How Crowdsourcing Translation Could Make The Internet More Inclusive
One of the founding principles of the internet is democratization: the idea that anyone, anywhere, can access the information on the web as long as they have an internet connection. But this principle doesn’t quite work if some internet users don’t speak the language or languages that much of the world does. If you do not speak a dominant tongue, it can be as hard to communicate and understand what’s going on online as offline.
Indeed, while the internet has made great strides in translating the many languages of the world, certain less common languages or types of content remain woefully underserved. The four most spoken languages online are English, Chinese, Spanish, and Arabic, but there are millions of people who don’t speak any of those. This could be why we’re seeing a growing trend of crowdsourcing when it comes to solving this problem, where big internet companies like Google are asking its users to help improve translation accuracy—often offering them little in return.
Google just released an application called Crowdsource in its Google Play store, which is an example of this trend. Crowdsource asks a user to report what languages they are fluent in, and then asks them to choose from a selection of tasks including “image transcription, handwriting recognition, translation and translation validation, as well as map translation validation.” Unlike Amazon’s Mechanical Turk program, which doles out micro monetary rewards for tasks like this, the Crowdsource app offers no rewards—monetary or otherwise—for people that opt do this work. They are simply banking on language speakers goodwill.
TechCrunch reported on that strange quirk app by writing that “It seems odd that Google would launch an app like this, without offering something to users that would actually encourage contributions. After all, the company is already familiar with the model thanks to Google Opinion Awards, which offers Google Play credits in exchange for answering free surveys. Throwing out a few credits in exchange for helpful contributions across transcription services, translation and more, seems like it would have made sense.”
But of course, this assumption doesn’t fully take into account the eagerness that some native speakers of less represented languages have to see their language more prominent across the web. Indeed, a Google spokesman said that ““people may be inclined to use [Crowdsource] because, for many languages, tools like Translate, Image Recognition, etc. aren’t very good right now.” For people of a minority tongue, an opportunity to translate the web can come with a sense of pride.
In fact, Google’s step into this area isn’t the first time a huge tech company—one that would ostensibly have the budget to pay people to do this—has asked its users for help out of goodwill. Last year, YouTube (also owned by Google) rolled out its crowdsourced video subtitling tools, which allows “viewers add or edit captions voluntarily. They can correct spelling errors, add punctuation, and include speaker IDs. Since automatically generated captions are only 70% accurate, there are usually plenty of mistakes for fans to edit.”
YouTube is a platform that seems especially suited for this, as mega fans of some channels would be delighted to help spread the videos to more speakers of more languages. Fans can add subtitles for any language, meaning that there are no longer any linguistic limits on the global audience that a YouTuber could potentially reach.
Indeed, language is a huge barrier on a global scale, but the power of the internet could help change that. It just remains to be seen whether minority language speakers will be willing to pick up the mantle of translation now that these huge tech companies are asking for their help.