The re-translated text is coherent and delivers the overall message. At times it tends to look more formal than the original, however, it is not always an advantage. For example, “smart government” is a widely known term, but Bing replaced it with “intelligent governance”, impeding its recognition. There are certain sentences that were seemingly influenced by Italian grammar, for example, the gendered names of inanimate objects have become animate in re-translation to English. Bing also failed to restore some pronouns that can be omitted in Italian because the affixes at the end of the verbs can replace them. The capitalization in re-translation is not the same as in the original, and some of it was lost in the initial translation into Italian. Speaking of capitalization, Bing has no notion of formatting rules and does not recognize the title, and the nouns in the said title are not capitalized.
The machine translator does not understand the context, distorting it with term replacement. In re-translation, some crucial words are substituted by synonyms (e.g., “prepare the way” became “prepare the ground”), which changes the meaning of the sentence. Another example can be seen in the case, in which “move forward” was replaced with “advance”. Aside from peculiarities, we can see that machine translators have, for the most part, learned the grammatical rules of the English language, so the text is easy to understand.
As it was mentioned before, the meanings of some phrases were distorted because a machine translator does not understand the importance of wording and precision in translation. For example, “qualitative leap” in the original turned into “quantum leap”, which is something completely unrelated to the topic of the original text. Legal terminology and language should also be translated by humans, for instance, the sentence in which the UAE Federal Penal Code is mentioned. The quote that follows the mention of this Code should also be translated or taken directly from the source by a human, as the paraphrasing of legal documents can have consequences for the translator.
In recent years, machine translators have advanced a lot, and their text output may often seem as it was translated by a human. However, translation algorithms tend to make mistakes, just as humans do. Most of the sentences seem as they were translated by a person, however, there are some mistakes at times that a human would not make. A machine cannot research the correct terminology or accurate proper names, and it does not recognize such words in the input text, thus making inexcusable errors. The given text is formal, which means there are many standard phrases, and Bing can translate those, but when it comes to literary works, it would be helpless. A machine translator cannot detect the tone of the text, figures of speech and thus will make any informal text just as robotic as a legal document.
On the one hand, an artificial intelligence device cannot be responsible for the wording it uses, so a person would need to go through a machine-translated legal text even when technology will be more advanced. On the other hand, artificial intelligence lacks the human soul to convey the mood of a written text, so people cannot trust machines with literary translation. Editing computer-aided translations are becoming more popular, but this process still does not eliminate the human factor from the process. In conclusion, machine translators are currently not ready for replacing human experts, and this job will still be relevant for a long period of time.