The potential usefulness of automatic computerized translation was recognized by the very first AI researchers in the 1950s. But it wasn’t until new algorithms emerged in the 1980s and 1990s that the field made significant progress.
Now, translation tools of great sophistication are playing a growing role in both everyday office use and for specialized fields, as the economy becomes increasingly globalized and companies sell products and services in multiple markets.
The poster child for computer translation is Google Translate, the easy-to-use, general purpose Web-based translation engine that can handle nearly 60 languages. Google’s Translate has the same 800-pound gorilla status in its world that the company’s namesake product does in search.
Not only is Google Translate the world’s most widely used translation system, it’s also the most technically advanced. What’s more, says Kevin Hendzel, spokesman for the American Translators Association, is that it has the advantage of being able to tap into the vast contents of Google’s search engine in its quest for “raw text.”
In part because of Google Translate’s growing popularity, the office of the future is getting a verb of the future: “gist.” To “gist” means to use Google Translate or some other program to get a quick sense—a gist—of something that’s been written in another language, says Don DePalma, an analyst at Common Sense Advisory, a firm in Lowell, Massachusetts, that follows the translation industry.
This might mean an Arabic web page being “gisted” to see if it’s about politics or soccer, or someone sitting down to “gist” a Beijing site in Chinese for the latest official views on yuan-dollar currency levels.
“Gisting” is expected to take on an ever-larger role in everyday office use, especially as the amount of information on the Web continues to expand.
And as home computer users appreciate the accuracy of Google Translate for casual Web surfing—like reading foreign newspapers when big breaking stories occur—they will bring those habits with them to the office and start doing the same thing at work.
Among office computer users, “I would bet they would be much more likely to click the ‘Google Translate’ button now than they would have five years ago,” says DePalma.
Government is already a big “gister,” especially for national security. American intelligence agencies, notoriously short on translators for strategically crucial languages like Farsi and Arabic, are already power users of machine translation technologies like Google’s, which they rely on to quickly decide if, say, a website deserves closer study. That frees up scarce human translators to focus on sensitive documents, where they are most needed.
Since even the best computerized translation programs still make obvious mistakes, all translation software, including Google’s, requires human beings to double-check their work. That sort of human backstopping, says DePalma, will be necessary for the foreseeable future, at least with important documents.
As a result, many companies sell a combination of computerized and human translation services that work in parallel; these companies make up the global “localization” industry, estimated by Common Sense Advisory to be worth $40 billion a year and growing—in spite of the anemic global economy—by 7 percent annually. (Google Translate is usually counted as a general-purpose Web service, rather than a specialized localization software provider.)
Typical localization customers are companies that need to translate instruction manuals from one language to another, or to bring out a product in another culture without unwittingly violating local customs and taboos. Many localization firms target technical fields, like medicine or law, with highly specialized dictionaries that work in conjunction with in-house translation software. IBM, which has been working on computerized translations for decades, licenses its software to localization companies such as Lionbridge of Waltham, Massachusetts.
The localization industry turns out to be a good proxy for global economic development. In its early days, European languages dominated the demand for localization products. Japanese, then later Chinese, became important, and today, Russian and Brazilian Portuguese are mainstays, as are the languages of the Indian subcontinent. There is even the beginning of a localization boomlet in Romanian and Vietnamese, on account of the emergence of those countries as what Wall Street investors call “frontier economies.”
To be sure, experts use the term “gisting” with good reason, since computers are still only consistently reliable for conveying the general sense of a passage of text.
No one would confuse even the best computer translation with the carefully considered phrases of a human translator, especially with long, dense paragraphs. And computer-generated text often isn’t very pleasurable to read, being to prose something like what computer-generated speech is to the human voice.
So for the foreseeable future, it will still require a certain amount of human intelligence to figure out what the translating computer “meant” to say.
A horrifying new AI app swaps women into porn videos with a click
Deepfake researchers have long feared the day this would arrive.
We can’t afford to stop solar geoengineering research
It is the wrong time to take this strategy for combating climate change off the table.
Meet Altos Labs, Silicon Valley’s latest wild bet on living forever
Funders of a deep-pocketed new "rejuvenation" startup are said to include Jeff Bezos and Yuri Milner.
The new version of GPT-3 is much better behaved (and should be less toxic)
OpenAI has trained its flagship language model to follow instructions, making it spit out less unwanted text—but there's still a way to go.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.