← Risk register SOC 31-9094 · reviewed 2026-08-11

Medical Transcriptionists

41,550 US workers · median $40,410/yr · Healthcare Support

COOKED

The core task — converting dictated physician audio into structured clinical documents — is the single most directly solved task in healthcare support: medical-grade speech recognition plus LLM formatting now produces drafts at or above human first-pass quality, and ambient scribing tools skip the dictation step entirely. What remains is editing and quality assurance: catching drug-name and dosage errors, flagging internal inconsistencies, and enforcing template and payer-specific documentation rules, which is real judgment but requires a fraction of the headcount. No license is required and the physician, not the transcriptionist, signs the note.

10-year outlook: Employment keeps shrinking sharply through the 2030s; the surviving jobs are retitled as documentation-integrity, coding, or AI-note QA roles inside health systems rather than transcription per se.

US employment, 2019–2025-25.5%
55,78041,550 workers

Part 2020 shock, part continued decline in the years since.

Median pay $33,380 → $40,410 -3.2% in real terms (nominal +21.1%, less ~25% US inflation over the period)

The job count is not the verdict

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. The marked year is 2020.

BLS projection, 2024–2034

-4.9% 43,900 → 41,800 on the projections basis

Exposed, and shrinking

Both signals point the same way: the tasks are largely automatable and the BLS projects -4.9% by 2034. This is the case where the score and the forecast agree, and it is the one worth taking seriously.

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.

~7,400 openings a year on average, including replacing people who leave.

One email if this score changes. Watch as many occupations as you like from the same address — no account, and nothing is sent on a schedule, only when a verdict actually moves.

Also known as — 24 job titles this covers

Titles reported by people doing this work, from the US Department of Labor's O*NET survey. If your job title is here, this page is about your work even though the name doesn't match.

ScribeTranscriberMedical ScribeClinical ScribeProvider ScribeData TranscriberTranscriptionistOphthalmic ScribeMedical TranscriberMedical StenographerMedical Records ClerkDocumentation SpecialistEmergency Medical ScribeMedical TranscriptionistTranscription SpecialistMedical Record TranscriberPathology TranscriptionistRadiology TranscriptionistMedical Language SpecialistMedical Transcription EditorMedical Language ProfessionalTranscribing Machine OperatorClinical Medical TranscriptionistCertified Medical Transcriptionist

Score — 15/100 resistance

Holding it up: judgment & accountability (8/20). Weakest point: embodiment (1/20).

Five dimensions, 0–20 each, summed. Higher means more protected. The arithmetic is shown so you can check it: 2 + 1 + 2 + 2 + 8 = 15. · Scored 2026-08-11, and re-examined when evidence accumulates rather than on a schedule.

Task resistance 2/20

Core tasks are already automatable Dictation-to-document conversion is what ASR was built for — Dragon Medical and Nuance DAX already deliver 98%+ accuracy on drug names and anatomical terms, and ambient listening tools generate the H&P or operative note from the room audio without anyone dictating at all, leaving no core task on your daily queue that a model does not already draft.

Embodiment 1/20

Fully desk- and screen-based The whole shift is a headset, a foot pedal, and a text window — often from home — with no patient contact, no specimen handling, and nothing that requires being in the building.

Liability shield 2/20

No licence, no signature requirement AHDI's RHDS/CHDS credentials are voluntary and no state licenses transcription; the attending physician attests and signs the note, so a mis-transcribed dosage becomes a documentation-error claim against the practice, not a license action against you.

Trust premium 2/20

Anonymous artifact production Work arrives as anonymous audio files in a queue, frequently through an offshore or domestic MTSO the dictating physician cannot name, and turnaround-time SLAs — not any relationship with the clinician — determine whether the account is retained.

Judgment & accountability 8/20

Meaningful discretion Editing an ASR draft means resolving genuinely ambiguous audio — hypo- versus hyper-, 15 versus 50 mg, laterality that contradicts the rest of the note — and deciding whether to fill the blank or flag it for the dictator, which is real discretionary weight on patient safety, but it sits at 8 rather than 14 because style guides (AHDI Book of Style), account-specific templates, and mandatory flag-don't-guess rules constrain nearly every call, and a physician reviews before signature.

Scored twice. An independent second run returned 17/100 — COOKED, agreeing with the verdict above.

Confidence: high · reviewed 2026-08-11 · how scoring works · 29 deployment reports on file

What this job involves — and which parts are yours

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.

AI already does these at usable quality

These still need a person

Active moats on the surviving side: judgment

How to future-proof this job

Training paths for your skill gaps: Apprenticeship.gov — industrial maintenance and millwright programs paid to train · Coursera — engineering and procurement courses, auditable without paying free to audit · MIT OpenCourseWare — circuits, systems and diagnostic method, free and ungated free · Coursera — customer service and client-facing skill courses free to audit

All 35 skills ranked by how many jobs they open →

Where this experience transfers — occupations you could move toward

Computed from U.S. Dept. of Labor O*NET skill and knowledge profiles: high overlap with what you already do, a materially higher resistance score, no large jump in required training, and no licence you would have to start a new pipeline to get. Targets that pay meaningfully less, that are themselves COOKED, or whose own headcount is falling are excluded — a move into a shrinking trade is not an escape.

Ophthalmic Medical Technicians EXPOSED · 55/100 · you already have ~56% of the skill profile

Skills to close: Equipment Maintenance, Equipment Selection, Troubleshooting, Service Orientation

What would move this back up — beyond any one person

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 28/100, still COOKED.

4 specific changes that would raise this score
  • already happening task resistance +4

    Task-mix shift: once first-pass drafting is fully automated, the surviving work is the QA tier — reconciling drug/dose mentions against the med list, catching ASR homophone errors in high-risk fields (mg vs mcg, 'hypo/hyper'), resolving contradictions between ambient-captured dialogue and the structured note, and applying payer- and specialty-specific documentation rules for E/M level support. That residual tier is genuinely harder for models because it requires cross-referencing the EHR record, not just the audio. This raises average per-worker task_resistance even as headcount collapses — the score describes the remaining job, not the number of jobs.

  • already happening judgment accountability +4

    If health systems formalize the editor role as a clinical documentation integrity (CDI) function with authority to hold or query a note before physician signature — the query workflow ACDIS/AHIMA already define for CDI specialists — the role owns a consequential call (does this note support the coded acuity, is this dose plausible) rather than transcribing. Watch for job postings retitling transcriptionist lines as 'CDI specialist' or 'AI documentation QA reviewer' with query authority.

  • plausible liability shield +3

    A Joint Commission or CMS Conditions of Participation requirement that AI-generated clinical documentation carry an attested human review step before entering the legal medical record — analogous to how California AB 3030 requires clinician review of AI-generated patient communications — would create a named reviewer role. Note the ceiling: unless the attestation must come from someone other than the signing physician, hospitals will satisfy it with the physician's own signature and no transcriptionist is needed.

  • plausible liability shield +2

    Malpractice insurers or EHR vendors requiring a documented independent QA pass on ambient-scribe output as a condition of coverage or indemnification — the same pattern insurers used to mandate double-check workflows for high-alert medications. This is a contract requirement, not a license, so it caps low, but it makes the reviewer headcount non-optional.

The limit. No plausible route to trust_premium: patients never see the transcriptionist and no buyer will pay extra for a human-typed note. No route on embodiment. Even with every lever above, this is a role that shrinks by an order of magnitude and survives as a small QA function inside CDI or HIM departments; the score rises for the survivors, not for the occupation's size. Absence of any licensure body for transcription is the hard cap on liability_shield — AHDI's CHDS/RHDS are voluntary credentials with no statutory signing authority, and no state has moved to license them.

These are conditions, not forecasts — what would have to happen, not what will. Specific rules, cases and bills are named so you can go and check whether they exist and where they stand; verify before relying on any of them. Nothing here is legal or financial advice.

Where this work is, and what it pays there

BLS metro figures for 154 areas. The verdict above does not change by city — the rubric judges what the work involves, not where it happens — but pay and headcount do, and the national median hides a very wide range.

Most of these jobs

New York-Newark-Jersey City, NY-NJ 2,910 $44,010 +9%
Los Angeles-Long Beach-Anaheim, CA 2,290 $38,650 -4%
Dallas-Fort Worth-Arlington, TX 1,300 $34,910 -14%
San Antonio-New Braunfels, TX 780 $27,400 -32%
Chicago-Naperville-Elgin, IL-IN 720 $38,700 -4%
San Diego-Chula Vista-Carlsbad, CA 710 $45,420 +12%
Atlanta-Sandy Springs-Roswell, GA 680 $28,330 -30%
San Francisco-Oakland-Fremont, CA 680 $58,240 +44%

Best paid

San Francisco-Oakland-Fremont, CA 680 $58,240 +44%
San Jose-Sunnyvale-Santa Clara, CA 200 $57,420 +42%
Norwich-New London-Willimantic, CT 30 $55,890 +38%

Percentages are against this occupation's national median of $40,410. Counts are jobs in that metro, not vacancies. Metros where the BLS suppressed the cell are absent rather than shown as zero.

Who is actually doing this

The score above is about what the work exposes. This is reporting about real deployments in this occupation — the difference between "could be automated" and "somebody automated it."

Health New Zealand · Cleveland Clinic · Te Whatu Ora · St. Luke's Health System · Mass General Brigham · Carle Health · NHS · Beth Israel Lahey Health · Walter Reed National Military Medical Center · Alberta Health Services / University of Alberta · Rush, McLeod Health, Franciscan Missionaries of Our Lady Health System · Health PEI · Northwest Territories Health and Social Services Authority · Nova Scotia Health · Sharp HealthCare · Cabrini Health · Seoul St. Mary's Hospital · New Zealand general practices · Penn Medicine · US Department of Veterans Affairs · Ardent Health

6 of 42 reported cases, with sources

36 more in the dispatch

Quick take — do you do this job?

Has AI actually changed your work? One tap, anonymous, and the running tally is public. Nothing else is asked of you.

Self-reported and unverified — a sentiment signal, not a survey. One response per person per occupation; you can change your answer.

Field reports — what people say has changed

No field reports yet. A written account takes a paragraph rather than a tap, goes to an editor before it appears, and is the one thing on this page the rubric cannot produce on its own.

File a field report

Concrete beats general: a tool that arrived, a task that moved, a headcount decision you watched happen. Don't include anything that identifies you or your employer if that would put you at risk.

Watch this verdict
Kept current

Rather than check back: get the digest and we'll tell you what changed — or watch a single occupation from its own page.