Today’s language translation Apps are like self-driving cars: incredibly useful, promising, nearing maturity, and powered almost entirely by machines. It’s amazing that the technology even exists.
Despite this, machine translation is still sometimes cumbersome, if not downright cumbersome.
Consider a conversation I recently had with my neighbor Andre, who immigrated from Russia last year. Andre speaks little to no English and navigates the American dream almost exclusively through Google Translate, the most popular speech-to-speech translation app that first launched 10 years ago.
Through his phone, Andrew and I can have surprisingly deep conversations about where he comes from, how he thinks, how we can help each other, and what he hopes to achieve. But on more than one occasion, Google Translate failed to convey what Andre was trying to convey, forcing both of us to shrug and smile during the glitch.
However, as computers become more intelligent, Google, Apple, Microsoft and others hope to completely eliminate the language barrier that Andre and I shared that day. But that requires faster neural machine learning, and that “could take a few more years,” admitted one developer I spoke to.
Not that waiting matters. In fact, many consumers are surprised at how good today’s translation apps already are. For example, this video shows three Microsoft researchers using the company’s live translation software to conduct a conversation across multiple languages. The video is seven years old. But when I showed it to some friends, they reacted as if they had seen the future.
“Translation technology has made great strides in a very short period of time,” says Erica Richter, spokesperson for DeepL, an award-winning machine translation service that licenses its technology to Zendesk, Coursera, Hitachi and other companies. “But this did not happen in parallel with consumer awareness.”
I am an example of that. Even though I’ve been writing about technology for nearly 20 years, I had no idea how adept Google Translate, Apple Translate, Microsoft Translator, and Amazon Alexa were until I started researching this story after my fateful encounter with Andre. The technology is still unable to provide instant translation like you would expect from a real human translator. But turn-based speech-to-speech, text-to-speech, or photo-to-text translation is incredibly powerful.
And it gets better every year. “Translate is one of the products we develop that is fully powered by artificial intelligence,” says a Google spokesperson. “Since launching Google’s Neural Machine in 2016, we’ve seen the greatest improvements in accuracy when translating entire sentences rather than just phrases.”
At the same time, half of the six apps I tried for this story sometimes mess up even simple greetings. For example, when I asked Siri and Microsoft Translator to convert “Olá, tudo bem?” From Portuguese to English, both answered correctly: “Hello, how are you?” Google Translate and Amazon Alexa, on the other hand, answered more literally and clumsily: “Hello, are you okay?” or “Hello, is everything okay?” Not a complete failure. But enough nuance to cause hesitation or confusion in the listener.