AI for Translators
Turn AI from a threat into your post-editing engine. Build a first-draft prompt, a project glossary, a quality checklist, and a rate line — 8 hands-on lessons.
For most translators, AI didn’t arrive as a tool. It arrived as a pay cut. In the UK Society of Authors’ 2024 survey, 36% of translators had already lost work to generative AI, 43% had watched their income fall, and 77% expected it to keep falling. Standard rates that used to sit around $0.15–0.30 per word have slid toward machine-translation post-editing bands closer to $0.05–0.15. By 2024, nearly half of all language-service projects ran through a post-editing workflow.
Here’s the part the doom headlines skip. The machine is fast, but it is not reliable — accuracy swings wildly by language pair and subject, and it confidently invents legal citations, mangles drug names, and quietly drops the soul out of a sentence. So clients are now shipping their own bad AI translations and paying humans to fix them. The translators who are surviving aren’t out-typing the machine. They’ve become the post-editor who drives the AI and stands behind the result.
This workshop teaches that exact workflow, start to finish, with prompts you copy and paste. You’ll write a briefed first-draft prompt, build a project glossary and lock it across files, run a real light-vs-full post-editing pass, set a confidentiality boundary that protects your clients’ NDAs, and write a rate line that prices the work honestly — including which offers to refuse. Every lesson ends with something you keep.
It pairs with our two companion reads: ChatGPT for Translators (the defensive-adoption playbook) and Build a Translation Glossary With ChatGPT (the glossary deep-dive behind Lesson 4). This course is where you stop reading about the shift and start building the kit that gets you through it.
What You'll Learn
- Explain how AI is reshaping translation rates and demand, and reframe your role as the post-editor
- Use a briefed first-draft prompt that controls domain, audience, tone, and language variant
- Apply the 5 risk zones to decide what AI output you can never trust unedited
- Build a reusable project glossary and run a consistency sweep across multiple files
- Evaluate a machine draft with a light-vs-full post-editing checklist
- Design a client-ready post-editing rate line and a confidentiality boundary
After This Course, You Can
What You'll Build
Course Syllabus
Prerequisites
- You translate, or want to, in at least one language pair
- A free ChatGPT, Claude, or Gemini account (no paid plan required to start)
- Willingness to treat AI as a junior assistant you supervise, not a replacement
Who Is This For?
- Working translators watching AI eat into their rates who want a way forward
- New and student translators who want to start with the modern workflow, not the old one
- Bilingual professionals asked to handle translation at work
- Freelancers being offered machine-translation post-editing (MTPE) jobs they don't know how to price
Frequently Asked Questions
Do I need to be a professional translator already?
No. If you translate in any language pair — professionally, as a side gig, or for an employer — this course fits. It assumes you know your languages but have never built an AI workflow around them. Beginners to AI are the target audience.
Which AI tool does this cover — ChatGPT, Claude, or Gemini?
The prompts and workflow work in all three. We use ChatGPT in most examples because it's the most common starting point, and we flag where Claude's longer context window helps on big documents. Everything is copy-paste and tool-agnostic.
Will this teach me to let AI do the translation for me?
The opposite. You stay the translator who stands behind the result. AI drafts; you post-edit with judgment, lock terminology, catch the errors it introduces, and price the work honestly. The course is about supervising the machine, not trusting it.
Is it safe to put my clients' documents into ChatGPT?
Not always — and Lesson 7 is entirely about this. Free consumer ChatGPT can use what you paste to improve its models unless you change settings; that can breach a client NDA. You'll learn exactly what never goes in, how to anonymize, and which tools have a no-training default.
Will I get a certificate?
Yes. Complete all 8 lessons, pass the quizzes, and submit your capstone post-editing kit to earn a verifiable certificate you can show clients.
