Computer-aided translation: What is it?

Machine translation has been around for some time and it's not perfect. CAT tools are designed to facilitate and speed up the process by providing context and suggestions from previously translated sentences. CAT tools are often used by professionals to speed up the translation process, as segments that have already been translated in previous documents are stored in the translation memory (translation database) and can therefore be reused. It is therefore not a case of machine translation, but rather an aid through recycled texts translated by humans.

Machine translation vs. CAT tools

This type of software is widely used by professional translation companies. Every professional translator must be familiar with CAT tools in order to use them properly. These tools are used to insert customised translation memories, glossaries and reference material for clients. This makes the translator's work easier, whether into English or Spanish. With the correct use of CAT tools, segments that have already been translated can be reused, thereby reducing translation costs. This is particularly helpful if, for example, you have translated instructions for use into Portuguese, Italian and French in the past, but have now updated them.

With CAT tools and the integrated translation memory, all segments that have not been changed are automatically recognised and blocked for the translator so that only the new or changed sentences are sent for translation. This maintains the terminology and consistency of the instructions for use and reduces translation costs through reuse. CAT tools are often confused with machine translation, but in this respect we would like to distinguish the fundamental difference between these two types of translation.

Machine translation involves no human intervention, i.e. the translation is produced exclusively by a machine. While it is true that technology has improved enormously over the last few decades, it is also clear that despite the neural network structure of texts, translation systems such as Google have their drawbacks. There are numerous cases in which the translation does not work perfectly, especially when it comes to highly specialised technical texts with their corresponding terminology or creative marketing texts. These are the weaknesses of machine translation. For example, if a company wants to translate material for medical products that affect human lives, machine translation is not recommended.

Full machine translation should generally not be used by translation companies, as each client could simply do this themselves on the Internet. A professional translation company will always use CAT tools in order to respect a client's individual wishes and terminology and deliver high-quality human results. One option that some translation agencies offer is post-editing. This includes both machine translation and human processing. This option can be interesting for some companies if their texts are not very complex and they need to keep translation costs down. It consists of pre-translating the text by a machine and correcting all possible errors by a human translator in a subsequent step.

The future of AI in translation

The future of AI in translation could be promising. The quality of translations is improving with the help of AI and will continue to do so in the future. The technology is already being used by large companies such as Google Translate and Microsoft Translator . Many software developers are also already working on API solutions for integration into CAT tools. This can make work easier at both levels and improve quality enormously.

Nevertheless, the extent to which a human being can completely rely on a machine for certain types of texts is questionable and a human translator should always be involved, particularly in the case of specialised translations, such as operating instructions for surgery. With this in mind, you must be careful and not blindly trust what can be done by a machine. In the meantime, the possibility of using such AI tools as support is extremely positive. The error rate has decreased and the quality of the translations has improved significantly.