Professor Mahmoud Kamel
A few years ago, the question “Will AI replace human translators?” sounded like a question from the future. Today, it is already a question that many translators, academicians and students have to think about.
Artificial intelligence has entered the translation world very quickly. Tools such as ChatGPT and other large language models can translate, rewrite, summarise and produce texts in different languages in only a few seconds. The quality is also becoming better, and this has made many people wonder about the future of the translation profession.
The first chapter of Translation Studies in the Age of Artificial Intelligence, edited by Sanjun Sun, Kanglong Liu and Riccardo Moratto, looks at this question from an interesting point of view. Written by Donald A. DePalma and Arle Lommel, the chapter is called “Transforming Translation: The Evolution and Impact of AI on Language Transfer and Communication.”
What makes this chapter interesting is that the authors do not see AI as something that suddenly appeared and changed everything. They place it within the long history of technology and translation. In other words, today’s AI revolution is another important stage in a story that started many centuries ago.
Translation has always changed
Translation is often seen as a very old human activity, and this is true. But the way people translate has changed many times.
The invention of writing made it possible to record and transfer information. Printing later made books and knowledge available on a much larger scale. Then came computers, the internet and computer-assisted translation tools. Translation memories, terminology databases and other digital tools gradually became part of the normal work of professional translators.
So, translators have never worked completely outside technology.
The difference today is the speed and power of the new technology. Artificial intelligence, especially neural machine translation and large language models, can deal with huge amounts of language data. It can produce a translation very quickly and can sometimes deal with complicated language in a surprisingly good way.
This does not mean that the machine understands language in exactly the same way as a human being. But it means that the traditional way of doing translation is changing.
What do we really mean by translation?
One of the interesting ideas in the chapter is that we may need to think again about the meaning of the word “translation.”
For many people, translation simply means changing a text from one language into another. A French text becomes an English text, or an Arabic text becomes an English one.
But communication today is much wider than this.
Information can move from written language to speech, from speech to subtitles, from video to text, and from one form of media to another. Translation can involve subtitling, dubbing, transcription, localisation and other forms of communication. The authors therefore suggest looking at translation as a wider process of transferring information between languages and also between different forms of communication.
This is especially important now because AI is not working only with traditional written translation. It is increasingly being used with speech, video, subtitles and multimedia content.
For Translation Studies, this means that the field itself has to become wider. Researchers cannot study only the relationship between a source text and a translated text. They also have to study the new relationship between humans, machines, language and communication.
The human translator is not finished
Perhaps the most important question is what will happen to the human translator.
There are two extreme views. One says that AI will soon replace translators. The other says that machines will never be able to do real translation.
The authors offer something between these two views.
They see the possibility of an “augmented translator.” This means a human translator who uses AI to improve his or her work. AI can take care of some repetitive and time-consuming tasks, while the translator remains responsible for the final result.
This is already happening in many areas of translation. A machine can produce a first version of a text. The human translator then checks it, corrects it and improves it. This process, usually called post-editing, is becoming an important part of professional translation.
But the authors suggest going beyond the idea of the human simply cleaning up machine output. AI can also give suggestions, provide terminology and help the translator find information. In this model, the human translator is still in control. The technology is there to support the translator, not to make all the decisions.
This is perhaps a better way to look at the future. Instead of asking, “Who is better, the human or the machine?” we should ask, “What can the human do better, and what can the machine do better?”
AI has obvious advantages. It is fast. It can work with large quantities of information. It does not become tired after translating hundreds of pages. It can also help with terminology and consistency.
But translation is not only about producing sentences.
A translator has to understand the situation, the purpose of the text, the culture behind it and sometimes the meaning that is not directly expressed in words. A literary translator has to understand humour, irony, metaphor and style. A translator working in law or diplomacy has to be very careful because a small change in meaning can have serious consequences.
These are areas where human judgment remains very important.
AI has its own problems
The chapter does not present AI as a perfect solution. There are still important problems.
One problem is the quality of the data used to train AI systems. Large language models are trained on enormous amounts of information, but this information is not equally available for all languages. English and some other major languages have much more digital material than many less widely used languages. This can affect the quality of AI translation.
There are also questions about accuracy, consistency, privacy and cultural understanding.
An AI system can produce a sentence that sounds very natural and still be wrong. This is one of the dangers of AI translation. A fluent sentence can give the impression that everything is correct, even when the original meaning has been changed.
This is why human checking remains necessary, especially for sensitive texts.
A new translator for a new market
AI will probably change the translation market. Some routine translation tasks may become less common, while other tasks may become more important.
The translator may spend more time on post-editing, quality control, terminology management and cultural adaptation. The translator may also become more of a language consultant or cultural mediator.
This means that the translator of the future will need new skills.
Good knowledge of languages will remain necessary, but it will not be enough. Translators will also need to understand AI tools and know how to use them correctly. More importantly, they will need to know when not to trust them.
A translator who simply accepts everything produced by AI may create serious problems. A translator who knows how to question, check and improve AI output can use the technology as a very useful assistant.
The same issue applies to universities. Translation students cannot be prepared for the profession in exactly the same way as students were prepared twenty or thirty years ago. Translation education has to include technology, AI, quality assessment and ethical issues, while keeping the strong language and cultural skills that have always been important.
A partnership, not a war
For me, the strongest message of this chapter is that the future of translation does not have to be a war between humans and machines.
Technology has been changing translation for a very long time. The computer did not destroy translation. The internet did not destroy translation. Computer-assisted translation did not destroy translation. Instead, translators learned to use these technologies.
AI is much more powerful, and its impact will probably be much greater. Some parts of the profession will certainly change, and some traditional jobs may decrease. The book itself recognizes these challenges. At the same time, it points to new opportunities for human translators in areas where cultural understanding, creativity, responsibility and professional judgment are needed.
So, perhaps the wrong question is: “Will AI replace translators?”
The better question is: “What kind of translator will be needed in the age of AI?”
The answer may be a translator who knows languages, understands cultures, has good professional judgment and knows how to work with intelligent machines.
AI can translate faster. It can process more information. It can help with many routine tasks.
But translation is ultimately about communication between people.
And as long as people have different languages, different cultures and different ways of expressing ideas, there will still be a need for people who can understand one world and explain it to another.
The future translator may be very different from the translator of the past.
But ‘different’ does not necessarily mean ‘unnecessary’.
By Dr Mahmoud Kamel
Professor at the Academy of Arts
