Dermatologists
29-1213.00Median wage $328,730/yr11,370 employed (US)Rank #663 of 923 scored · top 72% by substitution
Diagnose and treat diseases relating to the skin, hair, and nails. May perform both medical and dermatological surgery functions.
Sub-scores
0–100 · band = confidence interval from rater disagreement
Substitution20
Exposure22
Augmentation60
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
18 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%20
panel mean rating 1.8/5 → substitution pressure 20/100
Technical feasibility todayw 20%25
panel mean rating 2.0/5 → substitution pressure 25/100
Cost vs. human wagew 15%21
panel mean rating 1.9/5 → substitution pressure 21/100
Adoption barriersw 20%inverted — strong barriers lower the score14
panel mean rating 4.4/5 (barrier strength) → substitution pressure 14/100
Sector adoption velocityw 10%23
panel mean rating 1.9/5 → substitution pressure 23/100
Task breakdown (18 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.
Read current literature, talk with colleagues, and participate in professional organizations or conferences to keep abreast of developments in dermatology.
59CI 46–72 · exposure 55 · augmentation 88 · importance 4.3/5 · click for rater detail
Read current literature, talk with colleagues, and participate in professional organizations or conferences to keep abreast of developments in dermatology.
59| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Healthcare and professional services sectors show moderate adoption of AI literature tools and CME platforms, with pilots and early integration in large academic centers and hospital systems. Adoption remains uneven; many individual practitioners and smaller practices lag, and conference attendance remains largely human-driven social activity. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Medical professionals are adopting AI literature summarization and search tools at a moderate pace, with pilots and some routine use in staying current, though full integration into CME workflows is still developing. |
| Augmentation potential | claude-haiku-4-5-20251001 | 5/5 | AI dramatically assists dermatologists in staying current: automated journal alerts, rapid summarization of new studies, and personalized knowledge feeds significantly amplify what a human can absorb and integrate. The human remains the arbiter of clinical relevance, making this a high-augmentation scenario. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools substantially help dermatologists filter, summarize, and prioritize literature, saving significant time even though the full task of professional engagement remains human-driven. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | AI can rapidly scan, summarize, and synthesize current dermatology literature, conference abstracts, and professional guidelines with >50% time savings compared to manual review. However, the human judgment required to evaluate significance and relevance to one's practice, plus the social aspect of colleague conversations, means full end-to-end automation without meaningful human involvement is limited. |
| Task automatability | claude-sonnet-5 | 2/5 | AI can summarize literature and surface relevant papers, but the task inherently involves professional networking, conference participation, and discussion with colleagues that require human presence and judgment, limiting full automation. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | No licensing requirement or legal mandate mandates human performance of literature review or conference attendance. Organizational and professional norms favor human judgment in deciding what knowledge matters, but these are weak barriers compared to regulated clinical acts. CME credit requirements typically allow online or AI-assisted content, not blocking automation. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement mandates human-only literature review, but professional norms and value of in-person collegial exchange and CME conference credit create moderate friction against pure AI substitution. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | Literature scanning and summarization via AI costs a fraction of the dermatologist's loaded hourly wage. Integrating these tools into existing workflows is inexpensive, and the time savings are substantial, making the cost ratio strongly favorable for the information-gathering components. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI-assisted literature review tools are cheap relative to a dermatologist's time spent reading, but conference attendance and colleague discussion are not replaceable by cheaper AI alternatives, keeping overall cost comparable. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Deployed products (literature summarization tools, AI-assisted journal filtering, automated CME platforms, and conference alert systems) reliably perform parts of this task at scale in healthcare settings. Literature scanning and summarization are mature; however, meaningful professional networking and peer discussion remain poorly automated. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Products like AI literature summarizers, research digest tools, and PubMed AI search assistants are deployed and used by clinicians, but they only cover the reading portion, not the networking/conference aspects of staying current. |
Record patients' health histories.
46CI 41–50 · exposure 50 · augmentation 75 · importance 4.6/5 · click for rater detail
Record patients' health histories.
46| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Dermatology practices are moderate adopters of AI-assisted documentation; larger health systems pilot EHR AI features, but independent practices and smaller groups adopt more slowly. Production deployment is growing but still not deeply penetrated across the specialty. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Healthcare documentation AI is being piloted and adopted in outpatient clinics at a moderate pace, following broader but still cautious digitization trends in medicine. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI-assisted documentation tools significantly accelerate history capture by offering templates, auto-population from previous records, and real-time suggestions, allowing dermatologists to focus on clinical reasoning rather than manual data entry. This augmentation is widely valued and deployed. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI scribes and dictation tools meaningfully speed up history-taking documentation while the physician remains responsible for verifying and finalizing the record. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI can extract and summarize medical history from patient interviews or documents with moderate accuracy, but requires significant human review and correction due to medical accuracy requirements and the need to capture contextually important details. Meeting a 50% time-saving threshold is feasible with structured workflows but not yet reliable without oversight. |
| Task automatability | claude-sonnet-5 | 3/5 | AI scribes and NLP tools can capture, transcribe, and structure health history from patient intake forms or conversations, but full accuracy and clinical nuance still require physician review and correction. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Medical record accuracy is legally and clinically critical, with liability tied directly to documentation quality. Regulatory requirements (HIPAA, malpractice standards) and professional obligation to ensure record completeness create strong governance barriers that slow autonomous adoption. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No strict licensing requirement mandates the physician personally record history, but liability for inaccuracies, EHR compliance, and patient privacy rules create moderate friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI documentation tools require significant integration, training, and human oversight time that limits cost savings. The loaded cost of a dermatologist reviewing and correcting automated records remains substantial relative to the AI inference cost, making the all-in economics marginal or unfavorable. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI scribe subscriptions are cheaper than dedicated scribes per hour, but integration, oversight, and correction time reduce the net savings relative to a dermatologist's own documentation workflow. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | EHR systems with AI-assisted documentation exist and are deployed in many practices, but they typically require substantial physician review and editing rather than fully autonomous capture. Error rates on clinical detail remain material enough that most clinicians do not trust unreviewed AI-generated histories. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Ambient documentation tools (e.g., AI scribes) are deployed in some dermatology and general practice settings, but adoption is uneven and error correction is still routinely needed. |
Counsel patients on topics such as the need for annual dermatologic screenings, sun protection, skin cancer awareness, or skin and lymph node self-examinations.
29CI 25–34 · exposure 25 · augmentation 75 · importance 4.8/5 · click for rater detail
Counsel patients on topics such as the need for annual dermatologic screenings, sun protection, skin cancer awareness, or skin and lymph node self-examinations.
29| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Healthcare adoption of fully autonomous AI counseling remains low due to regulatory constraints, liability concerns, and clinician resistance; most deployment is limited to supplemental educational materials rather than substitution for physician counseling. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Healthcare overall remains a slower-adopting sector for patient-facing AI in clinical encounters, with pilots more common than production deployment of AI-driven counseling replacing physician-patient interaction. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can powerfully assist dermatologists by generating customizable educational materials, organizing evidence-based guidance on sun protection and screening, and preparing summaries—allowing physicians to deliver more comprehensive, consistent counseling with higher productivity. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can generate patient education handouts, personalized risk summaries, and reminder systems for screenings and self-exams that dermatologists can use to enhance and reinforce their counseling, meaningfully boosting efficiency while the physician remains central. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can generate informational content about sun protection and skin cancer awareness, counseling requires nuanced patient engagement, addressing individual concerns, and building trust—elements that current AI systems cannot reliably replicate in a meaningful, personalized way that constitutes ≥50% time savings at equal quality. |
| Task automatability | claude-sonnet-5 | 2/5 | General health counseling content can be generated by AI (e.g., chatbots delivering sun-protection education), but personalized counseling tied to a specific patient's exam findings and risk profile requires clinical judgment and rapport that current AI cannot fully replace end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Medical counseling involves clinical judgment and patient trust; regulatory frameworks and malpractice liability create strong barriers to full automation, and patients strongly prefer human-delivered health advice that carries professional accountability. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No strict licensing requirement mandates a dermatologist personally deliver generic educational counseling, but liability, patient trust, and standard-of-care expectations in a clinical encounter create meaningful friction against full delegation to AI. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Even accounting for AI content generation, the oversight required from a licensed dermatologist to verify accuracy, contextualize for individual patients, and maintain liability makes the all-in cost comparable to or higher than direct human counseling. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI-generated educational content is very cheap to produce, but the counseling task in practice is bundled into the visit and requires physician time for individualized risk communication, so realized savings are moderate rather than dramatic. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed products reliably perform comprehensive patient counseling on dermatologic topics; chatbots and educational content exist but lack the clinical judgment, personalization, and accountability required to substitute for physician-led counseling in real medical practice. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Patient-facing chatbots and AI health assistants exist and can deliver generic skin cancer awareness content, but no deployed product reliably substitutes for a dermatologist's individualized counseling session in production clinical workflows. |
Conduct or order diagnostic tests such as chest radiographs (x-rays), microbiologic tests, or endocrinologic tests.
26CI 20–32 · exposure 30 · augmentation 63 · importance 4.3/5 · click for rater detail
Conduct or order diagnostic tests such as chest radiographs (x-rays), microbiologic tests, or endocrinologic tests.
26| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Adoption of AI-assisted diagnostic interpretation (especially image reading) is growing in dermatology and radiology, but autonomous test ordering remains rare in production. Most adoption centers on interpretation assistance rather than replacement of the ordering decision itself. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Healthcare overall lags in AI adoption for diagnostic ordering/decision workflows due to regulatory, liability, and integration hurdles, with pilots more common than production use. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI significantly augments dermatologists by automating or accelerating image interpretation, summarizing findings, and flagging abnormalities, which helps dermatologists make faster, more informed ordering and clinical decisions while they remain in control of test selection and clinical management. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help flag relevant tests, summarize differential diagnoses, or pre-screen imaging, offering moderate assistance while the physician retains final decision authority. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can assist in interpreting some diagnostic images (e.g., radiographs via computer vision), the task includes ordering decisions that require clinical judgment about which tests are appropriate for a given patient presentation. Current AI cannot reliably make these ordering decisions independently without human oversight, and the full end-to-end task—deciding what to order, when, and why—remains primarily human-driven. |
| Task automatability | claude-sonnet-5 | 2/5 | Ordering tests and interpreting them requires clinical judgment integrated with patient history and physical exam; AI can support diagnostic reasoning but cannot end-to-end perform this task at equal quality with 50% time savings today. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Strong regulatory and liability barriers exist: dermatologists are legally responsible for test ordering decisions and their clinical appropriateness. Medical licensure and malpractice liability tie this task tightly to a licensed physician; autonomous AI ordering would face regulatory resistance and potential liability gaps. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Ordering diagnostic tests and acting on results is a licensed medical act requiring physician authorization and legal accountability, making this a hard regulatory barrier. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI-assisted image analysis can reduce the time for interpretation, but the ordering decision, clinical context-setting, and oversight still demand a dermatologist's time. Integration costs and the need for human judgment in test selection mean the all-in cost advantage remains marginal compared to a full physician cost. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI tools reduce some interpretation time but still require physician oversight, integration with EHR/lab systems, and liability coverage, keeping all-in costs close to or only modestly below physician cost for this specific task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Deployed AI products exist for image interpretation (chest X-ray reading, dermoscopy analysis), but ordering the right diagnostic test in the first place requires nuanced clinical reasoning that is not yet reliably automated at scale. Most real-world systems require a dermatologist to decide what to order; AI assists only in interpretation post-ordering. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | AI diagnostic aids exist (e.g., radiograph interpretation tools) but are narrow, decision-support only, and not deployed to autonomously order or fully interpret multi-modal diagnostic workups in dermatology practice. |
Conduct clinical or basic research.
25CI 25–25 · exposure 25 · augmentation 63 · importance 3.4/5 · click for rater detail
Conduct clinical or basic research.
25| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | While dermatology research labs use AI tools for analysis and mining, adoption of AI-driven autonomous research workflows remains in pilot phases. Most clinical research is still conducted under traditional human-led protocols with regulatory constraints limiting rapid automation adoption. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Academic medicine and clinical research adopt AI tools cautiously; pilots for literature review and data analysis exist but full research automation is rare and slow to scale. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI meaningfully assists dermatologists in research through literature summarization, statistical analysis, image processing for skin lesion studies, and hypothesis generation from large datasets. These tools improve productivity on research components, though the dermatologist remains responsible for design and interpretation. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI significantly assists with literature review, data analysis, image classification, and drafting manuscripts, meaningfully boosting researcher productivity while humans retain scientific judgment and oversight. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Clinical research involves complex hypothesis generation, patient recruitment, protocol design, and interpretation of nuanced biological findings that require deep domain expertise and judgment. While AI can assist with literature review, data analysis, and statistical modeling, end-to-end autonomous research execution with ≥50% time savings at equal quality remains beyond current capabilities. |
| Task automatability | claude-sonnet-5 | 2/5 | Clinical/basic research involves hypothesis generation, experimental design, patient recruitment, wet-lab work, and interpretation that require physical execution and creative judgment AI cannot fully replace end-to-end today.: |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Clinical research requires IRB approval, legal liability for patient safety, regulatory compliance (FDA, HIPAA), informed consent protocols, and institutional accountability. Dermatologists must sign off on research decisions and data integrity, creating substantial legal and regulatory barriers to autonomous AI research execution. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Clinical research is heavily regulated (IRB approval, human subjects protections, publication and credentialing norms), requiring licensed/qualified researchers to design and sign off on studies. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Dermatologists conducting research earn substantial salaries (often $200k+/year loaded), and current AI tools require significant computational infrastructure, domain-expert oversight, and integration labor that do not yet approach order-of-magnitude cost advantage for research-quality outputs. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI can cut costs on subtasks like data analysis or literature synthesis, but overall research still requires expensive human expertise, lab infrastructure, and oversight, keeping total cost comparable to human-led efforts. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed product system reliably conducts independent clinical or basic research end-to-end. AI tools exist for literature mining and statistical analysis, but design of novel research protocols, IRB interactions, patient management, and novel hypothesis generation remain human-dependent in production settings. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | AI tools (literature review, statistical analysis, image analysis for dermatology research) exist and are used, but no deployed product independently conducts full research studies reliably. |
Diagnose and treat skin conditions such as acne, dandruff, athlete's foot, moles, psoriasis, or skin cancer.
25CI 20–30 · exposure 30 · augmentation 63 · importance 4.7/5 · click for rater detail
Diagnose and treat skin conditions such as acne, dandruff, athlete's foot, moles, psoriasis, or skin cancer.
25| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Dermatology is a highly specialized, human-contact-dependent medical field with slow AI adoption. Most practices still rely on in-person examination; AI tools remain experimental or adjunctive in most settings. Adoption is concentrated in a few digital-forward telemedicine and research centers, not mainstream clinical practice. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Healthcare overall adopts AI diagnostic aids cautiously with regulatory review and slow clinical integration despite promising research in dermatology imaging. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI-powered dermoscopy and lesion-flagging tools do assist dermatologists by highlighting concerning lesions and offering differential diagnoses, improving review speed and consistency. However, augmentation is limited to image analysis; physical examination, patient interaction, and complex treatment decisions remain human-driven. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-assisted dermoscopy and image classification meaningfully help dermatologists triage and prioritize suspicious lesions, improving diagnostic speed and accuracy while the physician retains final judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI excels at image classification for skin lesions (melanoma detection ~90% accuracy in benchmarks), real-world diagnosis requires patient history, physical examination, palpation, differential reasoning across overlapping conditions, and treatment planning tailored to individual factors. Current AI cannot perform the full diagnostic and treatment workflow end-to-end at the 50%-time-saving threshold. |
| Task automatability | claude-sonnet-5 | 2/5 | AI can assist with image-based triage of skin lesions but full diagnosis-and-treatment requires physical exam, biopsy decisions, procedural treatment, and patient management that current systems cannot execute end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Dermatology diagnosis and treatment decisions carry liability risk; medical licensing requirements legally mandate a physician's judgment and sign-off. Patients also expect human examination and tactile assessment. Regulatory (FDA/equivalent) oversight of diagnostic AI and standard-of-care expectations create substantial adoption friction. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Diagnosing and treating disease is a licensed medical act requiring a physician's legal authorization, with high liability exposure and mandatory human sign-off. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI image analysis for skin lesions is cheap per image (~$0.01–0.10), but integration into clinical workflows, regulatory oversight, and the need for human dermatologist review to validate and treat means total cost per patient remains comparable to or higher than a dermatologist visit, especially for complex cases. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI image analysis is cheap per query, but the overall task includes procedures, prescriptions, and liability-bearing decisions requiring a licensed physician, keeping blended cost comparable to or above human cost. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | FDA-cleared AI products for skin lesion classification exist (e.g., DermAssist, SkinVision) and show good performance on narrow tasks like melanoma screening, but they operate as decision-support tools with significant oversight requirements, not autonomous diagnosis. Production deployment remains largely in dermoscopy-assisted workflows, not end-to-end dermatology. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | FDA-cleared dermoscopy AI tools exist for lesion classification support, but no deployed product independently diagnoses and treats the full range of conditions listed at scale in clinical practice. |
Recommend diagnostic tests based on patients' histories and physical examination findings.
24CI 20–29 · exposure 25 · augmentation 63 · importance 4.6/5 · click for rater detail
Recommend diagnostic tests based on patients' histories and physical examination findings.
24| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Dermatology practices are adopting AI-assisted imaging tools, but uptake of AI for test recommendation specifically remains limited and experimental. Most adoption is in large academic and specialized centers; broad deployment in primary and specialty practice lags. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Healthcare, especially specialty medicine like dermatology, is a slower-adopting sector due to regulatory, liability, and workflow integration hurdles, with AI use still mostly at pilot or advisory stage. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can usefully flag differential diagnoses and suggest relevant tests to consider, assisting dermatologists in comprehensive workup planning. However, the augmentation is constrained by the need for human validation and the complexity of individualizing test panels. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools (e.g., image analysis, symptom checkers, decision-support systems) meaningfully assist dermatologists in narrowing differentials and suggesting relevant tests, improving efficiency while the physician retains final judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | AI can suggest diagnostic tests given patient data, but dermatology diagnosis requires integration of subtle visual cues, patient context, and risk stratification that current systems handle inconsistently. The task requires judgment calls on test necessity that AI alone cannot reliably replicate at the 50% time-saving bar. |
| Task automatability | claude-sonnet-5 | 2/5 | AI can suggest differential diagnoses and relevant tests given structured history/exam data, but synthesizing nuanced physical exam findings and patient context into test recommendations still requires clinical judgment that current systems only partially replicate. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Dermatologists must legally own diagnostic decisions and sign off on test orders; liability and malpractice risk are high if AI-driven recommendations lead to missed diagnoses or unnecessary procedures. Regulatory oversight of clinical decision tools and patient expectation for human physician judgment create strong structural barriers. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Ordering diagnostic tests is a licensed medical act requiring physician authorization and liability accountability, making this a hard-barrier task that cannot be legally delegated to AI alone. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI inference costs for dermatological decision support remain modest, but integration with EHRs, clinical oversight, and the cost of false recommendations (unnecessary or missed tests) offset savings. The loaded cost per recommendation does not yet achieve substantial savings versus a dermatologist's time. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI-assisted decision support software has low marginal inference cost, but integration, EHR connectivity, and mandatory physician oversight keep total cost comparable to physician time rather than dramatically cheaper. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While AI models exist for dermatological image analysis and can support test recommendation, no production system reliably recommends the full spectrum of diagnostic tests (patch tests, biopsies, systemic workup) with clinical-grade accuracy across diverse presentations. Products are primarily narrow-scope assistants, not full decision-makers. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Clinical decision support tools and AI diagnostic aids exist and are used in some dermatology practices, but they function as advisory overlays rather than autonomously recommending tests in production workflows at scale. |
Provide dermatologic consultation to other health professionals.
24CI 20–28 · exposure 25 · augmentation 63 · importance 4.3/5 · click for rater detail
Provide dermatologic consultation to other health professionals.
24| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Dermatology practices are adopting AI-assisted image analysis for screening and triage, but actual clinical consultation between professionals remains a high-touch, high-judgment activity with slow organizational adoption of autonomous or minimally supervised AI alternatives. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Healthcare is adopting AI diagnostic aids at a moderate pace, with dermatology being an early use case for image-based tools, but full consultation workflows still involve pilots rather than broad deployment. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can augment consultation by rapidly generating differential diagnoses from images, summarizing literature, or flagging high-risk features, helping dermatologists structure and accelerate their consultation, though the core clinical communication remains physician-driven. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-assisted image analysis and differential diagnosis tools meaningfully speed up dermatologists' ability to advise colleagues, serving as a strong productivity aid while the physician remains responsible for the final consult. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | AI can assist in analyzing skin images and generating preliminary diagnostic suggestions, but dermatologic consultation to other health professionals requires nuanced clinical judgment, integration of patient history, and the ability to discuss differential diagnoses and treatment rationale—tasks that current systems cannot reliably perform end-to-end without significant human oversight. |
| Task automatability | claude-sonnet-5 | 2/5 | AI can provide diagnostic suggestions and image-based analysis to assist consultation, but synthesizing patient context, liability, and nuanced clinical judgment for peer consultation remains beyond full automation today. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Consultation between licensed professionals carries implicit legal and liability requirements; referring physicians expect accountability from a credentialed dermatologist, not an AI system, and malpractice frameworks assume human expertise and judgment in clinical consultation. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Providing formal medical consultation requires a licensed physician; regulatory and liability frameworks prevent AI from independently rendering authoritative consults. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | The cost of AI image analysis plus integration and physician oversight for high-stakes consultation is substantial relative to the alternative of direct physician-to-physician consultation, and the liability exposure means oversight remains heavily human-driven. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | While AI image analysis is cheap per query, a full consultation requires physician oversight and liability coverage, so total cost remains comparable to or only modestly cheaper than a specialist consult. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While AI tools exist for skin lesion classification and image analysis, no deployed product reliably performs the full consulting task (explaining findings, weighing alternatives, tailoring advice to specific clinical contexts) without material errors or narrow applicability that would be acceptable to referring physicians. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Dermatology AI tools (e.g., image classifiers) exist and are used as decision support, but no deployed product independently performs full physician-to-physician consultations reliably in production. |
Diagnose and treat pigmented lesions such as common acquired nevi, congenital nevi, dysplastic nevi, Spitz nevi, blue nevi, or melanoma.
23CI 20–25 · exposure 30 · augmentation 75 · importance 4.9/5 · click for rater detail
Diagnose and treat pigmented lesions such as common acquired nevi, congenital nevi, dysplastic nevi, Spitz nevi, blue nevi, or melanoma.
23| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | While telemedicine and digital dermatology are growing, adoption of autonomous AI diagnosis tools remains limited to pilot programs and research settings. Most practicing dermatologists use AI as a supplementary review tool rather than a primary diagnostic engine, reflecting cautious institutional and regulatory uptake. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Dermatology has been an early adopter of AI-assisted dermoscopy tools, but adoption remains in pilot/adjunct use in select clinics rather than deep, widespread deployment replacing physician judgment. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI-powered dermoscopy and image analysis substantially augment dermatologist productivity by rapidly flagging atypical lesions, organizing image archives, and prompting differential diagnoses. The technology demonstrably improves screening speed and recall without removing human judgment, making it a high-value assistive tool in clinical workflows. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-based lesion analysis tools meaningfully help dermatologists prioritize suspicious lesions and improve diagnostic accuracy, while the physician retains ultimate responsibility for diagnosis and treatment. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI image classification systems show promising accuracy on dermoscopy images in controlled studies, diagnosis requires integration with patient history, clinical examination, and biopsy decisions. Current AI cannot reliably perform end-to-end diagnosis and treatment planning with the necessary safety margin for a task where misclassification directly risks patient harm. |
| Task automatability | claude-sonnet-5 | 2/5 | AI can assist with image-based triage of pigmented lesions but full diagnosis-and-treatment (including biopsy decisions, excision, patient management) requires physical exam, procedural skill, and clinical judgment that current systems cannot fully replace. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Dermatologists are licensed professionals required by law and regulation to make diagnostic and treatment decisions, particularly for melanoma ruling-out. Malpractice liability and standard-of-care expectations create hard barriers: a physician must legally validate any AI-assisted diagnosis before treatment, preventing full substitution. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Diagnosing melanoma and performing biopsies/excisions requires a licensed physician; liability for missed melanoma is severe, and regulations mandate physician sign-off on diagnosis and treatment plans. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI dermoscopy analysis inference is cheap, but integration requires dermatologist oversight, possible confirmatory biopsies, and liability infrastructure. The all-in cost per lesion evaluated remains higher than task displacement would suggest, given the human specialist's bundled decision authority. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI image analysis software has a modest per-scan cost, but it doesn't replace the physician's time for exam, biopsy, and treatment, so overall cost savings versus the full task are limited. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Deployed AI products (e.g., FDA-cleared dermatology AI classifiers, smartphone apps) exist and show competitive sensitivity/specificity in benchmarks, but real-world use reveals material error rates on atypical lesions and limitations on non-Caucasian skin. These products function as aids rather than autonomous diagnostic engines in production settings. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | FDA-cleared dermoscopy AI tools exist for lesion classification support, but they are decision-aids used alongside dermatologists, not standalone diagnostic-and-treatment systems deployed at scale. |
Conduct complete skin examinations.
21CI 11–30 · exposure 22 · augmentation 75 · importance 4.9/5 · click for rater detail
Conduct complete skin examinations.
21| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Adoption of AI-assisted dermoscopy in clinical practice remains limited; most dermatologists still rely on in-person examination and standard magnification, and health systems have been slow to integrate AI tools into routine workflows due to liability concerns and reimbursement uncertainty. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Healthcare, especially hands-on clinical exams, is a slower-adopting sector for full-task automation; AI adoption is concentrated in diagnostic image support rather than replacing the physical exam workflow. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can meaningfully augment dermatologists by flagging atypical lesions, providing confidence scores on benignity, and highlighting lesions that warrant closer inspection or biopsy, substantially improving diagnostic accuracy and efficiency when the dermatologist remains in control of the clinical assessment. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-powered dermoscopy and total body photography systems significantly help dermatologists detect and prioritize suspicious lesions during exams, meaningfully boosting diagnostic productivity while the physician remains in control. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | AI can assist with image classification of skin lesions from photographs, but a complete skin examination requires visual inspection of the entire body surface, palpation to assess texture and depth, and clinical judgment about systemic signs—tasks that demand in-person human contact and multisensory assessment that current AI cannot replicate end-to-end. |
| Task automatability | claude-sonnet-5 | 1/5 | Complete skin examinations require hands-on physical inspection of the entire body surface, palpation, and real-time patient interaction that current AI cannot perform end-to-end without a human physically present.-so no meaningful time savings can be captured with off-the-shelf systems. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Legal and regulatory barriers are substantial: dermatologists must be licensed physicians, patient safety liability for missed diagnoses is high, and malpractice risk for AI-only or inadequately supervised skin cancer screening creates strong professional and legal requirements that a licensed dermatologist perform or directly supervise the examination. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Performing a physical examination on a patient's body is legally restricted to licensed medical professionals, and liability for missed diagnoses (e.g., melanoma) creates strong regulatory and legal barriers to full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI dermoscopy tools have low inference cost, but integrating them into complete examinations still requires a dermatologist's time for the physical exam, clinical interpretation, and documentation—so the all-in cost per complete examination remains comparable to or higher than human-only performance. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI image analysis tools are cheap per image, but since a licensed clinician must still conduct the physical exam, the overall cost is dominated by human labor, making AI at best a marginal cost addition rather than a substitute. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Dermoscopy AI tools and lesion classification systems exist and perform well on curated images in clinical settings, but they operate on isolated lesion photos rather than complete examinations, and require dermatologist oversight to integrate findings into diagnostic and treatment decisions. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | AI-based dermoscopy image classifiers exist and are deployed for lesion analysis, but no product performs a complete physical skin exam autonomously; these tools assist rather than replace the exam itself. |
Prescribe hormonal agents or topical treatments such as contraceptives, spironolactone, antiandrogens, oral corticosteroids, retinoids, benzoyl peroxide, or antibiotics.
20CI 20–20 · exposure 25 · augmentation 75 · importance 4.3/5 · click for rater detail
Prescribe hormonal agents or topical treatments such as contraceptives, spironolactone, antiandrogens, oral corticosteroids, retinoids, benzoyl peroxide, or antibiotics.
20| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | While dermatology is a digitization-forward specialty, adoption of AI for prescription automation is minimal because of strict regulatory and liability constraints. Most adoption focuses on diagnosis support, not prescription delegation. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Healthcare, especially clinical prescribing, is a highly regulated sector with slow AI adoption for autonomous decision-making, though diagnostic support tools are gradually being piloted. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | Current AI systems meaningfully assist dermatologists by suggesting evidence-based treatments, flagging drug interactions, and helping organize patient history—tasks that accelerate prescribing decisions while the physician retains full authority and accountability. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools can assist dermatologists by analyzing skin conditions, suggesting differential diagnoses, and recommending evidence-based treatment protocols, meaningfully speeding up the prescribing decision process. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | AI can assist in recognizing when a prescription may be indicated and drafting recommendations, but dermatologists must exercise clinical judgment about patient-specific contraindications, drug interactions, comorbidities, and treatment history. Current AI lacks the contextualized reasoning to autonomously prescribe; the full task requires physician decision-making and legal accountability. |
| Task automatability | claude-sonnet-5 | 2/5 | Prescribing requires clinical examination, diagnosis, and legal accountability for the specific patient; AI can suggest options but cannot independently prescribe, so full end-to-end automation with equal quality is not achievable today. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Prescribing is legally restricted to licensed physicians; state and federal regulations require a healthcare provider to sign and take liability for each prescription. Autonomous AI prescription would violate medical practice acts and drug dispensing laws, creating an insurmountable regulatory barrier. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Prescribing medication is strictly regulated and requires a licensed physician's authorization; this is a hard legal barrier that cannot be bypassed by AI systems. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | The cost of AI-assisted prescription drafting (oversight, verification, liability management) remains comparable to or exceeds the marginal cost of a dermatologist's time for this task, since the physician's cognitive review is mandatory and cannot be eliminated. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | While AI-assisted diagnostic suggestions are cheap to run, the requirement for licensed physician review and legal prescribing authority means the overall cost of the task pathway still centers on physician time. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Deployed AI systems can suggest appropriate medications based on condition classification, but no production system independently prescribes controlled or restricted drugs for dermatological use. Clinical decision support tools exist but stop short of autonomous prescription generation that meets medical-legal standards. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Clinical decision support tools exist that suggest treatment options based on symptoms/images, but no deployed product autonomously prescribes medications for dermatologic conditions in production. |
Refer patients to other specialists, as needed.
17CI 14–20 · exposure 25 · augmentation 63 · importance 4.1/5 · click for rater detail
Refer patients to other specialists, as needed.
17| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Adoption of AI for autonomous referral decisions in dermatology is minimal; practices remain heavily dependent on dermatologist judgment due to liability, regulatory oversight, and lack of proven, validated AI systems for this decision point. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Healthcare, especially clinical decision-making tasks, adopts AI slowly due to regulatory, liability, and safety concerns despite growing use of documentation tools. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can provide useful assistance by suggesting relevant specialists based on lesion characteristics, differential diagnoses, or clinical flags, helping dermatologists recognize referral opportunities more systematically and reducing cognitive load in complex cases. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can meaningfully assist by summarizing patient records, suggesting relevant specialists, and drafting referral documentation, saving physician time while they retain final judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can assist in identifying when specialist referral is clinically appropriate (e.g., flagging oncology-relevant lesions), the final referral decision requires nuanced medical judgment integrating patient history, preferences, and coordination that remains largely human-dependent. Current AI cannot reliably make this end-to-end decision with 50% time savings at equal quality. |
| Task automatability | claude-sonnet-5 | 2/5 | Referral decisions require clinical judgment integrating patient history, exam findings, and specialist availability, which AI cannot fully execute end-to-end today despite drafting support. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Strong legal and regulatory barriers exist: dermatologists hold medical licenses and malpractice liability, and the referral decision is embedded in the clinical care pathway that requires physician judgment and accountability. Healthcare regulation mandates that a licensed physician direct patient care transitions. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Referrals are a licensed medical act requiring physician judgment and legal accountability, making this a hard regulatory and liability barrier to automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The cost of integrating, maintaining, and overseeing an AI system for referral recommendations, combined with liability considerations, likely exceeds the cost of a dermatologist's few minutes per patient to make this decision themselves. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI could cheaply draft referral letters, but the clinical judgment and liability of the referral decision still require physician time, limiting overall cost savings. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed AI system independently makes specialist referral decisions in production dermatology practice. AI tools exist to support triage and suggest referral pathways, but human dermatologists must initiate and justify referrals; the task is not yet reliably automated in real-world clinical settings. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some clinical decision-support tools flag potential referral needs, but no deployed product autonomously makes and executes physician referrals reliably in production. |
Evaluate patients to determine eligibility for cosmetic procedures such as liposuction, laser resurfacing, or microdermabrasion.
14CI 3–25 · exposure 13 · augmentation 50 · importance 2.9/5 · click for rater detail
Evaluate patients to determine eligibility for cosmetic procedures such as liposuction, laser resurfacing, or microdermabrasion.
14| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Dermatology has moderate digitization, and cosmetic procedure screening remains a high-touch, high-liability task where practitioners adopt AI as assistive tools rather than replacements. Adoption is slow in this specific sub-domain despite broader dermatology tech growth. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Clinical dermatology involving direct patient evaluation and procedural decision-making has seen limited AI adoption beyond diagnostic imaging aids; production-level adoption for eligibility determination is essentially absent. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can usefully flag skin conditions, contraindications, and high-risk candidates from images or patient data, helping dermatologists structure their evaluation more efficiently. However, the task's emphasis on holistic clinical judgment and patient suitability limits transformative augmentation potential. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can assist with reviewing patient history, flagging contraindications, analyzing skin images, or supporting documentation, improving efficiency without replacing the clinical evaluation itself. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | AI can assist with image analysis and risk screening but cannot replace the full clinical evaluation. Dermatologists must assess medical history, medications, skin type, realistic expectations, and psychological fitness—nuanced judgments requiring direct patient interaction and medical liability assumption that AI cannot fully automate today. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires hands-on physical examination, medical history assessment, risk stratification, and clinical judgment about procedure eligibility that cannot be performed end-to-end by current AI systems. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Legal and liability barriers are high: dermatologists must personally evaluate and sign off on cosmetic procedure eligibility, and medical boards hold them accountable for adverse outcomes. No AI system can absorb that legal responsibility or substitute for the physician's clinical judgment and signature. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Determining eligibility for invasive/semi-invasive cosmetic procedures requires a licensed physician's medical judgment and carries direct liability, making this a hard-barrier task legally requiring human authorization. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | A dermatologist's loaded cost for a procedure eligibility consultation is substantial; AI image and risk tools cost less per unit but require dermatologist oversight, review, and legal sign-off, making the all-in cost structure unfavorable compared to direct human evaluation. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | Since AI cannot perform the actual patient evaluation and eligibility decision, there is no valid cost comparison—the physician must still perform this task, so AI adds cost without substituting. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While AI tools exist for skin image classification and some risk flagging, no deployed product reliably performs comprehensive cosmetic procedure eligibility evaluation end-to-end. Clinical decision support exists but human dermatologists remain essential for the multi-factor assessment and sign-off. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product independently evaluates patients for cosmetic procedure eligibility; AI tools at best support image analysis or documentation, not the clinical eligibility determination itself. |
Perform skin surgery to improve appearance, make early diagnoses, or control diseases such as skin cancer.
4CI 0–7 · exposure 5 · augmentation 38 · importance 4.8/5 · click for rater detail
Perform skin surgery to improve appearance, make early diagnoses, or control diseases such as skin cancer.
4| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Dermatological surgery is performed primarily in hospital and clinic settings with strong professional gatekeeping and risk aversion; adoption of autonomous surgical AI is minimal and limited to niche research contexts, not mainstream clinical deployment. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Healthcare/surgical fields adopt AI slowly for physical procedures due to regulatory, safety, and liability constraints, though diagnostic imaging AI adoption is faster elsewhere in the field. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI assists dermatologists in image analysis for diagnosis and biopsy site selection, but provides minimal real-time augmentation during the manual surgical act itself. Pre-operative planning support exists, but intraoperative augmentation remains limited. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can assist with pre-surgical diagnosis, lesion risk assessment, and imaging analysis to guide decisions, improving accuracy and efficiency without replacing the surgical act itself. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Dermatological surgery requires manual dexterity, real-time visual feedback, three-dimensional spatial judgment, and adaptive decision-making during the procedure itself. Current AI systems cannot perform surgical interventions end-to-end; robotics in dermatology remain experimental and require continuous human control. |
| Task automatability | claude-sonnet-5 | 1/5 | Skin surgery requires physical dexterity, real-time tactile judgment, and hands-on manipulation of tissue that current AI systems cannot perform; no robotic system autonomously performs dermatologic surgery today. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Dermatological surgery is legally restricted to licensed medical doctors; malpractice liability, tissue damage risk, and regulatory oversight (FDA for surgical devices) create hard barriers. The human physician must remain legally and clinically responsible for the surgical outcome. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Surgical procedures require licensed physician performance, direct liability, sterile technique, and regulatory oversight, representing hard legal and safety barriers to automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The capital cost of surgical equipment, maintenance, and integration infrastructure far exceeds the savings from any partial automation; a dermatologist's surgical labor remains cheaper than the amortized cost of surgical AI systems with required oversight and liability. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI substitute performing the physical surgery, so cost comparison is moot—human surgeons remain the only option, making AI effectively infinitely costlier for this task as automation. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While AI excels at dermoscopic image classification for diagnosis support, no deployed product performs actual skin surgery autonomously or semi-autonomously in routine clinical practice. Surgical robots in dermatology exist in research settings but lack production-ready clinical integration for the full spectrum of skin surgeries. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs autonomous skin surgery; AI is used only for adjunctive image analysis (e.g., lesion classification), not the procedural task itself. |
Instruct interns or residents in diagnosis and treatment of dermatological diseases.
1CI 0–3 · exposure 0 · augmentation 63 · importance 4.1/5 · click for rater detail
Instruct interns or residents in diagnosis and treatment of dermatological diseases.
1| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Medical education remains heavily dependent on human mentorship and direct oversight; teaching role automation is not occurring in practice. Adoption of AI as a teaching tool (not replacement) is nascent and limited to supplementary content generation. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Medical education is a slow-adopting, highly institutionalized sector where AI is used for supplementary learning materials but not to replace mentorship. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist dermatologists in preparing teaching materials, summarizing literature, or generating case libraries for discussion, improving the efficiency of lesson planning. However, the instructional interaction itself remains fundamentally human-driven. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can meaningfully support teaching via case simulations, diagnostic decision aids, and quizzing residents, enhancing the education process while the attending remains central. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Teaching and mentoring requires real-time interaction, adaptive explanation, and evaluation of trainee understanding and clinical reasoning—tasks that demand human judgment and presence. Current AI cannot replicate the Socratic method or personalized feedback loop that characterizes effective clinical instruction. |
| Task automatability | claude-sonnet-5 | 1/5 | Teaching clinical judgment, bedside manner, and hands-on procedural skills to trainees requires live supervision, mentorship, and adaptive interaction that current AI cannot replicate end-to-end." |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Medical licensure and accreditation standards legally require that qualified human physicians supervise and instruct residents; regulatory bodies (ACGME, boards) mandate human oversight of clinical training. Substituting AI for human instruction would violate training program requirements. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Medical education and supervision of trainees is tightly regulated, requiring licensed attending physicians to teach and certify competency. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | A dermatologist's hourly rate for teaching is substantial, and AI systems cannot yet replace this role cost-effectively because the task requires genuine expertise assessment and adaptive guidance that human supervisors provide. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute for this instructional role, so cost comparison favors the human entirely. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product reliably performs clinical instruction and mentorship. While AI can generate educational content or answer questions, it cannot assess trainee competence, adjust teaching strategy based on performance, or bear responsibility for trainee development—core elements of the task. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product substitutes for a supervising dermatologist instructing residents; AI tools exist only as reference aids, not as autonomous instructors. |
Perform incisional biopsies to diagnose melanoma.
0CI 0–0 · exposure 0 · augmentation 25 · importance 4.8/5 · click for rater detail
Perform incisional biopsies to diagnose melanoma.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | This is a core clinical procedure with no active automation trend; dermatologists continue to perform incisional biopsies as standard practice with no evidence of AI-driven displacement. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Surgical and procedural aspects of medicine remain among the least automated in healthcare, with physical intervention tasks seeing minimal AI adoption compared to diagnostic imaging or documentation. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI can help pre-procedure by identifying suspicious lesions on dermoscopy images or imaging, but offers minimal real-time assistance during the actual incisional procedure itself beyond pre-operative planning. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can assist with pre-procedure planning, lesion risk scoring (e.g., dermoscopy analysis) to decide whether to biopsy, but offers little direct assistance during the physical incision and tissue excision itself. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Incisional biopsy requires skilled surgical technique, precise instrument handling, sterile field management, and real-time clinical judgment about lesion margins and depth—all performed on live tissue. Current AI cannot physically perform invasive procedures or adapt to intraoperative complications. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a physical surgical procedure requiring manual dexterity, sterile technique, and hands-on tissue manipulation that no current AI system can perform end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | A licensed physician must legally perform and take responsibility for the biopsy procedure itself, including anesthesia, sterile technique, and specimen handling. Regulatory and liability requirements create hard barriers to automation. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Performing an incisional biopsy is an invasive medical procedure requiring a licensed physician, informed consent, and legal accountability, making this one of the most protected tasks against automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Incisional biopsy is a clinical procedure requiring a licensed dermatologist's direct execution. The human cost is inherent to the task definition; AI cannot substitute for the procedural component. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI system capable of substituting for the physical procedure, so no cost comparison favoring AI exists; a human physician remains the only option. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI system performs incisional biopsies autonomously. While AI assists in melanoma diagnosis from imaging, the surgical procedure itself remains entirely human-performed in clinical practice. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs incisional biopsies; robotic surgery systems exist for other procedures but none autonomously execute skin biopsies for melanoma diagnosis. |
Provide therapies such as intralesional steroids, chemical peels, or comodo removal to treat age spots, sun damage, rough skin, discolored skin, or oily skin.
0CI 0–0 · exposure 0 · augmentation 25 · importance 4.0/5 · click for rater detail
Provide therapies such as intralesional steroids, chemical peels, or comodo removal to treat age spots, sun damage, rough skin, discolored skin, or oily skin.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | There is no meaningful adoption of AI for performing these therapies because the task is medically and legally bound to licensed practitioners. No sector trends toward automation of hands-on clinical dermatology procedures. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Procedural, hands-on dermatologic care shows minimal AI displacement; adoption in this specific physical-treatment niche is essentially nonexistent despite AI diagnostic tools advancing elsewhere in dermatology. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI might assist with pre-procedure imaging analysis, patient triage, or treatment planning documentation, but it offers minimal augmentation for the actual therapy delivery itself, which remains entirely manual and tactile. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can assist with pre-procedure diagnostics, treatment planning, or documentation, but offers negligible support during the actual manual execution of these therapies. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires in-person administration of intralesional injections, chemical peels, and mechanical removal procedures that demand physical manipulation of patient skin. Current AI systems cannot perform these hands-on clinical interventions, and no automation could achieve the 50% time-saving threshold for the core work. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a hands-on physical procedure requiring injection, chemical application, and manual extraction directly on a patient's skin; no current AI system can physically perform these interventions. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Medical licensing law explicitly requires that a licensed dermatologist perform or directly oversee these invasive procedures. There are strict regulatory requirements around drug administration, informed consent, and liability that create hard legal barriers to substitution. |
| Adoption barriers | claude-sonnet-5 | 5/5 | These are invasive or semi-invasive medical procedures requiring a licensed physician, direct physical contact, and legal accountability for patient safety and outcomes. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | AI has no cost advantage here because the task is not automatable at all. The procedure delivery is inseparable from the dermatologist's billable time, and any oversight or quality assurance would add cost rather than reduce it. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | Without any functional AI system capable of performing the physical procedure, there is no viable AI cost comparison—human labor is currently the only option. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI system can autonomously administer injections, apply chemical peels, or perform comedone extraction. These are skilled procedural tasks requiring real-time tactile feedback, patient assessment, and sterile technique that remain firmly in the human-only domain. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs intralesional injections, chemical peels, or comedo extraction; robotics for such delicate dermatologic procedures remain research-stage at best. |
Provide dermabrasion or laser abrasion to treat atrophic scars, elevated scars, or other skin conditions.
0CI 0–0 · exposure 0 · augmentation 25 · importance 3.2/5 · click for rater detail
Provide dermabrasion or laser abrasion to treat atrophic scars, elevated scars, or other skin conditions.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Procedural dermatology remains human-dependent; no significant adoption of autonomous AI systems for laser or dermabrasion treatments is occurring, nor is it plausible given regulatory and safety constraints. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Procedural dermatology is a hands-on, in-person medical specialty with minimal AI-driven automation of the physical procedure itself, reflecting slow adoption in physical/manual clinical tasks. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could modestly assist in pre-procedure planning (analyzing skin imaging, recommending parameters) but offers minimal real-time augmentation during the procedure itself, which remains largely dependent on the dermatologist's expertise and judgment. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can assist with pre-procedure planning, scar assessment, or laser parameter recommendations via imaging analysis, but offers limited real-time assistance during the actual physical performance of dermabrasion or laser treatment. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task involves precise surgical/procedural intervention using specialized medical equipment on live human tissue, requiring real-time tactile feedback, safety margins, and immediate complication management—capabilities far beyond current AI systems. No AI can autonomously operate dermatological laser or dermabrasion equipment on patients. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a physical, hands-on procedural task requiring manual dexterity and real-time tactile/visual judgment during abrasion of skin tissue; no AI system can perform the physical manipulation of dermabrasion or laser devices end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | This task has hard legal barriers: dermatologists must be licensed physicians, and state medical boards explicitly prohibit unlicensed entities from performing invasive procedures. Malpractice liability, patient safety requirements, and regulatory oversight (FDA for devices) create insurmountable adoption barriers. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Performing invasive skin procedures requires a licensed physician (or supervised practitioner), with strict medical licensing, liability, and safety regulations that legally require human performance and judgment. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The equipment, training, malpractice liability, and clinical oversight required for this procedure make human dermatologists performing it far cheaper than any theoretical AI alternative, which would require custom robotics, regulatory approval, and extensive validation. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI system that can substitute for the physician performing this procedure, so no meaningful cost comparison for full task replacement exists; the physician's cost remains necessary. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI product can perform this procedural task independently. While diagnostic AI exists for skin conditions, the execution of ablative procedures requires licensed medical professionals with specialized training and cannot be delegated to autonomous systems today. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product autonomously performs dermabrasion or laser resurfacing on patients; robotic/AI-assisted laser systems exist only as physician-operated tools, not autonomous performers of the procedure. |
Related occupations — Healthcare Practitioners & Technical
How to read this
A high substitution score does not mean this job disappears — it means a large share of its current tasks face replacement pressure, so the mix of tasks is likely to change. High augmentation alongside substitution typically means the occupation reorganizes around the protected tasks. Wide confidence intervals mean the rater panel disagreed: treat those scores as open questions, not verdicts.
What would change this score
New model capabilities (automatability, feasibility), falling inference costs (cost ratio), regulation and licensing shifts (barriers), and measured sector adoption (velocity) all re-enter at every index release. Each release is recomputed, versioned and kept queryable — scores are claims with a date on them, not permanent labels.