machine translators don't make mistakes

Machine translators work impeccably. They don’t make mistakes. They are perfect in their core competencies. However, what people often forget is that machine translators’ skills are purely mathematical. When an automatic translation is involved in some sort of problematic situation the machine can’t be held responsible. The user has made the decision to use and publish raw machine translation without post-editing. The machine is only a tool.

Machines do what people want them to do. They don’t understand the tasks. Machine translators basically just calculate and compare. They use existing databases and dictionaries in the search of a translation with the highest probabilities. Giving the available resources, machine translators always perform their best work. But like in any type of gambling, the odds aren’t always implemented and the highest probability doesn’t guarantee that the translation is grammatically correct.

Everyone makes mistakes but it’s easy to blame the machine for your bad choices. The user is the active player. Someone always makes the decision to use machine translation. If you skip proofreading and post-editing, it’s your choice. However, one should realize that when the translation needs to be perfect, raw machine translation won’t meet the requirements. Machine translation is rarely 100% flawless in terms of grammatical and linguistic correctness. Human languages don’t fit well on mathematical models.

However, this doesn’t mean that automatic translation would be useless. Machine translation is the best option in the right places. Sometimes it’s enough to just understand the message. The question whether we even should try to fit human languages into the mathematical environment or not is a source of a never ending discussion. The current level of automatic translations show that machines can produce quite good quality for understanding various texts in dozens of languages.

Problems occur when people misuse automatic translations. It’s a mistake to use raw machine translation when the goal is to do more than deliver a message. Imperfect translations may have an effect on how the audience perceives the message and the messenger. The higher the cost of a translation mistake is the greater is the need for perfect translation. Machine translation itself isn’t a mistake, using it in the wrong situations is.

Overall, it is clear that everyone should understand the limits of machine translation before using it. By knowing how machine translators work one can learn how to make get the best possible quality out of the machine.

Qualified machine translation Decrease the post-editing time and costs by using qualified machine translation. MT-Qualifier evaluates available machine translations and selects only the good enough translations for you automatically. Learn more.