If the AI bubble bursts, do you get your old translation workload and rates back? – 23 September 2026 (AI Mastery for Translators Insights)

Hello translators with bills beyond the bubble!

NAVIGATING THE PARADIGM SHIFT

If the AI bubble bursts, do you get your old translation workload and rates back?

Wouldn’t we all love to say yes?

But just because an AI executive warns that their technology could threaten humanity doesn’t make it so.

I don’t dismiss the risks. But warnings that also flatter company valuations deserve scrutiny.

And for us as translators, it’s probably irrelevant: the hype can collapse without wiping out the underlying trends.

Take those bland AI articles nobody asked for. I’m increasingly reluctant to read AI slop myself.

But generating content is different from translation.

A good source text already supplies the ideas, information, and structure. AI can help translate it without turning it into empty prose (though, of course, professional translators can make it better).

And pressure on rates? That predates ChatGPT. MTPE and clients’ appetite for cheaper translation won’t disappear because investors lose money.

Neither will useful software. Even if progress stalls, today’s capabilities are still impressive and will continue to be integrated into workflows.

That leaves us with a better question than “When will this bubble burst?”

“What can I improve so that I stay competitive no matter what happens?”

Keep working hard, but give that effort a direction. Improve your craft. Test technology on suitable assignments. Notice which client problems still require your expertise. Try different methods. Compare the time spent and the quality delivered. Keep what works, and discard what doesn’t.

As a working Korean-to-English translator, I’d rather build on results I can verify than wait for relief nobody can promise.

My optimism is about translators adapting, not the market rewinding.

Reject exaggerated claims. Just don’t make their failure your business strategy.

Keep improving your work, even when the headlines change.

EFFECTIVE PROMPTING

Getting reference material out of your CAT tool and into your AI prompt

Your CAT tool contains useful guidance for your AI workflow: approved terminology, translated segments, and examples of your client’s preferred style.

But while working inside a CAT tool, how can you easily get that material into your translation prompt so you can use it?

Here are three ways.

1. Iterate your prompt as you go

For terminology, get the entries you need directly from your termbase… or from the document you’re working on. You can do this manually, or by exporting to an Excel file. Then select the source and target columns and copy them into the prompt. And keep adding as you work.

Keep each pair on its own line, with a tab or clear separator between the languages.

2. Capture selected bilingual rows with OCR

Want just a handful of approved segments, without exporting the whole document? Or working in an interface where you can’t export the document?

One workaround I’ve used is ABBYY Screenshot Reader. Capture the relevant source and target rows showing in your CAT-tool editor, extract the text, and paste it into your prompt.

Remember: Each source segment needs to stay clearly paired with its translation.

3. Pull from an exported bilingual Word or Excel file

If you’ll be collecting reference materials repeatedly on a longer project, export a bilingual file (source on the left and target on the right) and keep it open alongside your project.

You can then copy relevant sections as needed instead of repeatedly capturing the translation grid with OCR. The initial export may save you several rounds of fiddling later.

Whichever route you choose, extract only what’s relevant. A reasonable number of accurate, clearly paired examples can beat a screenful of scrambled references.

COTRANSLATORAI TIP

Is your AI question on this list?

What’s costing you time with AI? An ignored glossary? A timeout? A prompt that grows longer with every client request?

You’re not the only translator asking.

At our AI Drop-In Clinic for Translators last week, we discussed questions like these:

How do I update CotranslatorAI for newer models and top up API credit?

Which models suit different translation tasks, clients and documents?

How can I make my glossary stick without repeating terminology instructions?

Which parts of my growing client prompts can I safely remove?

Why does a task time out in CotranslatorAI and what can I do about it?

How do I save reusable prompts and entire workflows?

How can I strip timecodes and line breaks from an .srt file without changing the wording?

Where should I keep client guides and regulatory wording?

After the beginner masterclass, what should I study next?

I appreciated hearing how differently translators work. Some want one strong AI suggestion to improve. Others prefer several alternatives so they can write their own version. Neither approach requires surrendering your judgment and neither is right or wrong; you have to find what works for you.

Start writing down your questions about CotranslatorAI or AI generally.

When an output disappoints, save the prompt, a short sample, and the result. Add a sentence explaining what you wanted instead. And then email it to me before next month’s AI Drop-In Clinic for Translators.

It is free for CotranslatorAI Advanced users. Dates and times are coming soon.

Not a CotranslatorAI Advanced user? You can get a three-month subscription here: https://cf.cotranslatorai.com/main

Keep your judgment sharp and your questions flowing.

Steven
For Stanislav, Dave, and Anne