Today I want to talk about Machine Translation. This can be a controversial topic for translators and I hadn’t addressed it so far.
But before talking about what it means for translators, we should talk about what exactly is Machine Translation.
Machine Translation or MT has made appearances in the Artificial Intelligence (AI) world with apps like Google Translate. In a larger format, the MT industry is more complex. It involves translating text and content from one native language to another.
It is a sub-category of computational linguistics that borrow from computer science, AI, information theory, and statistics. In the past few decades, there has been incredible progress in MT quality, and its evolution has created a huge industry.
While MT alone is not perfect, the use of human translators along with machine learning has many practical applications. Today, the MT market is thriving with the combination of human and machines translations, offering faster service and lower costs.
Of course, MT has many flaws and remains inferior to human translators. Gender bias, distorted and mistranslated words and phrases, and bizarre word phrasing, are among some of the biggest flaws.
But perhaps the worst flaw is a machine’s inability to process human thoughts and emotions. Using translators as post-editors rectifies this as they can find errors that the machine missed.
Have you ever post-edited a machine-translated document? You definitely see these flaws right away. With just one word or very short phrases, it might be okay, but with longer sentences and paragraphs, you can see right away that they don’t make sense.
Of course, the problem is that MT translates literally. A machine cannot notice the context of the original document. It will translate word per word, but without sense. Here is where human translators come in. They post-edit the translation of the MT to correct all the errors.
I’ve done a few Machine Translation Post-Editing (MTPE) projects. It is interesting to correct the errors made by the machine. This also means that the rates for MTPE are lower than for translating since the translation is already done and all you need to do is review the document and make any necessary corrections. Almost like a proofreading project.
But I know that many translators reject these kinds of projects because they don’t support MT. I understand that but MT is not going anywhere, on the contrary, it is growing and getting more sophisticated.
I don’t think machines will ever take over human translators. No matter how sophisticated they get to be, they will never have thoughts, understatement, feelings like humans do.
Where do you stand? Do you take on MTPE projects or do you reject them, and why? Let me know in the comments, I’d really like to know 🙂
Thank you again to Day Translations for sponsoring this blog post. Check out more about them and their services here: https://daytranslations.com/
Don’t forget to thank the women in your life for their support!
Until the next time, take care and stay safe!