■ CRITICAL RISK ■ Administrative
Yes — dictation machines started this replacement, and modern speech recognition finished it. Shorthand as an occupation is essentially extinct outside court reporting and a few parliamentary and journalistic holdouts, and AI transcription is pressing on those too.
“Speech-to-text is faster and actually legible.”
Our AI replacement risk score — how we score jobs
Shorthand writers — stenographers in the office sense — captured spoken words at speed: dictated letters from executives, meeting minutes, interviews, and proceedings, all rendered in Pitman, Gregg, or Teeline squiggles and then transcribed into clean text. Speed certification was the credential; 100-plus words per minute was table stakes, and elite writers touched 200. It was once a mandatory skill for an entire tier of office work.
The tools came for it in stages. Dictaphones removed the need for a human present during dictation. Word processors removed the retyping. Then speech recognition removed the transcription itself: modern engines convert clear speech to text in real time, free, on the phone in everyone's pocket, and meeting platforms now generate transcripts and AI summaries of every call automatically. When the capture, the transcription, and even the summarization are software features, there's no task left to certify at 120 words per minute. Our 96 describes an occupation that survives mainly as a skill listed on older CVs.
The holdouts are instructive. Court and parliamentary reporters still use machine shorthand (stenotype) because legal proceedings demand verified, speaker-attributed, accountable records — and crosstalk, jargon, and mumbled testimony still trip automated systems. Some journalists keep pen shorthand for interviews where recording is impractical or forbidden; in a few countries it remains a journalism-school requirement. But these are adjacent occupations with their own training pipelines, not a market for shorthand writers. Even court reporting faces digital-recording-plus-AI-transcription pressure, defended mostly by accuracy requirements and, frankly, by a shortage of stenographers making the argument moot.
Automatability: our editorial assessment of current and near-term AI capability
The occupation's collapse is decades old — dictation machines and word processors gutted it, and consumer-grade speech recognition removed the last rationale in the 2010s. What's happening now is the siege of the final redoubt: courts and parliaments weighing AI transcription against certified human reporters. Expect verbatim-reporting niches to persist for years on accuracy and accountability grounds, while general shorthand work stays extinct.
As an office occupation, yes — speech recognition and automatic meeting transcription eliminated its purpose. As a skill it survives in pockets: stenotype-based court reporting and captioning, some journalism traditions where recording isn't always possible, and personal note-taking. Nobody is hired as a 'shorthand writer' anymore; a few professions still use shorthand inside a different job.
Accountability and accuracy under bad conditions. Legal records must be certified, speaker-attributed, and defensible on appeal, and courtrooms feature crosstalk, mumbling, jargon, and emotional speech that still degrade automated transcription. A certified reporter can interrupt and ask for repetition; a microphone can't. That said, many jurisdictions now run digital recording with human review, so the moat is narrowing.
Only for the specific court-reporting and captioning path — and there it's genuinely undersupplied, since training pipelines shrank faster than demand. Realtime captioners and court reporters can earn well precisely because so few people finish the training. Learning shorthand for general office work, though, is like learning telegraphy for a communications career.
Trade on the adjacent strengths: precision, discretion, and documentation speed. Legal-support roles, compliance documentation, medical scribing, and executive support all value those. If speed certification is current, broadcast captioning and CART (realtime captioning for deaf audiences) are growing, meaningful applications of the same core skill.