A novel exists in one language until somebody translates it — contained, in the strict sense, to the people who read that language — and then, if the translation is good, it exists for everyone else too. A streaming platform ships a show with subtitles in thirty languages on the same day it releases at home. A hospital cannot treat a patient it cannot understand. A refugee's asylum claim cannot be heard by a tribunal that does not speak their language. Translation, interpreting and localisation is the field built entirely around one operation: taking something that exists in one language and making it exist, faithfully, for people in another. Whatever the specific role — a literary translator rendering a novel's voice into English, an NHS-contracted interpreter in an asylum tribunal, a localisation engineer shipping a video game into forty markets on the same release date, a terminologist keeping a pharmaceutical company's safety vocabulary consistent across sixty languages — the shared shape of the work is the same: something contained becomes something widespread, and a person with genuine fluency in two systems of meaning is the mechanism that makes the move possible. The structural pull is Spread, and it is one of the cleanest instances of that gradient in the entire corpus, because unlike most Spread work — which scales an idea or a product that was always meant for everyone — translation exists because the thing being spread was never designed to cross the barrier at all. Someone has to build the bridge, word by word, every time.
The field is not one career; it is at least three, and they have almost nothing in common except the language pair. Translation is written work — a text exists, and the translator has time, reference materials, and the ability to revise before anyone sees the result. Interpreting is spoken and lives entirely in the present tense — there is no draft, no undo, and the interpreter's output is final the moment it leaves their mouth. Localisation is neither, really: it is the industrial process of taking a product — software, a game, a website, a set of legal terms — and re-releasing it, simultaneously, in dozens of languages and markets, which is a systems and project-management problem that a great many localisation professionals solve without being fluent translators themselves at all. A student who says "I'm good at languages, maybe translation" is choosing among three genuinely different working lives, and the field's own literature rarely makes the distinction clearly enough for a seventeen-year-old to act on it.
🦊
There's a guide here if you want one
Kitsune can talk through anything on this page — whether it might suit you, what to do next, questions this page doesn't answer. Everything here is yours to read either way.
MT-plus-post-editing lowering the bilingual-fluency bar for entering commodity translation work
Current fact
The barrier
Producing publishable commodity translation historically required years of building translation-specific speed and fluency on top of bilingual competence.
What changed
MT provides a fluent first-draft scaffold; a linguist with strong subject-matter knowledge but less translation-specific speed can review and correct MT output profitably at 20-30% higher productivity than translating from scratch.
Behaviours involved
Reviewing and correcting MT output rather than translating from a blank page, earlier in a career.
EvaluatingRefining
Building translation judgement through post-editing rather than years of unassisted from-scratch practice.
Reviewing
What this is based on
15% industry workforce growth in 2023 attributed to lowered entry barriers
Widespread LSP adoption of MT-plus-post-editing as a standard service tier
How this could age
Low risk on direction; moderate risk on magnitude, given contested questions about rates and career progression from this on-ramp.
What it does not cover
The same mechanism compresses the classic entry-level from-scratch commodity-translation role; the newly viable path is post-editing, a genuinely different and typically lower-rate job than the one it replaces.
Assessed August 2026
MT-first pipelines letting smaller localisation teams cover far more languages and markets simultaneously
Inference
The barrier
Shipping a product into dozens of languages simultaneously historically required a large in-house or vendor network of professional translators per language.
What changed
MT-first pipelines with human review for market-critical content let a single manager or small team coordinate dozens of languages, extending simultaneous multi-market release to smaller studios and companies.
Behaviours involved
Small-team multi-language release management previously requiring a much larger vendor network.
DistributingSystematizing
What this is based on
Growth of translation-management-system platforms serving small-and-mid-size companies
Industry shift toward hybrid AI-plus-human-review localisation teams
How this could age
Low risk on direction; moderate risk on magnitude.
What it does not cover
Getting the MT-first/human-first split wrong in the wrong market remains genuinely costly; this widens who can attempt multi-market localisation more than it removes the judgement risk of doing it well.
Assessed August 2026
Jobs that did not exist five years ago
These are real jobs that exist now and did not exist before the current wave of AI.
Familiar title, new shape
MT Post-Editor / Linguistic QA Specialist
Reviews and corrects machine-translation output rather than translating from scratch; a standard, widely-hired LSP service tier.
Documented 20-30% productivity advantage over from-scratch translation; standard industry service offering. · current fact
Decides what content goes through MT-first vs human-first pipelines and manages the resulting mixed workflow.
Described directly in Mode 1 as the localisation manager's actual daily decision. · current fact
Familiar title, new shape
AI Dubbing / Synthetic-Voice Localisation QA Specialist
Reviews and directs AI-generated dubbed audio for accuracy, performance quality, and rights compliance.
Tracks the growth of AI dubbing in streaming/game localisation. · early signal
Familiar title, new shape
AI-Literate Linguist / Reviewer
Broader hiring-emphasis shift toward reviewers, specialists, and AI-literate linguists as the field's composition changes.
Named directly in industry trend reporting on the field's evolving hiring emphasis. · seen in the wild
The first two roles are well-evidenced and widely documented industry-wide. The AI dubbing QA role and the broader AI-literate-linguist hiring shift are directionally well-supported but not independently verified against specific named employers this session.
Bars above the line are the parts of this work that still need a person. Bars below it are what AI can already do. Tap any column to see the actual work behind it.
high ground · holds stronglydeep water · reaches furthest
yours, by strengthAI reach, by depth
The honest read. One of the most robustly-human fields assessed this session for its protected segments -- four dimensions strong, anchored by the field's defining ambiguity-resolution act and genuine accountability structures. Protection is sharply uneven by content type: strongest for interpreting and literary work, essentially absent for the commodity bulk translation Section 2 shows as AI-transformed.
Language & Reasoning is unambiguously the field's core and defining capability -- neural MT and LLM-based translation are the mechanism the whole field's AI story runs through. Predictive ML and Agentic AI matter at the workflow/pipeline layer (deciding what gets machine-translated and orchestrating multi-language releases). Generative media (AI dubbing) is a real, fast-growing, genuinely dual-use application concentrated in media localisation specifically.
How AI is changing the way in
3 ways into this field, and AI is not doing the same thing to each of them.
Commodity/bulk written translationMuch harder to enter
Measured MT-usage-vs-employment-growth correlation; ~28,000 fewer new translator positions 2010-2023 (CEPR/VoxEU).
Localisation engineering/management and terminologySlightly harder to enter
These roles manage and enable the MT transition rather than compete with it; Mode 1 states the terminologist role has become more important, not less, as MT output has scaled.
Literary, media, and high-stakes interpreting (legal/medical/diplomatic/asylum)Largely unchanged
MT reported useless for real-voice prose; human interpreters essential for court/healthcare/EU/conference settings; protected by both quality gap and accountability structure.
That is everything we currently know about AI in Translation, Interpreting & Localisation. It shows where things are moving so you can choose which way in suits you.
People drawn to Translation, Interpreting & Localisation are often drawn to these. Most sit in a different part of the terrain.