SHIFT — THE AI JOB-CHANGE WIRE —
Vol. 1 · No. 19 MON · 4 MAY 2026
DISPATCH No. 34 · TRANSLATION & LANGUAGE FROM SHIFT WIRE · DESK: TRANSLATION

AI IMPACT ROLE DOSSIER · TRANSLATOR

Heavy pressure ahead for translators.

Automation is stable. Augmentation is medium-high and edging down. Human resilience is edging down. By 2030, the role will look noticeably different — but not gone.

Translation is the role where the wide-and-shallow AI revolution went furthest, fastest. By 2026 general-purpose translation is essentially solved for most language pairs.

Automation is HIGH and stays high. Document translation, subtitling, captioning, marketing localisation, UI strings, customer-support translation — all production-grade machine output. Many language-services agencies cut their full-time translator headcount by half between 2023 and 2026.

Augmentation does its work in the QA and post-edit layer. Style adaptation, register tuning, machine-output review. The translator of 2027 is closer to a QA editor than a writer.

Resilience is low and falling. What survives: literary, poetic, and high-craft translation. Legal and medical certified translation where liability matters. Live interpretation under courtroom or summit pressure where presence and judgement are the product. Everything else is hard work for a shrinking market. The pivot paths are narrow and specialised — but they exist.

— On the instruments —

2028 outlook · 5-band scale · interpretive
moderate ·
3 of 5

— overall reading · 2028 outlook —

HEAVY

3 of 5 · high confidence · last verified 2026-05-04

Automation stable · augmentation med-high and edging down · resilience edging down.

VIEW EDITION

automation pressure

how much AI is taking over

High
5 of 5

stable High High

augmentation pressure

how much AI is changing the workflow

Med-High
4 of 5

edging down Med-High Medium

human resilience

how much stays stubbornly human

Low
1 of 5

edging down Med-Low Low

Where the pressure lands

Skills automating

  • General-purpose document translation high ↗ Stanford HAI
  • Subtitling and captioning high
  • Routine localisation for marketing and UI high

Skills augmenting

  • Style adaptation and register tuning medium
  • QA workflows on machine output medium

Skills holding

  • Literary, poetic, and high-craft translation high
  • Legal and medical certified translation high
  • Live interpretation under courtroom or summit pressure high
— Adjacent dispatches — BRIDGES BY SKILL OVERLAP
— Sources cited in this dispatch —
  1. Artificial Intelligence Index Report 2025 — Stanford HAI, 2025