Imagine: you are in a conference room with colleagues from China, Japan, Germany and Egypt. You each speak in your native tongue and the other participants hear it in their native tongue. Humans have sought the ability to understand each other regardless of their respective languages throughout history. That time is here.
In the biblical tale man’s hubris to build a tower to reach heaven is punished by the tower’s destruction and the people now unable to understand each other because they are speaking in different tongues. It’s a powerful myth and there are estimated to be almost 7000 living languages. The world continues to get smaller and just 23 languages account for half of the world’s population.
Imagine a device that would enable you to understand and converse in any of those languages that are not your native tongue? Artificial Intelligence (AI) powered by cloud computing now makes this possible. The latest solution was announced just a few days ago. Google presented its Pixel Phone Bluetooth headsets that come with the ability to translate across 40 different languages.
But Google is not the first. Just months earlier the Bragi Dash Pro earbuds were also launched and provide translations in 40 languages using an Apple iPhone. The technology has been building for a while. Last year a crowd-funded startup called Waverly Labs announced an in-ear translator named Pilot. The industry is bubbling with more start-ups like TimeKettle, “Travis” and “ili” with different approaches, degree of cloud dependency and languages.
Scientists, engineers and linguists have been trying to use computers to automatically translate one language into another almost from the beginning of the computer age. As early as 1954, the NY Times had a front page article announcing: “Russian Is Turned Into English By a Fast Electronic Translator”. Unfortunately, until lately the results have been limited and not totally satisfactory.
Why? The translation of one language into another turns out to be a hard problem. Writing a program with “if-then” algorithms just could not handle all the nuances and inflections common to human language but that our brains easily process. It was not until the advent of cloud computing and the concepts of natural language processing and deep learning that emerged in this century were better, timely translations possible.
Search engines with their vast databases and acres of compute power first tackled the challenge. Translation of text has now been around for a while. Google Translate (launched in 2006) and Bing Translator (launched in 2009) are notable examples. They kept getting better but let’s face it they were clunky to use and certainly tough to use in a face-to-face conversational setting. The leap forward has been to combine the advances in voice recognition, which is one AI function with translation, another AI function.
Today, the US is effectively a two-language culture (English and Spanish). In the world’s second largest economy over a billion people speak some dialect of Chinese. What will be the impact of this truly Sci-Fi universal translator capability? This is AI extending the capability of any (every?) human. Could language as a barrier be lowered, perhaps eliminated? Will the learning of “foreign” languages atrophy, much like our ability to do math by hand because of our dependence on calculators? Or, will there be a flourishing of appreciation for “foreign” theater and songs?
We are witnessing the Consumerization of a new and powerful technology that addresses one of mankind’s deepest desires. Roy Amara observed: “We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run.” Look how the smartphone has affected so many of your life’s aspects and it is only ten years old. How do you think your life will change when language is not a barrier?