SAFE
Lesion image classification is the one dermatology task where AI genuinely competes — CNNs already match dermatologists on curated melanoma photo sets — but that is a slice of the job, not the job. The median dermatologist spends the day doing full-body skin exams, dermoscopy, punch and shave biopsies, cryotherapy, excisions, cosmetic injectables and lasers, plus prescribing systemic immunosuppressants and biologics that require a licensed prescriber and personal liability for outcomes. AI will most likely arrive as triage that raises the volume of referred lesions the dermatologist must physically inspect and cut.
Headcount grew steadily across the period.
This line is counted by the Bureau of Labor Statistics — the one figure on this page that isn't a judgement of ours. Headcount moves on demand, offshoring, demographics and the business cycle, and automation is one term among several, often not the loudest.
So a falling line is not evidence that AI did it, and a rising one is not evidence that it won't. Both happen in this register: some occupations resist automation and shrink anyway, others are highly automatable and keep growing.
BLS projection, 2024–2034
+6.4% 10,900 → 11,600 on the projections basis
Hard to automate, and growing
The work resists current AI and the BLS projects +6.4% more of these jobs by 2034. Note that safe does not mean well paid — several of the fastest-growing resistant occupations are among the lowest paid on the register.
Different clocks. The score is what current AI could do to this work today. The projection is how many of these jobs will exist in 2034. Everything between the two — how fast employers actually adopt, whether demand grows in the meantime — is why they can point opposite ways without either being wrong.
~400 openings a year on average, including replacing people who leave.
DoctorMohs SurgeonDermatologistDermatopathologistMD (Medical Doctor)Dermatology PhysicianGeneral DermatologistMedical DermatologistClinical DermatologistDermatological SurgeonDermatologist PhysicianPediatric DermatologistPracticing DermatologistProcedural DermatologistMohs Micrographic SurgeonBoard Certified DermatologistDermatologist MD (Dermatologist Medical Doctor)DO Physician (Doctor of Osteopathic Medicine Physician)
Holding it up: liability shield . Weakest point: task resistance .
Mixed — a routine tier and a judgment tier A 13 rather than 17 reflects that image-based triage of pigmented lesions and teledermatology store-and-forward reads are genuinely contestable by CNNs, while the same day's punch biopsies, Mohs-adjacent excisions, intralesional steroid injections and palpation of scaly plaques for texture and induration have no digital substitute — so roughly a fifth of billable encounters are exposed, not the core.
Hands-on in uncontrolled environments Full-body skin checks under Wood's lamp, dermoscopy contact plate on the patient's back, liquid nitrogen sprays, shave and punch biopsies with suture closure, and laser and injectable work all happen with gloved hands on a live patient in an exam room — it sits at 15 rather than 19 because the environment is a controlled clinic, not a roadside or a crawl space.
Licensed human required and personally liable Prescribing isotretinoin under iPLEDGE, dosing methotrexate and cyclosporine, and signing off on biologics like dupilumab all require an unrestricted state medical licence plus board certification, and the named physician carries the malpractice exposure when a missed melanoma metastasises — 19 rather than 20 only because some cosmetic procedures can be delegated to supervised NPs and PAs.
Exists to be accountable for ambiguous calls Deciding whether an atypical nevus warrants excision with 5 mm margins or annual photography, whether a rash is a drug eruption or early cutaneous lymphoma, and whether to start a systemic immunosuppressant in a patient with latent TB are calls with no protocol that resolves them and real downside either way — 16 rather than 19 because staging and treatment of confirmed melanoma follows NCCN pathways.
The verdict above describes this occupation as a whole. Almost nobody does the typical version of a job — tick what's actually in your week and see how your own mix sits.
Your task mix speaks to task resistance (13/20 here) — how much of the day's work current AI already does. That is the dimension the boxes above are about.
It cannot move the other three. Liability shield (19/20) is whether the law requires a licensed human to sign. Trust premium (15/20) is whether buyers specifically pay for a person. Judgment and accountability (16/20) is whether the role exists to own consequential calls. Those are facts about the occupation's standing, not about which tasks are in your week — a paralegal who does only trial exhibits still holds no licence. Together they are 50 of this occupation's 78 points (64%).
Embodiment (15/20) is also a property of the work rather than the worker, but we don't tag individual tasks as physical or not, so the picker can't tell you anything about it. That's a limit of this tool, not a claim.
Did we get the list right? Tell us what's missing — the tasks are written from the outside, and you're reading this from the inside.
The moves above are yours to make. This is the other half: what would have to change in the world for the occupation itself to score higher. None of it is in any one person's gift, but it is where the floor actually comes from. Scores here are not a one-way ratchet. Only two of the five dimensions — task resistance and embodiment — track what machines can do. The other three track law, what buyers will pay for, and who is answerable, and those move in both directions, often in response to the same pressure AI creates. If every lever below landed, this occupation would score around 87/100, still SAFE.
State medical boards or FDA labeling that explicitly requires a board-certified dermatologist to review and sign every AI-flagged lesion triage output before biopsy or discharge — the pattern already set by FDA's De Novo clearance of DermaSensor and SkinVision-type tools as adjunct-only, plus state teledermatology rules (e.g., Texas, Arkansas) barring store-and-forward diagnosis without a licensed physician of record. Medical malpractice insurers adding riders that void coverage when an AI dermatology triage result is accepted without documented physician review would push the same direction.
Task-mix shift: if AI triage absorbs the routine 'is this benign nevus' read, the residual day is dermoscopy on ambiguous lesions, biopsy technique, immunosuppressant/biologic selection and monitoring, and complex inflammatory and pediatric disease — the tier CNNs on curated photo sets do not touch. Watch for AI-first triage in large dermatology groups (e.g., private-equity-consolidated practices) shifting scheduled slots toward procedures and systemic-therapy management.
Volume shift toward procedural work — Mohs surgery, excisions, cryotherapy, laser and injectable cosmetics — if AI triage increases referred-lesion throughput as expected. These are unpredictable-field manual procedures on live tissue with no near-term robotic substitute; a rising procedural share of RVUs would be the observable marker in Medicare billing data.
Cosmetic dermatology is cash-pay and already brands on the specific physician; if med-spa AI/device scandals drive state legislation requiring physician (not NP/PA/aesthetician) performance of energy-based procedures — bills of this kind have been introduced in Florida and Nevada — the named-human premium hardens legally as well as commercially.
Ownership of the biologic and immunosuppressant decision under ambiguity — infection risk, malignancy screening, pregnancy, insurance step-therapy appeals — grows if prior-authorization reform requires a named prescribing specialist to attest to medical necessity rather than a plan algorithm.
The limit. At 78 the headroom is small and mostly in embodiment and task-mix; liability_shield at 19 and judgment_accountability at 16 are already near the practical ceiling for a licensed procedural specialty. The realistic downside risk is not displacement but scope competition from NPs/PAs and AI-assisted primary care handling routine reads, which compresses volume without touching any of these five scores.
| New York-Newark-Jersey City, NY-NJ | 2,040 | $374,000 +14% |
| Houston-Pasadena-The Woodlands, TX | 370 | $207,000 -37% |
| Charlotte-Concord-Gastonia, NC-SC | 180 | $276,170 -16% |
| Minneapolis-St. Paul-Bloomington, MN-WI | 160 | — |
| Cincinnati, OH-KY-IN | 130 | $266,290 -19% |
| Albany-Schenectady-Troy, NY | 100 | — |
| Dayton-Kettering-Beavercreek, OH | 100 | — |
| Portland-Vancouver-Hillsboro, OR-WA | 90 | $239,200 -27% |
| Greensboro-High Point, NC | 80 | $457,770 +39% |
| Kiryas Joel-Poughkeepsie-Newburgh, NY | 80 | $405,000 +23% |
| New York-Newark-Jersey City, NY-NJ | 2,040 | $374,000 +14% |
We have no reported case of a named organisation automating this occupation. Not one deployment, not one announcement.
That is worth saying out loud next to a score of 78. The verdict above is about what the work exposes — what current AI could do to these tasks. It is not a claim that anyone has done it. For this occupation those two things have come apart completely: the capability argument is on this page, and the evidence column is empty.
Has AI actually changed your work? One tap, anonymous, and the running tally is public. Nothing else is asked of you.
Rather than check back: get the digest and we'll tell you what changed — or watch a single occupation from its own page.