By Sarah-Claire Jordan
Social networking sites and artificial intelligence might seem like a strange combination in many ways, but if you take a moment to think about it, it makes perfect sense actually. Artificial intelligence is what non-computer engineers call anything a machine, in this case computer, does that is similar to what a human can do, more or less. Social media and social networking sites are designed with the specific purpose of bringing people closer together via the Internet. Translation that utilizes some form of AI fits right in there as a great tool for making communication across language boundaries on social media platforms simpler.
The best example of a social media platform that is taking serious steps to bridge the gaps that language barriers create is Facebook. Since it is one of the oldest and most widely-used social media sites, it makes sense that the people in charge would have a vested interest in making sure that anyone who uses it has as similar experience as possible to everyone else. Perhaps due to its success, Facebook has been able to set aside time and money to really work on translation issues.
Everyone who is up on their tech news is aware of the fact that Facebook has decided to use a “neural” machine translation engine, or MT engine. It is called neural because it uses some AI technology that ends up creating translated content that is much more natural sounding than regular MT engines. Before, when we would see anything posted in something other than English, we would have the option to click and see a translated version of the content. More often than not, this was translated using a regular MT engine, and the results were not always, shall we say, easy on the ears.
The idea behind the updated AI translation feature is to allow users to have the option for automatically translating anything they post into a variety of languages. This doesn’t sound much different from previous translation features, but here’s where it gets good: if you post something in English but your family in Argentina wants to know what you are up to as well, they will be able to read it in a much more natural Spanish as long as they have their preferred language set to Spanish. This means a much more personalized experience for Facebook users all over the world, which in turn will help us feel more connected and be able to overcome the invisible barriers that language puts between us.
Not only is the neural MT engine used by Facebook much better than a regular one in terms of producing more natural translations, but it is also very accessible in that native speakers will be able to go in and correct any mistakes they see. This is exactly what Facebook wants its users to do, actually, so as to help the engine “learn” and be able to apply corrections to future translations. Sure, this kind of crowdsourcing isn’t a perfect way to handle translation correction, but it is far better than any other alternative out there. Besides, Facebook is a company built around getting people involved. Why would their newest translation feature be any different?