■ CRITICAL RISK ■ Administrative
Yes — the software took the name first and the job second. Between everyone typing their own documents and AI drafting them from a prompt, the centralized typing role has almost no remaining habitat.
“Your job title is literally a software application.”
Our AI replacement risk score — how we score jobs
Word processors and typists turned other people's words into finished documents: transcribing dictation and handwritten drafts, formatting reports and correspondence to house style, producing tables, proofreading, and cranking through revision cycles. The role assumed a world where professionals generated content but someone else operated the keyboard — typing pools in corporations, document production centers in law firms, transcription desks everywhere.
That world dissolved in stages, each one obvious in retrospect. Personal computers put a keyboard on every desk, and typing stopped being a specialist skill somewhere around the time email made everyone their own correspondent. Speech recognition matured into genuinely accurate dictation, taking the transcription core — medical transcription, the field's last stronghold, has been hollowed out by voice-to-text with human editors surviving only as a review layer. Templates and styles automated formatting. And now generative AI drafts the letter, the report summary, and the meeting minutes from a sentence of instruction, which attacks not just the typing but the document production itself. The job title survives mainly in government classification systems, where postings for 'word processing operator' read like archaeological findings.
The habitable niches are narrow and specific: legal document production centers at large firms still employ specialists for complex formatting under deadline (a table of authorities at midnight remains genuinely skilled work), court transcription retains certified humans because the legal record demands accountability, and accessibility services need real-time captioners — though AI presses hard on all three. For the general role, our 92 might be generous; this is one of the few occupations where the replacement literally shares its name.
Automatability: our editorial assessment of current and near-term AI capability
Mostly historical — the typing pool died with the PC, and speech recognition finished transcription's decline years ago. The current chapter is generative AI eliminating the residual document-production work, drafting and formatting from prompts, which squeezes the remaining legal and specialist niches this decade. Surviving positions concentrate where regulation or complexity demands a certified or highly skilled human, and those keep narrowing.
A dwindling number, mostly in large law firms' document production centers and in government offices whose job classifications update slowly. The typing pool disappeared with the personal computer, and speech recognition plus AI drafting removed most of what remained. Where the title persists, the actual work has usually narrowed to complex legal formatting and overnight deadline coverage.
As general-purpose work, no — automatic speech recognition handles routine transcription at near-zero cost, and rates for human transcription collapsed accordingly. The viable remnants are specialized: court reporting and certified legal transcription (accountability requires a human), difficult audio that defeats the models, and real-time captioning. All require certification or exceptional skill; commodity typing speed no longer sells.
It moves the automation from the keyboard to the author. Earlier waves made typing self-service; generative AI now drafts and formats the document itself from a prompt, meaning fewer documents need a production specialist at any stage. The remaining value sits in verification, complex formatting the AI fumbles, and legal contexts where a human must stand behind the record.
Attach the skill to judgment. Court reporting and CART captioning pay well and require certification that shields them. Legal document specialists remain in demand at big firms. More broadly, proofreading and editing AI-generated drafts is growing work — the models produce volume, and someone detail-oriented must catch what's wrong with it. Typing alone, though, is no longer a career; it's a feature.