The Consequences of Global Outreach with Artificial Intelligence
Submitted by: Jose R. Burgos, senior project manager at Innerlingua Translation Services
We have all heard about how AI is changing the world to increase productivity and get it at the speed of light. Globalizing content is no exception. It is a fast, inexpensive way to increase exposure in international markets when considering the communications challenges from translation and interpretation. But, at this point in time, can we rely on AI?
In the early development of language AI, it was not the best tool to rely on when it came to translation. Most of the engines were around 60-80% on target. Machine translation engines and AI in general are becoming faster and more accurate. However, even with the improvement of AI engines like Google neural machine translation, Deepl, Yandex, Baidu or Chat GPT, just to name a few, and their staggering skewness of 95%, there are still major issues with translation quality.
Poorly translated terminology can lead to misinterpretation of product assembly and operations, which could cause disastrous results. For instance, with AI translation that might boast 99% accuracy, a 300-page manual could have up to 90 errors. That’s the best-case, error-free scenario. So, would you want to take that risk? And do you want to take on the potential responsibility should a terrible error occur?
We will continue to see a rise in AI, however, translation work done by machines should not overshadow the human element.