Don’t disregard the Human Element in Language Development, even in an increasingly AI-enabled world!
We have some pretty good examples of why…
I think it’s fair to say that we have all experienced the speed and convenience of AI tools in work or at play. Assisted workflows, brainstorming inputs, copy editing and refinement, report or research summaries, as well as creating your kids’ birthday invitations or your own avatar – these are among the most obvious.
However, have you ever tried using AI for translation? It is not a far stretch of the imagination, given that users already tap into google translate every time they visit a foreign language website or require other foreign language inputs. Take tourists for example! You paste a paragraph into a browser or transcribe a conversation, click a button, and instantly get a fluent-looking back translation in almost any language.
But is it authentic and contextually correct? Communication in a foreign language has never been about direct translations, that is, basically swapping words from one dictionary to another. In fact, that’s exactly where it has gone horribly wrong with a plethora of examples existing all over the internet (just do a quick spot of research).
Real human, connected communication is shaped by culture, context, nuance, personal identity and lived experiences. Machines don’t quite carry these insights organically. Today, AI is trained on data and that limits its capability to replicate true human communication and insight.
A technically accurate word-for-word AI translation can still completely miss the intended meaning if it doesn’t understand the culture and people behind the message. Heck, even community specific advertising campaigns for communities that are fluent in English, such as the Indian community, need to be adapted and tested for cultural relevance.
So taking all that into account, true communication pieces targeted at diverse audiences are not about translations or word-for-word copy, it is about transcreation, copywriting in a way which is culturally relevant whether that be in a foreign language or not!
Here are some important considerations when using AI as a translation tool:
1. AI can be confidently wrong, human oversight is essential
Translation by generative AI tools can sound super smooth and confident at first view. And let’s face it? If you are not familiar with the language, I mean, really have a lived experience associated with that culture, how can you even judge the authenticity of the translation?
This means critical mistakes are often missed by non-language experts. The risks can be far-reaching in business communications such as, legal contracts, medical consent forms, public safety notices, and diplomatic negotiations, subtle errors in numbers, proper names, technical acronyms, or negation where a small error can lead to real-world legal consequences.
For responsible brand communications or public sector social campaigns, communication must be both sound in terms of the information presented and culturally equivalent. Human language experts trump machines in this regard.
2. AI can’t really “transcreate”, capture brand soul or reflect cultural nuance
When it comes to marketing, brand strategy and creative advertising, straightforward translation falls short and always has. Hence, the birth of Multicultural Marketing and Cross-Cultural communications where cultural relevance is paramount.
What is transcreation? It is basically creative copywriting in a foreign language, which while being a correct translation also, critically, understands and correctly conveys relevant cultural context and nuance.
Generative AI is limited for several reasons:
- Generative AI is data driven and can’t generate original creative concepts.
- Effective in-language copywriting (transcreation) is designed to evoke specific emotions and connect with communities. AI can’t perceive, think, or feel. So, it can’t create copy that will emotionally connect with the audience. That is why there is an influx of generic AI Slop in the market nowadays!
- Multicultural experts run foreign language copy through cultural relevance testing with native language experts and community members to ensure cultural consistency across slogans, imagery and tone. AI cannot run this and misses true cultural insight as a result.
3. If you don’t believe me, then check out some of these bloopers
Here are some funny real-world bloopers of generative AI outcomes that ignore cultural context:
- Pepsi: Famous brand slogans suffer when translated literally, such as Pepsi’s slogan “Come alive with the Pepsi Generation” being translated in overseas markets as “Pepsi brings your ancestors back from the grave”.
- Ford: A Ford Arabic advertisement translated “high-quality body (of the car)” using the word for corpse instead of vehicle, that is, “high-quality corpse”.
- KFC: KFC’s slogan “Grill with confidence” was perceived as a rigid command in Spanish rather than persuasive marketing copy, making it offensive and uneffective.
- Powerade: Powerade’s “Chikara Mizu” in Japanese translated as aggressive or forceful water.
- Nike “Get Fat” (?): This has to be my absolute favourite! A Nike Chinese New Year shoe customisation combining “fa” (get rich) and “fu” (luck) resulted in “get fat” when read together.
4. The future is probably somewhere in the middle
In the current AI World, where mass produced or “generated” content is running wild, the role of language professionals is evolving into language intelligence specialists and language strategy advisors.
Humans play a key role in oversight, infusing authenticity, cultural insight and nuance into the work produced. We see it time and time again across industries.
When clarity, cultural relevance and brand reputation are at stake, human judgement should still remain the ultimate quality standard.
Language and communication is, and will always remain, a defining human endeavor.