Can we rely on a machine translation tool in 2020?

These days, machine translation tools are everywhere, from instantaneous translators or voiceover translators to context dictionaries. However, since its existence, its reliability has been the subject of controversy.

Where are they today, in 2020? Can we trust an automatic translation tool in 2020? What are your limits? How have they progressed?

1. The advances are more than admirable

In its early days, machine translation made many mistakes that made it unreliable, even for translating a single word (there are words that are polysemic, that is, they have several meanings depending on the context).

However, in recent years, some of the latest technological advances have been applied to machine translation. This has led to a dramatic improvement in machine translation, especially through the use of machine learning and artificial intelligence.

These two technologies literally allow a system or a machine to learn a language, its grammar, its syntax, its idiomatic expressions, etc. Thanks to techniques such as mapping, syntactic or statistical analysis, machine translation tools have changed radically in recent decades. Just think of Deepl, an automatic translator that is overshadowing Google's own translator, for example.

To all this, we must add a constant enrichment of the databases and laboratories of computer giants such as Facebook, Google, Microsoft or Baidu, which also contribute to this enrichment.

2. There are still insurmountable limits

The advances made in recent years allow machine translation tools to perform translations in various languages. Their translations are relatively reliable and surprisingly fluent, but they are far from perfect.

In fact, machine translations cannot always replace a professional, flesh and blood translator . Although grammar, syntax and even context translation have improved enormously, machine translation still makes numerous errors and its tools can still be greatly improved.

To obtain a really perfect translation, it is not enough to have technical skills. It is also essential to have a knowledge of the culture in order to be able to take into account the localization or the global meaning of the text.
Nowadays, automatic translation works quite well with the translation of certain technical texts, while when it comes to more literary or poetic texts, it does not achieve the desired results.

3. Establish your needs

So, today, in 2020, unlike a few years ago, we can indeed rely on an automatic translation tool, but always depending on our needs.

Don't forget that these tools are mainly aimed at individuals, for example, in the case of people who go on holiday abroad and need to translate a text for their own understanding.

So it all depends on your needs and expectations. If you want to translate a text, even a specialized one, in order to understand it, but without wanting to spread it around, it may be enough to use an automatic translator. This will allow you to understand the main points of the text's content immediately.

On the other hand, any specialised translation aimed at third parties (collaborators, clients, hierarchy, etc.) must be of impeccable quality. It will have to respect the meaning and structure of the source text by adapting it to the target audience. It is recommended to use a specialized translator for this purpose.

In general, machine translators are much more reliable today than they used to be. They are very useful especially for personal use. However, they cannot always replace the skills and talent of a specialist translator. Before using an automatic translation tool, it is essential to establish the needs and purpose of the translation.