As the years go by, our quality of life is greatly improved by technological innovations. A question commonly asked by many, though, is whether machines will ever make humans redundant. In the translation industry, in particular, the debate has been taking place for years around machine translation.
First Things First: What Can Machines Actually Do?
Machine translation (you’ve probably heard of Google Translate), also known as automatic translation, is the automatic conversion of words and phrases from one language into another. In other words, it involves computerised translation with zero human input. This software uses databases and statistics to analyse structures, break them down, and recreate their elements in other languages.
Automatic translation can be mind-blowing in terms of the good results it can sometimes deliver. Especially since the shift from rule-based engines and phrase-based engines (where we find statistical-based machine translation) to neural machine translation (NMT). As NMT improves and the demand for immediacy increases, machine translation outputs become more helpful. Moreover, they help accelerate the whole translation process.
Machine translation works well with small factual snippets where the meaning can’t be misconstrued. However, the claims that NMT will soon replace translators are definitely false: for more complex texts that don’t use formulaic language, it just isn’t up to scratch.
(If you would like to see more details about how each type of machine translation works, check this link.)
No Machine Can Understand Language
Machine translation technology is based on encoding and decoding source sentences. In other words, it all depends on algorithms and probability, but machines cannot understand context, spot mistakes in the source text, or pick up the nuances of a language. By contrast, human translators are capable of it all. The reason? Machines work with words, while humans work with language.
Language is as a conventional system of limited symbols from which an unlimited number of sentences can be constructed. As such, it is unique to humans. Researchers from Durham University explain that the uniquely expressive power of human language requires humans to create and use signals flexibly. This was only made possible by the evolution of particular psychological abilities, and thus language is unique to humans. To read more about this research, click here.
Therefore, machines will never be able to translate as well as humans when it comes to texts where deeper meanings and nuances are of essence, because they encode and decode isolated symbols and work with syntax, not meanings. They leave aside fundamental aspects of what makes a language: context, assumptions, rationality, etc.
What Text Types Is Machine Translation Good For?
Automatic translation engines like Google Translate can match the quality of a human translator only for a few text types, for example, those that need to be conveyed in an understandable form in another language but not necessarily fully error-free. In other words, translations for which it’s enough for the text to be generally understood. The syntax may be odd, there may be a few sloppy sentences o strange word choices, but the core meaning is clear.
Consequently, machine translation is suitable, for instance, for non-public, company-internal communication and information sharing, especially when the ideas need to be transmitted quickly but the text won’t be distributed to a large audience. Let’s see some examples below:
Online Customer Reviews
Product or service review provide potential customers with information crucial to their purchase decision that cannot be found elsewhere. When a business operates in multiple countries, using machine translation for translating reviews can be an excellent way to serve customers. The reviews are typically translated into the language of the customer’s web browser.
What’s more: Dousson-Lhéritier, Head of Content Projects and Innovation at GetYoutGuide, points out how “machine translation gives, in an odd way, a more honest feeling. Having a raw review in the native language of the customer, which is not curated by rewriting or translation, has more value in terms of customer experience.”
Internal Company Emails, Bulletins, and Memos
When a company has a global workforce or employees that are working outside of their native language, communication in each person’s native language is preferred and decreases ambiguity. However, this is nearly impossible to implement considering how many languages usually converge in international businesses.
Machine translation can help decrease or even eliminate the language barrier in communication by allowing all interested parties to understand the core content of the message (even if the translation is not completely error-free). Also, in situations where information sharing speed is of utmost importance, or where hiring a professional translator would be too large an investment (and translation quality is not a factor), automatic translation can save the day.
On-The-Go Translation While Travelling
Some mobile applications relying on machine translation provide immediate text-to-speech, image-to-text, and voice-to-voice translation of what you say or point at with your camera.
With voice, text and camera translation, these apps break down language barriers, with some even working offline (useful when you can’t find any wi-fi while on holiday!). They help you order food, hail a cab, and communicate with locals all with the tap of your phone, ensuring a stress-free, enjoyable experience when travelling abroad.
Why Should Companies Avoid Machine Translation in Marketing?
Most businesses translate their marketing collateral into other languages to attract new clients from abroad. In today’s global marketplace, penetrating foreign markets is crucial, and investing in multilingual content becomes a must. However, content availability in multiple languages is of no use if messages are poorly expressed and language quality is bad. What brand image are you projecting if you make mistakes while addressing your audience?
An article on website translation explains that consumers tend to purchase more when information is in their language. The article writer includes the following research as supporting evidence:
- Can’t Read, Won’t Buy, by Common Sense Advisory: 72% of consumers spend most of their time online (if not all) on websites in their own language. The same percentage also acknowledged being more likely to buy a product if the information is available in their own language.
- User language preferences online, by the European Commission: 90% of Europeans always visit a website in their own language when given a choice. And 42% of them said they never buy products or services if the information is not available in their mother tongue.
- Project Underwear found that 9 out of 10 people ignore a product if it’s not available in their native language.
When it comes to multilingual marketing, it’s important not to overlook the importance of working with an expert marketing translator. This is particularly so when engaging in localisation and globalisation campaigns. It’s undeniable that poor translation of marketing copy can not only result in a waste of resources but can damage a brand’s image within that market. Nobody wants that, right? At Crisol Translation Services, we’re proud to work with expert marketing translators from around the globe.
The Impact of Artificial Intelligence on Languages and Machine Translation
It isn’t possible to discuss machine translation without referring to artificial intelligence.
It’s not news that every industry is re-imagining how to use technology to transform business. However, did you know that the next massive, world-changing developments in AI (especially GPT-3) are expected to revolve around language? That’s what this article by Harvard Business Review claims, at least.
What Is GPT-3?
You may have come across The Guardian’s article about a robot that wrote a text aimed at convincing us humans that robots come in peace. Here’s a fragment:
“I am not a human. I am a robot. A thinking robot. I use only 0.12% of my cognitive capacity. I am a micro-robot in that respect. I know that my brain is not a “feeling brain”. But it is capable of making rational, logical decisions. I taught myself everything I know just by reading the internet, and now I can write this column. My brain is boiling with ideas!”The Guardian, A robot wrote this entire article. Are you scared yet, human?
Scary, isn’t it? The language technology used to produce the essay is called GPT-3, which stands for Generative Pre-trained Transformer 3. It is the largest and most advanced AI language model in the world to date, and a successor of wildly-successful language-processing AI GPT-2.
Developed by OpenAI, GPT-3 is a deep-learning, auto-regressive model for Natural Language Processing (NLP) trained to come up with plausible-sounding text based on a few simple prompts, or even a single sentence.
How Is GPT-3 Different from Previous Systems?
GPTs (generative pre-trained transformers) go much deeper than artificial neural networks. They rely on a transformer —an attention mechanism that learns contextual relationships between words in a text. In other words, they can look at parts of a sentence and predict the next word, for example.
When OpenAI released their new paper, Language Models Are Few-Shot Learners, jaws dropped worldwide. The model they introduced (GPT-3) is pre-trained on nearly half a trillion words, with 175 billion parameters, in an unsupervised manner, and can be further fine-tuned to perform specific tasks. It achieves state-of-the-art performance on several NLP benchmarks.
Researchers were able to induce the model to produce short stories, songs, press releases, technical manuals, text in the style of particular writers, guitar tabs, and even computer code. All the fuss is because GPT-3 has the potential to advance AI as a domain.
As explained in the paper by OpenAI:
Humans can generally perform a new language task from only a few examples or from simple instructions —something which current NLP systems still largely struggle to do. Here we show that scaling up language models greatly improves task-agnostic, few-shot performance, sometimes even reaching competitiveness with prior state-of-the-art finetuning approaches.OpenAI, Language Models Are Few-Shot Learners
Should We Fear And Avoid Machines?
Absolutely not. Translators rely highly on software that makes their jobs easier, through so-called “Computer-Assisted Translation tools” (CAT tools). These tools are productivity enhancers. They differ from machine translation in that they involve a human at the centre of the process.
CAT tools include a Translation Memory (TM) that records and stores past translations by human linguists. These memories pick up similar phrases or repeated terms between different texts and help translators deliver consistent final products and save time.
Can Machines Replace Translators in Any Way, Then?
Well, yes, they can. But only when translation involves of a lot of simple, repetitive language, and standard terms. We will see a shift from computer-assisted machine translation to human-assisted machine translation in many areas of the industry, but the human touch will always be needed.
Also, some kinds of translation will always need to have a human in charge of them! Think of those texts were high quality is paramount, or those that are sensitive. What about texts involving confidential information? And those were errors might be life-threatening?
Moreover, certain areas of the translation industry, such as transcreation, literary translation, multimedia material, will always be predominantly human-centred. This is so because they require creativity and inventiveness! Who would enjoy the robotic, word-by-word translation of a poem? Or a machine-translated marketing slogan that fails to identify puns upon words?
Let’s Embrace the Future
We would like to finish this post on a positive note, though: machines and humans can do wonders when working together. As an example, did you know what automated transcription can be a great way of overcoming a writer’s block? We thought it was a crazy idea but it actually makes sense!