Translators and AI: Post Editing with Judgement
Use machine translation and AI well, post edit efficiently and know when to translate from scratch
A taste of a lesson
The AI translation of a contract reads perfectly in English. Do I still need to compare it line by line with the German?
Yes, especially for a contract. Machine output can read perfectly while omitting a clause, flipping a negation or rendering a legal term with a near synonym that changes meaning. Reading only the target hides those errors because the text flows. Compare source and target segment by segment, check defined terms against the glossary, and verify every number and date. Want a quick test? What would you check first in a German clause starting 'Sofern nicht schriftlich anders vereinbart'?
Written by the teacher as an example. In your lesson the tutor answers your own questions, and like any AI it can be wrong.
What you will be able to do
- Compare machine translation engines and language models for translation tasks
- Post edit to an agreed level using glossaries, memory and QA checks
- Spot omissions, terminology drift, register and gender errors
- Decide when AI is inappropriate and handle confidentiality and pricing honestly
Lesson plan
- 1 How machine translation works now Understand NMT engines and language models, and why quality varies by pair and domain. Start
- 2 Error patterns Spot the typical errors machine translation introduces. Start
- 3 Post editing levels and workflow Post edit efficiently to the level agreed with the client. Start
- 4 Prompting language models for translation Use instructions, glossaries and examples to steer language model translation. Start
- 5 When not to use AI Identify texts and contexts where AI translation is inappropriate or not permitted. Start
- 6 Value, pricing and clients Talk honestly with clients about AI, effort and quality. Start
Try asking
About this tutor
For professional translators, students of translation and bilingual staff who translate at work. We cover how neural machine translation and language models differ, light versus full post editing, building glossaries and style guides into your workflow, catching typical errors (terminology, omissions, false fluency, gender and register), and deciding when AI help is inappropriate: literary voice, legal or medical documents needing certified translation, confidential client material. We also discuss rates and client conversations about AI honestly, confidentiality terms, and protecting your professional value.
Reviews
4.7
3 ratingsSample
- Beatriz S.Sample
As a student I appreciated the honest look at the market. The language model prompting lesson was a real eye opener.
- Ines W.Sample
The false fluency examples were exactly what I needed to convince a client that post editing still takes real time. Very practical.
- Kazuki M.Sample
Good on workflow and pricing conversations. Japanese to English examples were limited, but the principles transferred.
About the teacher
Translator, editor and digitisation hand teaching AI for words, archives, buildings and travel
9 tutors 307 lessons taught Sample
I began as a technical translator for engineering and architecture firms, worked on a regional news desk as a sub editor, and later joined digitisation projects with libraries and archives. Translating for tourism businesses brought me into hospitality too. What links all of it is careful work with text and documents where a small error travels a long way. I...
See Zoltan's profile and tutorsMore like this
Other tutors on the same or nearby topics.