice liability is severe. Regulatory bodies and standard of care explicitly require human clinical judgment.
Adoption barriersclaude-sonnet-55/5Risk-of-harm assessments carry high liability, require licensed professional judgment, and are often legally mandated (e.g., duty-to-warn, involuntary commitment criteria), creating hard regulatory and licensure barriers to automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5While AI inference costs are low, the integration, validation, clinical oversight, and medicolegal liability review required to deploy AI in high-stakes risk assessment substantially elevate all-in costs. The human clinician's loaded cost remains lower given the minimal displacement achievable without unacceptable error.
Cost vs. human wageclaude-sonnet-51/5Given the liability, need for licensed sign-off, and inadequate AI reliability for high-stakes risk determinations, any AI cost savings are negated by required human oversight, making AI not meaningfully cheaper for equivalent trustworthy output.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed clinical product reliably performs independent risk assessment for self-harm or violence at the standard required for clinical decision-making. AI tools exist for risk screening and structured decision support (e.g., actuarial input), but they operate as narrow adjuncts, not autonomous performers, and have not demonstrated comparable diagnostic or predictive validity to clinical judgment in production settings.
Technical feasibility todayclaude-sonnet-51/5No deployed clinical product autonomously conducts suicide/violence risk assessments as a substitute for a licensed clinician; existing AI tools (chatbot screeners, sentiment analysis) are adjuncts, not production-grade replacements.

Direct, coordinate, and evaluate activities of staff and interns engaged in patient assessment and treatment.

1

CI 0–3 · exposure 0 · augmentation 38 · importance 3.9/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Healthcare and clinical settings are regulated sectors with slow AI adoption for core clinical functions; supervision and evaluation of licensed practitioners remain firmly human-controlled.
Sector adoption velocityclaude-sonnet-52/5Healthcare and mental health services are historically slower AI adopters, especially for supervisory and administrative-clinical hybrid tasks involving human accountability.
Augmentation potentialclaude-haiku-4-5-202510012/5AI might assist with scheduling coordination or documentation aggregation, but the evaluative, supervisory, and decision-making core of this task offers limited productivity gain from AI assistance.
Augmentation potentialclaude-sonnet-53/5AI tools can help track staff performance metrics, treatment outcomes data, and scheduling, offering moderate assistance to the supervising psychologist without replacing judgment.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires real-time judgment about clinical competence, ethical oversight, and treatment decisions—fundamentally human supervisory and evaluative roles. Current AI cannot meaningfully perform clinical supervision, staff performance assessment, or intern evaluation at the quality required in healthcare.
Task automatabilityclaude-sonnet-51/5Directing, coordinating, and evaluating staff/intern performance requires managerial judgment, mentorship, and clinical supervision that AI cannot perform end-to-end today.-
Adoption barriersclaude-haiku-4-5-202510015/5State licensing boards and clinical governance standards explicitly require a licensed psychologist to supervise clinical staff and interns; this is a hard regulatory and liability requirement that cannot be delegated to AI.
Adoption barriersclaude-sonnet-55/5Clinical supervision of interns and staff in patient care typically requires licensure and legal accountability for oversight, creating hard regulatory and liability barriers to automation.
Cost vs. human wageclaude-haiku-4-5-202510011/5The oversight, human review, and legal liability required to validate AI coordination of clinical staff far exceed the cost of a human supervisor, making AI substantially more expensive in practice.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing this supervisory function, so no meaningful cost comparison exists—human labor is the only option.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI product reliably performs clinical supervision, staff coordination, or treatment evaluation. While AI can assist with scheduling or documentation, the core supervisory and evaluative functions remain entirely manual in production.
Technical feasibility todayclaude-sonnet-51/5No deployed product performs clinical supervisory or personnel-management functions in psychology practice settings; this remains firmly human-only work.

Supervise and train interns, clinicians in training, and other counselors.

1

CI 0–3 · exposure 0 · augmentation 38 · importance 3.8/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Clinical training and supervision remain heavily regulated, relationship-driven, and tied to licensed professionals. Adoption of AI for this task is virtually nonexistent; organizations continue to rely on credentialed supervisors.
Sector adoption velocityclaude-sonnet-52/5While healthcare and professional services are adopting AI tools generally, direct AI-led clinical supervision has essentially zero adoption due to licensing and liability constraints.
Augmentation potentialclaude-haiku-4-5-202510012/5AI might assist with documentation, scheduling, or providing reading materials, but current systems offer minimal augmentation of the core supervisory functions of assessment, feedback, and clinical judgment development.
Augmentation potentialclaude-sonnet-53/5AI can help supervisors by reviewing session notes, flagging skill gaps, generating training materials, or summarizing case discussions, meaningfully aiding but not replacing the supervisory relationship.
Task automatabilityclaude-haiku-4-5-202510011/5Supervision and training of clinicians requires real-time judgment, relationship-building, adaptive feedback based on trainee development, and ethical accountability. Current AI cannot replicate the dynamic, personalized mentorship and clinical oversight this task demands.
Task automatabilityclaude-sonnet-51/5Supervision and training involves relational trust, live clinical judgment calls, ethical modeling, and mentorship that current AI cannot perform end-to-end, regardless of setup time.
Adoption barriersclaude-haiku-4-5-202510015/5Licensing laws and ethical standards require that clinical supervision be conducted by a qualified, licensed mental health professional who assumes legal and ethical responsibility for supervisee outcomes. No unlicensed entity—human or AI—can legally perform this function.
Adoption barriersclaude-sonnet-55/5Clinical supervision is typically legally required to be performed by licensed practitioners for licensure/certification purposes, creating a hard regulatory barrier.
Cost vs. human wageclaude-haiku-4-5-202510011/5Supervisor salaries are substantial and the task requires expert-level judgment; any AI tool for support would still require human supervisory review and sign-off, making full substitution cost-prohibitive compared to a licensed clinician.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing this task, so cost comparison favors the human supervisor entirely.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI system reliably performs clinical supervision or training of mental health professionals. This requires understanding of clinical nuance, professional judgment, and accountability that current products do not demonstrably deliver in production.
Technical feasibility todayclaude-sonnet-51/5No deployed product independently supervises or trains clinical interns; this remains outside the scope of any production AI system.

Plan and develop accredited psychological service programs in psychiatric centers or hospitals, in collaboration with psychiatrists and other professional staff.

1

CI 0–3 · exposure 0 · augmentation 38 · importance 3.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Healthcare and psychiatric program development are heavily regulated sectors with slow digitization of strategic planning roles. Adoption of AI for autonomous program planning in this context is minimal; human experts remain mandatory decision-makers.
Sector adoption velocityclaude-sonnet-52/5Healthcare administration and clinical program design are slow-adopting areas for AI, with regulatory and organizational inertia limiting deployment of AI in program planning roles.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could modestly assist with literature review, data analysis, or compliance documentation during program planning, but the core collaborative design and accreditation work remains fundamentally human-led with limited augmentation potential.
Augmentation potentialclaude-sonnet-53/5AI can assist with drafting documentation, summarizing accreditation standards, or organizing program materials, providing moderate support even though the core planning and collaboration remain human-driven.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires strategic planning, multi-stakeholder collaboration, professional judgment about clinical service design, and accreditation compliance that involve nuanced human decision-making. Current AI cannot autonomously develop accredited programs or meaningfully replace the collaborative clinical leadership required.
Task automatabilityclaude-sonnet-51/5This is a high-level strategic and organizational task requiring accreditation knowledge, interdisciplinary negotiation, and clinical judgment that AI cannot perform end-to-end today.
Adoption barriersclaude-haiku-4-5-202510015/5Accreditation bodies legally require qualified, licensed human professionals to design and oversee psychological service programs. Liability for program design failures, regulatory mandates, and the fiduciary responsibility of healthcare institutions create hard barriers to autonomous automation.
Adoption barriersclaude-sonnet-55/5Program development in psychiatric/hospital settings requires licensed psychologists and often institutional accreditation bodies, with legal and professional liability requiring qualified human sign-off.
Cost vs. human wageclaude-haiku-4-5-202510011/5The task demands the expertise and judgment of multiple senior licensed professionals (psychologists, psychiatrists, administrators); AI cost per output would not approach the value of a human expert team's work, and oversight costs would be substantial.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing this task, so cost comparison favors the human professional entirely; any AI involvement would only be supplementary at added cost.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI system can independently plan and develop accredited psychological service programs; this requires licensed psychologists to make clinical and organizational decisions. AI tools may assist with documentation or analysis, but no product performs this task end-to-end in production.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product plans or develops accredited hospital psychological service programs; this remains firmly a human administrative and clinical leadership function.

Related occupations — Life, Physical & Social Science

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.