Insights November 17, 2020

Machine translation – helping the disadvantaged?

Living Word
By Living Word

In light of the current pandemic, it is interesting to see how machine translation (MT) technologies can assist health communication in rural areas. But just how effective are they, and how do they help disadvantaged communities access important information?

Many experts agree that MT can help people access relevant information where language might be a barrier. Let’s take Covid-19 as an example: while more and more people become infected across the world, there are plenty of regions and communities where the local language becomes a barrier to timely information about the latest news on the disease, as translations from English, for example, simply aren’t produced quickly enough.

In response to this particular issue, Dr. Rejwanul Haque, who obtained his PhD at Dublin City University, decided to get a team together to work on building eight multilingual MT engines trained to translate Covid-19 materials between German, French, Italian and Spanish into English and vice versa. Now, users simply need to go online, choose their target language, paste the text they want translated, and voilà – the MT server translates the text for them.

Problems faced by MT developers

As machine translation is a massively complex field, however, it takes time to get it right. Unfortunately, even the most professional machine translation devices and programmes are still not even close to the standard that human translators provide. After all, millions of sentences and phrases need to be fed into programmes, which in turn requires a lot of training and experimenting. Current MT systems still lack sufficient accuracy – something that might be particularly problematic in health contexts.

A bright future?

While it might not be perfect yet, machine translation can indeed help those disadvantaged communities that don’t speak one of the world’s main languages to access important general health information. But it must be pointed out, of course, that the most marginalised groups don’t always have access to running water, let alone a computer with internet access, so it’s questionable whether these particular groups would benefit much from such developments.

In the future, however, MT systems are likely to get both better and more impactful. As well-developed machine translation algorithms improve, they might not only be useful in healthcare settings and in preventing the spread of diseases; they might even help to transform aspects of everyday life…