Translator / Interpreter
Moves meaning between languages — and decides what to do when the meaning does not move cleanly.
This is not a probability of losing your job. It combines how much of the role's task load is exposed to automation with how far adoption has actually gone — useful for comparing occupations on one consistent basis, and for nothing else.
Written for general commercial translation. Certified legal, medical and simultaneous conference interpreting sit on materially different curves, mostly because of liability and real-time constraints.
What is actually changing#
The unit of analysis is the task, not the job title. A role is not replaced — its task mix shifts.
First-draft translation of routine text
Automating✓ Evidence-backedManuals, product copy, correspondence, internal documents.
Machine translation quality on high-resource language pairs crossed the threshold where a first draft is faster to edit than to write. That single fact restructured the pricing of the entire industry.
Quality is highly uneven across language pairs and domains. Low-resource pairs and specialised terminology remain much weaker than the headline impression suggests.
Post-editing machine output
New task✓ Evidence-backedFixing what the machine got wrong, and knowing where to look for it.
This is now the dominant paid activity in commercial translation. It is a genuinely different skill from translating: the hard part is spotting fluent text that is confidently wrong.
It is real work, but it is usually priced per word at a fraction of translation rates. More of the industry's hours, less of its income.
Translation where being wrong is expensive
Being augmented≈ Platform inferenceContracts, regulatory filings, clinical material, safety documentation.
Machines draft, but a named person certifies. As with accounting sign-off, the binding constraint is who is liable rather than who is capable.
Deciding what to change, not just how to say it
Still human-led≈ Platform inferenceTranscreation, market adaptation, and telling a client their message will not land.
Requires knowing the target audience as a lived thing and having the standing to push back on the client's brief. The output is a judgement about the market, not a rendering of the source.
Interpreting in the room
Being augmented≈ Platform inferenceReal-time, in a negotiation or clinical setting, where you also read what is not being said.
Real-time speech translation works and is improving, but the interpreter is also managing turn-taking, register and trust between two parties. In high-stakes rooms, participants still want a person accountable for the rendering.
Which technologies matter here#
Four separate signals. They are deliberately not added together — a job exposed to two technologies is not twice as exposed.
How it got here#
The index is not a static number. This is where it would have sat at each capability checkpoint since ChatGPT — reconstructed, and labelled as such.
● 1 verified event for this occupation, plotted at the date it happened — the parts of the line near a marker are anchored to something checkable.
Starts high because machine translation was already good before 2022 — translation memory and neural MT had been in the workflow for a decade. What changed after 2023 is that the draft became good enough that the paid step moved from translating to certifying. The flattening in late 2025 is the pushback: buyers who cut human review discovered the failure mode is invisible until it is expensive.
A flat line is not a forecast of safety. It says which tasks automation has reached so far — the occupations that moved least here are the ones where the constraint is physical or regulatory, and both of those can change.
Recent changes#
Entertainment subtitling, one streaming platform, one episode (Necronomico and the Cosmic Horror Show, ep. 1, 1 July 2025). The company's president had said in 2024 it was testing generative AI for subtitling; the failure was unreviewed vendor output.
Failure, rollback, regulation or cost is suppressing adoption. Can lower an assessment or widen its uncertainty.
Engadget ↗What this means for you#
This is the hardest first-batch occupation to enter right now. The volume work that used to let a new translator build speed and a portfolio is largely gone, and post-editing rates are too low to learn on. If you are entering, aim at a specialism with liability or regulation attached, or at localisation roles that sit inside a product team rather than at an agency.
Your per-word income is under structural pressure and that is unlikely to reverse. The defensible positions are subject-matter authority, certification, and client relationships where you are advising on the market rather than delivering words. If your current mix is mostly volume work at agency rates, treat that as a signal to move rather than to work faster.
Your options#
Four directions, each with its real constraints and one thing you can test this week. Continuing as you are is a legitimate choice — it just has to be a chosen one.
Specialise into a liability-bearing niche
Certified legal, medical and regulatory translation still needs a named, qualified person, and prices accordingly.
Certification takes real time and money, and requirements are jurisdiction-specific. Verify the exact local requirement before starting.
Find the actual certification requirement in your market and write down the cost, the time and who accepts it. Most people never do this and assume it is harder than it is — or easier.
Localisation inside a product team
In-house localisation is about making a product work in a market — terminology systems, cultural review, working with engineers — and is paid as a product role rather than per word.
Requires comfort with tooling and version control, and usually a specific target market where you are genuinely native.
Take a product you use, find three localisation errors in your language, and write up why each one is wrong. That document opens doors.
User research or international operations
The underlying skill — understanding what people in another market actually mean — is the scarce part, and it is worth more when applied to research than to text.
The title change makes your CV harder to read. Expect to need one project that demonstrates the new skill before anyone will hire you for it.
Interview three users in your language about a product, and write a one-page findings note. That is the artefact, and it takes a week.
Reposition from translating to advising
Clients still need someone to tell them whether the message works in the target market. That conversation is not priced per word.
Only works with clients who have a market decision to make. Agency intermediation usually blocks it.
On your next job, send one unsolicited paragraph on what will not land in the target market and why. See whether the client engages.
Common questions#
Studying a language is not the same as training to be a translator, and the two have diverged sharply. Deep command of a language plus a second domain — law, medicine, product, research — is more valuable than ever. Training aimed purely at producing translations is aimed at the part of the work that has moved fastest. If you are already in the programme, the useful move is to add the second domain now rather than to switch out.
Studying towards this?
These majors lead here. Their pages break down which of their competencies transfer and what graduates typically lack.
Method and sources#
- Assessment date
- 2026-09-09
- Basis of the task judgements
- 2 evidence-backed · 3 platform inference · 0 not enough evidence
- Verified events
- 1