Don’t Trust AI: Use Humans at a Translation Company
by Chad Richardson
You might have heard that artificial intelligence is improving. Services like Google Translate make it easy to copy and paste text in a box online and get a full translation in a matter of seconds. Unfortunately, these translations have a reputation for being laughably bad.
But new artificial intelligence has some people reconsidering artificial translation. In November 2016, a new, advanced translation service from Google Translate showed improved translation results in 16 different languages. The new results have understandably impressed people around the world.
Whatever progress artificial translation has made, a recent competition between Human and AI translators shows that artificial intelligence still has a ways to go. The competition pitted professional human translators against three AI translation services. In the competition, all of the entrants had 50 minutes to translate literature and non-literature passages from Korean to English. Two professional translators judged the results, not knowing whether human or machines did the translations.
The results did show some promise for the advanced Google Translate, the highest scoring AI software with a score of 28 out of 60. However, the human professional translators defeated them by a landslide, with 49 out of 60 points.
AI translation services might help laypeople quickly get the gist of a document, but the services still have serious limitations. They miss key slang and proper names, and often they miss even basic grammatical tenses in the translated languages. With more sophisticated types of translation, such as technical science terms and legal documents, the problems grow exponentially. For a strong translation, using a professional, human translator remains the only way to get the job done.
Contact us to receive sophisticated, human done professional translation services in a wide variety of languages.
Related Articles
Why Generic Machine Translation Fails Enterprise Scale
Generic machine translation (MT) engines frequently fall short when enterprise teams attempt to deploy global content at scale because they fail to preserve specialized corporate tone or distinct target audience requirements. Standard MT outputs lack the direct controls needed to differentiate between communication intended for technical experts versus broad consumer bases, creating friction in international…
Why are CAT tools vital for enterprise cost control?
In a traditional localization workflow, every document – whether a product manual, marketing campaign, or legal contract – is treated as a completely new artifact. This leads to redundant expenditures, inconsistent brand voice, and long post-delivery revision cycles. CAT tools transform translation from a transactional expense into an accruing digital asset. By embedding these platforms…
E-Commerce Transcreation for Global Growth
Discover why deep cultural localization and e-commerce transcreation are essential for engaging the booming middle-class markets across Asia and beyond. Key Takeaways: The global consumer class has surpassed 4 billion, with Asia now holding a massive 50% share of the middle class. To capture market share in these emerging hubs, brands must pivot from generic,…
here
for you
We’d love to learn more about your translation and localization needs.