Machine Translation and Where it Fits in the Translation World

Machine translation is becoming more powerful with each passing year. It is worth considering whether it is now capable of completely replacing human translation. The most obvious example is Google Translate, a free product that has undergone steady updates as its developers improve the software. However, there is still room for human translators in a few specific areas.

Translation needs vary from person to person and business to business. This is what provides the different opportunities for both machine and human support. The most important differences involve speed of translation, cost, the format of the content, and the complexity of the content.

Speed

Humans will never be able to translate as quickly as software can. Thought a machine, even with modern AI machine translation technology, can’t replace the art of communication that native human translators can produce. There is always a quality difference between machine and human localization. Though in some instances, machine translation may be a better solution. A human translator has to be able to read through the content, decide on a translation, and then type up that translation. Computers are capable of doing all of those tasks much faster than people. Moreover, recent advances have leveled the gap in quality and capacity between humans and machines. It used to be the case that machines could only translate by words and phrases, but in 2016 Google Translate became able to handle multiple sentences at a time, and other services also have this ability. Ongoing work in teaching software how to understand context, colloquial language, and tone have made software more accurate. The number of supported languages and size of vocabulary has also increased. Even with these advancements though, humans will be the best way to accurately represent brands on a global scale.

Cost

The question of cost is similar to that of speed. Software tends to be inexpensive or even free to use, while human translators have higher costs. This is in large part due to the difference in process. For software, there is little difference between one sentence or ten pages. Both can be scanned and translated quickly. For a human, the longer the content, the more complex and time-consuming the translation job becomes. This will lead into a subsequent point about complexity of content, but in terms of cost alone, software is a cheaper solution. There is a variety of cost levels among humans depending on quality and experience, but none of them will do the work for free.

Format

Format is one of the areas where humans retain an advantage. For written content, it is simple for software to process and translate any amount of text. However, translating audio is much more challenging for a machine. A good human translator can actually translate audio faster than written content. This applies in several contexts. The most obvious one is conversation. Whether it’s talking on the phone to an international supplier or making contact with a new contact who doesn’t share a language, humans can facilitate clear, high-quality communication where software falls short. This is important in business because high-level communication between executives and managers involves a lot of speaking- those conversations can’t all happen over email. Aside from direct conversation, humans are also adept at handling translated transcriptions of audio or video recordings. A task like copying down the dialogue of a training video and translating it into another language for subtitling is still more human-friendly than software-friendly. It’s something that software is gradually improving on, but it’s just easier for people to accurately hear and record spoken language than it is for a computer. So any translation that needs to be done “live” or requires transcription is much better suited to human translators.