Imagine you’re a chef preparing a gourmet meal. You wouldn’t start with a pre-cooked dish and then try to transform it into a culinary masterpiece, would you? All you can really do with someone else’s meal is put a little ketchup on it and hope for the best!

The same principle applies to translation.

Starting with a machine translation and then post-editing it (MTPE, also known as PEMT) is efficient only if a “good enough” output is acceptable. MTPE inevitably leads to mediocrity, not excellence.

Here’s the fundamental problem with machine translation post-editing and how you, as a translator, can overcome it:

1. Understand path dependency

Translation is inherently path dependent. Starting with a machine-generated translation locks you into a “rut” that doesn’t align with your unique voice and style. It doesn’t matter whether the machine was properly trained or not, when you post edit, you’re not working in your own habitat.

In this situation, transforming a machine translation into something you can be really proud of takes at least as long as doing it from scratch.

This is why so many translators say that working with MT and AI doesn’t help them in their work…

However, this is not the end of the story!

2. Engage early

Don’t let the machine dictate your starting point.

If your client is forcing you to work from their machine translation, then it’s their loss; you can make do with the Generative AI MTPE (GAIM) workflow.

But if you have any say in the matter, you can deliver better work faster by switching to the Generative AI Iterative Translation (GAIT) workflow.

3. Leverage AI as a partner

Engaging early does not mean batch-translating your job with AI.

Yes, AI can help you translate better. But you have to work holistically, commanding the AI to align with your voice.

Through an iterative process with the GAIT workflow, you have a choice.

  1. You can work twice as fast and deliver 80% quality with a “turn-key MTPE” service that’s still better than legacy MTPE.
  2. You can deliver even better work than you did before by following the best-practice GAIT workflow.

Let’s rethink this; legacy MTPE is not the future of quality translation in our industry.