Native vs Automated Translation Services UK: Pros and Cons
As technology evolves, we see solutions such as AI-powered artwork tools and automated writing platforms like ChatGPT becoming more widely accessible and possibly offering businesses and organisations easier, cheaper ways to manage content creation.
However, one area where automation is unsuitable is in a professional translation context, where software and apps don’t have the human ability to spot nuance, context, or tone – and a direct dictionary translation may not mean the same thing at all.
Absolute Translations, industry-leading certified translation company explains how automated translations work, where they may be of benefit, and scenarios where they strongly advocate for a professional, native translation to avoid reputational damage.
What Is an Automated Translation?
An automated translation might, to a non-speaker, look fairly similar to that produced by a qualified, highly experienced translator, but look a little closer, and you begin to see why the translation process is entirely different.
Automated software and tools use digital dictionaries. They can also use algorithms to try and identify grammatical inconsistencies, where the placement of symbols in one language differs from the next.
They’re quick, cheap, and on-demand, but unfortunately, rarely accurate enough to pass the quality expectations of a customer, service user, or overseas buyer.
Here are a couple of high-profile examples that illustrate the point perfectly:
- Electrolux, the Swedish home appliance brand, translated a slogan into English when launching a vacuum cleaner in America. The direct translation, while technically correct, failed dramatically, stating that ‘nothing sucks like an Electrolux’, of course not realising that if something ‘sucks’ in US English, it means performing very badly.
- A Florida hospital admitted a patient, misinterpreting his medical history in Spanish, which showed that he was ‘intoxicado’. This did not mean intoxicated, as the doctors had assumed, but he had suffered a reaction to food. The outcome was a multi-million-dollar damages lawsuit.
While most low-quality automated translations may not end in such expensive embarrassments, it is all too easy to assume a computerised translation is 100% accurate, only discovering the meaning, intention, or message has been changed beyond recognition.
How Do Native Translations Differ From Software Translations?
Native professional translators spend years perfecting their crafts, and improving their skills, even if they are already fluent in multiple languages and dialects.
Rather than translating individual words and sentences, they digest the whole article, manuscript, instruction, or marketing translation text before beginning to ensure they clearly understand what the material says, who it is intended for, and what it means in the original language.
Only then will they begin the translation work, adjusting sentence structures, terminology, and particularly technical or slang phrases to ensure they are culturally correct, sensitive, and mean the same thing as intended.
The limitations of automated translations are that a machine, no matter how advanced, cannot recognise nuance and may replicate a neutral phrase in one language that is potentially offensive in another.
How Accurate Are Automatic Translation Tools?
Some of the more well-known translation tools, such as Google Translate, claim to be 94% effective, but that is based on word-for-word translations, meaning it cannot find the correct word 6% of the time.
As we’ve explained, it isn’t simply picking the correct word that makes a professional, effective, and on-brand translation. Google Translate uses a three-part translation model where quality is ranked according to the following:
- Phrase-based translation
- Neural translation
- Human translation
An update released in 2016 means that the algorithm looks at whole sentences rather than individual words, which has undoubtedly improved performance. However, it remains incomparable to a professional translator when addressing human translation.
Some more widely used languages, such as English to Spanish translation, approach the top end of the accuracy spectrum. Still, in other languages, such as English to Armenian translation, it scores 55% – meaning that 45% of the content is expected to be inaccurate.
A peer-reviewed study conducted by researchers from UCLA Medical Centre and the Memorial Sloan Kettering Cancer Centre found that automated translations were, on average, 67.5% accurate in English to Farsi and 81.7% in English to Chinese.
These statistics show why an automated translation, while suitable for personal use or conversations, cannot reach the 100% flawless translation you receive when hiring a professional.
Note that Google Translate is considered one of the top digital translation tools, and these accuracy metrics are likely to be substantially lower elsewhere.
What Are the Potential Benefits of an Automated Translation?
While we do not recommend using automated translations for professional projects, particularly technical translation where one error can be extremely costly, they can be a useful tool.
Travellers and visitors in different countries can use translation apps to find out how to ask for directions, engage in a discussion, or read road signs, for example, and can access immediate results.
Automated translations can also be an effective way to establish a rough understanding of a document, or translate a label, provided you acknowledge the potential for inaccuracies.
Advantages include speed and availability, where online translation tools can respond at the touch of a button. However, be cautious about using the right spelling if entering a short piece of text or phrase in a language you do not speak or guessing the spelling from a spoken word.
For example, in Greek, the word for yes is ‘ναι’, pronounced ‘nai’, which is commonly misspelt or mistaken for no – with the opposite outcome.
When Should You Prioritise a Native Translation?
The key areas where an exact translation and precise language are essential include safety warnings, medical labelling, healthcare records, contractual and legal translation, and other translations of a technical nature or where an error could cause negative outcomes.
However, it is also imperative that companies, brands, and commercial enterprises ensure that any globalisation efforts are supported by native translation services.
Launching a product with obviously substandard translation can appear, at best negligent and, at worst, inappropriate. Where you cannot afford a margin of error or are unwilling to invest in publishing materials that contain mistakes, a professional native translation is the only viable choice.
