MT_banner

MACHINE TRANSLATION:
WHEN TO USE AND WHEN NOT TO USE

The most common MT systems are Babelfish and Google Translate. They remain excellent to get the “gist” of what a text means, either from or into English (or any other language combination). But at best they still are only 70-80% accurate. The issue becomes which 20-30% of your text and your reputation for excellence you want to jeopardize with wrong MT word usage, improper grammar, and horribly comical results.

The Challenge for MT

The quality of machine translation has taken quantum leaps since its beginnings in 1954 at Georgetown University.  MT remains excellent to get the sense of what certain lines, sentences, paragraphs mean … all for personal use. But converting the MT version to proper speech requires tremendous effort.  Even with all the latest developments, MT still does not understand the tone, nuances and assumptions in any language.  It frankly is easier, faster and more accurate to use professional linguists to do a proper human translation right from the start.


>>See Examples of How MT Performs
vs. Human Translation
:
Using the popular Google Translate Tool


 
Our advice: Don’t ever trust machine translation when the text is critical. Your own reputation is at stake: Do you want your customers to wonder whether you will cut corners on them, looking for the cheapest method and not caring what end result you will deliver?

Our methods over time have saved some clients around $500,000. That includes substantial discounts within the normal human-professional process. But the results of MT can cost clients that much if applied incorrectly.  The following is a simple reference guide to help you understand when to use MT.