Proofreaders and Copy Markers: COOKED

COOKED

The core work — catching spelling, grammar, punctuation, and typographical errors, checking type against copy, marking corrections with standard proof marks — is the exact task class LLMs and integrated editorial tools already do at usable quality and near-zero marginal cost. What remains is thin: house-style enforcement across long manuscripts, legal/factual sensitivity reads, and being the named human who signs off before a print run or regulated filing goes out. Employment is already tiny (4,580) and shrinking as the function gets absorbed into editors' and production staff's workflows.

10-year outlook: By the mid-2030s standalone proofreading roles will be largely gone, with the surviving work folded into a smaller number of editorial QC and release-approval jobs in high-liability publishing.

18/100 risk resistance · ±3 on re-score

Score — 18/100 resistance

Five dimensions, 0–20 each, summed. Higher means more protected. The arithmetic is shown so you can check it: 3 + 1 + 0 + 5 + 9 = 18.

Task resistance 3/20

Core tasks are already automatable. Reading galleys against manuscript for transposed letters, bad hyphenation, widow/orphan lines and misapplied proof marks is character-level pattern matching that spellcheck, PerfectIt and an LLM prompted with a style sheet now do in one pass, which is why the score sits at 3 rather than 7 — even the residual queries-to-author step is being drafted by the same tools.

Embodiment 1/20

Fully desk- and screen-based. The only physical act left is handling a hard-copy galley or a blueline with a red pencil, and most shops moved to PDF markup in Acrobat or tracked changes years ago, so 1 rather than 0 reflects that a few print and packaging proofers still check press sheets under a light booth.

Liability shield 0/20

No licence, no signature requirement. There is no state licence, no board, and no statute naming a proofreader — an uncaught error in a prospectus or a drug label lands on the editor, the production manager or the filing attorney, and your name appears nowhere on the sign-off that matters.

Trust premium 5/20

Anonymous artifact production. Freelance proofers do accumulate repeat clients who trust them with a specific house style and a known author's tics, which is why this is 5 and not 0, but the deliverable is a clean file — the publisher can swap in another proofreader mid-list and no reader will ever know your name.

Judgment & accountability 9/20

Meaningful discretion. Deciding when a serial comma changes meaning, when to query the author versus silently fix, and when a factual slip is grave enough to stop the print run is real discretion, but it is exercised inside a style manual with an editor above you who can overrule any mark you make.

Confidence: high · reviewed 2026-08-11 · how scoring works

Tasks already automatable

spelling, grammar, and punctuation correction
flagging typographical errors in typeset proofs
style-guide consistency checks
comparing typeset copy against source manuscript
marking standard proofreader symbols on copy

What survives

final human sign-off before print or regulated release
catching factual and legal landmines machines miss
enforcing bespoke house style and brand voice judgment
reconciling conflicting author/editor/legal markup
quality control on complex layouts, tables, and figure callouts

Active moats: judgment

How to future-proof this job

Move to the sign-off tier: own final release approval for print runs, packaging, or regulated documents where an error is expensive and someone must be accountable.
Specialize in domains where errors carry liability — pharma labeling, financial prospectuses, legal filings, medical copy — where subject knowledge plus verification beats generic grammar checking.
Become the editorial QC lead who configures and audits AI proofing passes, then adjudicates the flagged items and maintains the house style rulebook.
Add production/layout skills (InDesign proofing, typesetting checks, figure and table verification) so your value spans copy and page, not copy alone.

Training paths for your skill gaps: Structured problem solving partner link

Escape hatches — adjacent fields with better verdicts

Computed from U.S. Dept. of Labor O*NET skill profiles: high overlap with what you already do, materially higher resistance score.

Judicial Law Clerks EXPOSED · 35/100 · you already have ~64% of the skill profile

Skills to close: Active Learning, Critical Thinking, Active Listening, Complex Problem Solving

Locker Room, Coatroom, and Dressing Room Attendants EXPOSED · 40/100 · you already have ~64% of the skill profile

Skills to close: Service Orientation, Social Perceptiveness, Negotiation, Operations Monitoring

Field report — do you do this job?

Has AI actually changed your work?

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

From people who do this job

Nobody has filed one yet. If you do this work, you know things the rubric can't see.