Once widely considered to be nothing more than a sci-fi movie trope, artificial intelligence is now a cutting edge technology with many potential applications in the translation industry. However, to understand how AI reached this point in its development, it first helps to have a basic understanding of its history.
The History of Artificial Intelligence
Since the 1950s - As computer technology began to develop, innovators started to create programs that relied on Narrow AI, like the Ferranti Mark 1’s chess-playing program. Narrow AI programs are able to perform basic tasks that follow a set of rules or steps.
Since the 1990s - Following rules doesn’t exactly require an AI network to learn anything new. That’s why Narrow AI gave way to Machine Learning developments. Machine Learning occurs when algorithms allow AI to analyze massive amounts of data, separate that data into groups, and make predictions from it. For instance, Machine Learning is at play when a video streaming service offers you recommendations based on past behavior.
Since the 2000s - In recent years, AI specialists have focused on the concept of Deep Learning. In this branch of the AI field, algorithms are based on the actual structure of the human brain. These sophisticated algorithms allow programs to solve problems they may not have been designed for, learning from past actions to adapt to new situations. Thanks to Deep Learning, we have AI programs that can read lips more accurately than humans.
Right Around the Corner - As the technology continues to advance, it won’t be long before General AI is a reality. This form of AI is essentially what’s been depicted in sci-fi movies: smart robots with human characteristics. These programs could theoretically perform general intelligent actions as naturally as humans can.
What it Means for the Translation Industry
Advances in Deep Learning and Machine Learning have substantially improved the accuracy of machine translators. In the past, these programs have typically used Narrow AI to break a sentence down into fragments. It would then research the individual words within those fragments in a foreign language dictionary and apply post-processing rules to put the fragments back together in the form of a cohesive sentence.
Anyone who has ever used an online translator knows this often results in mistranslations.
Luckily, AI may be changing that. Google recently switched their Google Translate service over to a new system, known as Google Neural Machine Translation. The purpose of this move wasn’t simply to leverage AI and improve the accuracy of Google’s online translator. The goal was also to determine if AI could learn to translate between multiple languages.
For example, Google researchers wanted to find out if, after teaching the service to perform a Korean to English translation, as well as a Japanese to English translation, the AI would be able to perform a Korean to Japanese translation (or vice versa) with no additional training. The test was successful.
This development has wide-ranging implications for anyone who works in an industry that requires the efficient translation of documents and content. Law firms that regularly engage in international litigation can comb through foreign-language documents more quickly.
Businesses can use AI-based chatbots to offer customer service in more languages than ever before. Translation firms can leverage the technology to finish client projects more quickly. It opens the door to make multimedia translation even more efficient than ever. This technology has interesting implications for machine-conducted cross-cultural translations of information, photos and videos included.
Granted, it will still be important to rely on human translators when preparing important documents or materials. While the ability of AI to translate content accurately may be improving, AI still can’t address the cultural nuances of a foreign audience.
It can, however, simplify the translation process. In our digital world, when it’s easier than ever to reach people throughout the globe, everyone from CEOs to online shoppers benefit.
