Recently, Facebook shut down one of the artificial intelligence (AI) programs they were working on because it started to develop and use its own language to communicate between two AI agents. As a member of the translation industry, and a principal with Para-Plus Translations in New Jersey, I thought it was humorous that the AI found it easier to create a new language rather than try to follow the rules, subtleties and nuances of our existing one. That very action by the AI illustrates the complexity of language and how difficult it can be to effectively communicate your thoughts, even for a machine.

The AI system originally started with English instructions. If that system was frustrated enough with just one language, image what it would be like if another language was introduced with its own set of rules, cultural norms and intricacies. This confirms that written translation is not a simple equation where A always equals B. There is a skill and an art to translation which cannot easily be defined in a derived algorithm.

With advances in modern technology, we often hear about machine translation such as Google Translate and how it can revolutionize the industry. To some degree, it has. The sheer volume of information readily available has grown exponentially over the past several decades. This information can be as robust as entire library catalogs or as simple as a YouTube video or a Twitter feed. People have access to information and can communicate with others half way around the word with literally a click of a button. That access to information, however, still has to overcome the hurdle of language accessibility.

With the volume of information being distributed on the web on a daily basis, there are not enough human translators in the world who can keep up with the need. Machine translation and translation memory software is now relied upon worldwide. We rely on it to help with something as complex as sorting through a mountain of documents needing translation for an up and coming lawsuit or something as simple as a funny meme that was posted in a different language.

Machine translation may be able to, at some point, take over human work but based upon the AI example, the real question is whether the machine translation will even want to. At least for the here and now, there can be cohesive synergy between machine and human translation services.

At Para-Plus Translations, we understand the complexities of language. Our translators understand language and cultural nuances and appreciate the importance of using translation tools to help expedite translation output, decrease turnaround times, but more importantly, reduce translation costs to clients. Our linguists have the ability to utilize such tools but also inject their own knowledge, skills and experience to effectively communicate our clients’ message as it is intended. And we'll be sure not to make up our own language just because we feel like it.