We haven't heard much from the software giant in Redmond lately, but they certainly don't sit still at Microsoft Research. An old science fiction dream, real life translation of spoken text, is now becoming reality. Just a few seconds is enough for the software to convert spoken English into spoken Mandarin Chinese, a totally unrelated language. The Chinese audience was deeply impressed ...
Language classes may have had their day with this technical breakthrough on Microsoft's behalf. Not only can this new translation software from Microsoft translate spoken English into spoken Chinese with just a few seconds delay, the software also uses the intonation and voice of the speaker to do this. The technology was demonstrated by research director Rick Rashid in the Chinese city of Tjianjin on October 25, 2012. The translation from English into Mandarin starts in the video at 7.30am.
The implications, also for the Netherlands, are enormous. Telephone sales and other services that require spoken knowledge of another language can now be truly outsourced to low-wage countries. Or, more likely, entrepreneurs in the low-wage countries will now start selling en masse by telephone in the European countries that were previously unreachable due to the language barrier. Services such as secretarial support and advice can also be rapidly taken over by well-educated Indians, Chinese and Southeast Asians. Conversely, this also means that almost the entire world can be served from the Netherlands, including areas where English is not spoken. Politicians who say different things for domestic use than for other countries are also immediately exposed. This is the beginning of opening the intellectual boundaries between the various language areas. The linguistic tower of Babel comes to an end. Language areas that are currently cut off from the rest of the world, such as the Arabic language area, will within a few years be inundated with information from the rest of the world. Because Rashid expects that language barriers will be a thing of the past within a few years. We will then be able to flirt literary with frustrated Saudi housewives, ask an Inuit from Nuuk for advice on the best way to smoke salmon, understand Aztec poetry in classical Nahuatl and sing Gregorian chant in classical Latin.
The possibilities only really become unprecedented when these skills are linked to a semantic system. This makes it possible to build artificial intelligence systems that are able to answer questions with a high abstraction level. People can then carry a digital personal assistant in their pocket, who selects and arranges things for them. In short: living a life as a noble lord or lady a few centuries ago, but without the constant risk of being murdered or dying from a simple infectious disease.
Microsoft Research (2012)