Salvador Ordorica is the CEO of The Spanish Group LLC, a first-class international translation service, that translates over 90 languages.
Speaking to investors and running a business in the translation industry requires one to interact with a diverse group of people from a wide array of backgrounds. Still, no matter where you are or who you talk to, the general idea that technology will one day, perhaps soon, eclipse the translator is pervasive.
The thought is anything but new. Humanity has long dreamed of the day that technological advancements will produce the metaphorical battering ram with which we shall smash the language barrier. The idea of using machines for translation was proposed as early as 1947, and science fiction has long assumed it was an inevitable outcome of our technological advancement. To the credit of these dreamers, the ideas of lab-grown meat, space stations and the defibrillator were all once just science fiction and are now a part of our reality.
Google, Microsoft and other companies have invested millions into automatic online translation. It does seem inevitable from the outside that, like dozens of other professions, the translator may soon become obsolete thanks to technology.
What is the reality? I will briefly explore the future of technology and translation.
Consumer Versus Professional Translation
In our evermore intertwined and interconnected world, language has become an increasingly common hurdle. We are presented with more foreign languages in our daily lives than at any point in history. It isn’t rare for videos from China or India to go viral in the West and vice versa, nor is it uncommon to take classes or lessons on Zoom from someone on a different continent. Immigrant marriages have been consistently on the rise for decades as well.
Most of the consumer-level translation technology being worked on today is being made to handle these needs, to help facilitate day-to-day interactions between millions of people and to ease general understanding. However, this is where the confusion between this technology and what trained translators can do begins.
Skilled Translation: More Than Word Replacement
You don’t need sophisticated language skills to compliment someone, but what about asking them about aspects of a complicated legal case? There are different degrees of complexity required in language for different tasks.
While machine learning translation apps like Google Translate have proven useful for fundamental cross-language interactions, they can be insufficient for sophisticated health, legal and business dealings. Scans, handwriting and acronyms are all also current hurdles machines have yet to successfully overcome. Most currently required translation work is for civil or legal purposes, where it is legally necessary to have a party responsible for the translation and not AI.
Literal translation has often failed to convey correct meaning, often with humorous or disastrous results. In 2019, this idea was tested in an academic medical setting. Google Translate got 92% of the sentences correct, while 2% of the mistranslations could have led to severe harm for the patient. This test was with a ubiquitous pairing of English-Spanish and after years of neural machine translation learning from Google Translate. Mistakes will increase with less common pairings, and such outcomes may only worsen the divide for underserved communities and minorities if adopted en masse. For example, many languages from Africa are not currently present in Google Translate, such as Oromo, a language spoken by roughly 34 million.
Even if machines could correctly translate the language word for word, it may still not be entirely of use. Language, even in a basic conversation, is more complex than we typically realize. Speaking with a purpose to someone requires relating to them. Therapists, speechwriters and salespeople spend years perfecting tone and word choice and creating a pattern of speech that is wholly unique to their goals. Languages evolve, syntaxes shift and machines will need to “unlearn” proper usages to better fit an audience. Slang, memes, humor and emotional context are essential to our everyday interactions, and marketers understand this. Accordingly, these are all areas that machines will be unable to imitate or replace for the foreseeable future.
For now, machines have neither the accuracy to take over for a medical translator or the personality to recalibrate an emotional pitch for a new cultural context.
How Translation Agencies Can Still Thrive
So, what steps can translation agencies take to thrive today amid the myriad of technological advances we are faced with? The first is to transition into the fields that technology cannot replace. By diversifying your offerings, you can ensure that no new app suddenly undercuts your entire business model. One option is to focus on growing but underserved markets. There are millions in emerging markets around the world coming online and looking for professional language solutions.
For example, Indonesia is considered a top emerging market for 2020 and has almost zero effective automated translation services. Part of the reason is that daily conversation among Indonesians is mixed with local ethnic dialects. Becoming integral to the functioning of government agencies and businesses in that area could provide a way to diversify revenue.
Moving into medical and legal services is also an option, as we covered earlier. Grow your services into that area by first focusing on niche needs. Serving specific immigrant groups and offering specialized services such as translations for new USCIS or immigration requirements will ensure that your organization offers critical services that are unlikely to disappear in the near future.